Authors: Tingkai Zhang, Chaoyu Chen, Cong Liao, Jun Wang, Xudong Zhao, Hang Yu, Jianchao Wang, Jianguo Li, Wenhui Shi
Abstract: Text-to-SQL conversion is a critical innovation, simplifying the transition from complex SQL to intuitive natural language queries, especially significant given SQL's prevalence in the job market across various roles. The rise of Large Language Models (LLMs) like GPT-3.5 and GPT-4 has greatly advanced this field, offering improved natural language understanding and the ability to generate nuanced SQL statements. However, the potential of open-source LLMs in Text-to-SQL applications remains underexplored, with many frameworks failing to leverage their full capabilities, particularly in handling complex database queries and incorporating feedback for iterative refinement. Addressing these limitations, this paper introduces SQLfuse, a robust system integrating open-source LLMs with a suite of tools to enhance Text-to-SQL translation's accuracy and usability. SQLfuse features four modules: schema mining, schema linking, SQL generation, and a SQL critic module, to not only generate but also continuously enhance SQL query quality. Demonstrated by its leading performance on the Spider Leaderboard and deployment by Ant Group, SQLfuse showcases the practical merits of open-source LLMs in diverse business contexts.
Authors: Sean Wu, Michael Koo, Li Yo Kao, Andy Black, Lesley Blum, Fabien Scalzo, Ira Kurtz
Abstract: Open-source LLMs have shown great potential as fine-tuned chatbots, and demonstrate robust abilities in reasoning and surpass many existing benchmarks. Retrieval-Augmented Generation (RAG) is a technique for improving the performance of LLMs on tasks that the models weren't explicitly trained on, by leveraging external knowledge databases. Numerous studies have demonstrated the effectiveness of RAG to more successfully accomplish downstream tasks when using vector datasets that consist of relevant background information. It has been implicitly assumed by those in the field that if adversarial background information is utilized in this context, that the success of using a RAG-based approach would be nonexistent or even negatively impact the results. To address this assumption, we tested several open-source LLMs on the ability of RAG to improve their success in answering multiple-choice questions (MCQ) in the medical subspecialty field of Nephrology. Unlike previous studies, we examined the effect of RAG in utilizing both relevant and adversarial background databases. We set up several open-source LLMs, including Llama 3, Phi-3, Mixtral 8x7b, Zephyr$\beta$, and Gemma 7B Instruct, in a zero-shot RAG pipeline. As adversarial sources of information, text from the Bible and a Random Words generated database were used for comparison. Our data show that most of the open-source LLMs improve their multiple-choice test-taking success as expected when incorporating relevant information vector databases. Surprisingly however, adversarial Bible text significantly improved the success of many LLMs and even random word text improved test taking ability of some of the models. In summary, our results demonstrate for the first time the countertintuitive ability of adversarial information datasets to improve the RAG-based LLM success.
Authors: Rikhiya Ghosh, Oladimeji Farri, Hans-Martin von Stockhausen, Martin Schmitt, George Marica Vasile
Abstract: The healthcare industry is currently experiencing an unprecedented wave of cybersecurity attacks, impacting millions of individuals. With the discovery of thousands of vulnerabilities each month, there is a pressing need to drive the automation of vulnerability assessment processes for medical devices, facilitating rapid mitigation efforts. Generative AI systems have revolutionized various industries, offering unparalleled opportunities for automation and increased efficiency. This paper presents a solution leveraging Large Language Models (LLMs) to learn from historical evaluations of vulnerabilities for the automatic assessment of vulnerabilities in the medical devices industry. This approach is applied within the portfolio of a single manufacturer, taking into account device characteristics, including existing security posture and controls. The primary contributions of this paper are threefold. Firstly, it provides a detailed examination of the best practices for training a vulnerability Language Model (LM) in an industrial context. Secondly, it presents a comprehensive comparison and insightful analysis of the effectiveness of Language Models in vulnerability assessment. Finally, it proposes a new human-in-the-loop framework to expedite vulnerability evaluation processes.
Authors: Nilanjana Das, Edward Raff, Manas Gaur
Abstract: Previous research on testing the vulnerabilities in Large Language Models (LLMs) using adversarial attacks has primarily focused on nonsensical prompt injections, which are easily detected upon manual or automated review (e.g., via byte entropy). However, the exploration of innocuous human-understandable malicious prompts augmented with adversarial injections remains limited. In this research, we explore converting a nonsensical suffix attack into a sensible prompt via a situation-driven contextual re-writing. This allows us to show suffix conversion without any gradients, using only LLMs to perform the attacks, and thus better understand the scope of possible risks. We combine an independent, meaningful adversarial insertion and situations derived from movies to check if this can trick an LLM. The situations are extracted from the IMDB dataset, and prompts are defined following a few-shot chain-of-thought prompting. Our approach demonstrates that a successful situation-driven attack can be executed on both open-source and proprietary LLMs. We find that across many LLMs, as few as 1 attempt produces an attack and that these attacks transfer between LLMs. The link to our code is available at \url{https://anonymous.4open.science/r/Situation-Driven-Adversarial-Attacks-7BB1/README.md}.
URLs: https://anonymous.4open.science/r/Situation-Driven-Adversarial-Attacks-7BB1/README.md
Authors: Saurav Muralidharan, Sharath Turuvekere Sreenivas, Raviraj Joshi, Marcin Chochowski, Mostofa Patwary, Mohammad Shoeybi, Bryan Catanzaro, Jan Kautz, Pavlo Molchanov
Abstract: Large language models (LLMs) targeting different deployment scales and sizes are currently produced by training each variant from scratch; this is extremely compute-intensive. In this paper, we investigate if pruning an existing LLM and then re-training it with a fraction (<3%) of the original training data can be a suitable alternative to repeated, full retraining. To this end, we develop a set of practical and effective compression best practices for LLMs that combine depth, width, attention and MLP pruning with knowledge distillation-based retraining; we arrive at these best practices through a detailed empirical exploration of pruning strategies for each axis, methods to combine axes, distillation strategies, and search techniques for arriving at optimal compressed architectures. We use this guide to compress the Nemotron-4 family of LLMs by a factor of 2-4x, and compare their performance to similarly-sized models on a variety of language modeling tasks. Deriving 8B and 4B models from an already pretrained 15B model using our approach requires up to 40x fewer training tokens per model compared to training from scratch; this results in compute cost savings of 1.8x for training the full model family (15B, 8B, and 4B). Minitron models exhibit up to a 16% improvement in MMLU scores compared to training from scratch, perform comparably to other community models such as Mistral 7B, Gemma 7B and Llama-3 8B, and outperform state-of-the-art compression techniques from the literature. We have open-sourced Minitron model weights on Huggingface, with corresponding supplementary material including example code available on GitHub.
Authors: Michael C. Stern, Maria M. Pi\~nango
Abstract: We propose and computationally implement a dynamic neural model of lexical meaning, and experimentally test its behavioral predictions. We demonstrate the architecture and behavior of the model using as a test case the English lexical item 'have', focusing on its polysemous use. In the model, 'have' maps to a semantic space defined by two continuous conceptual dimensions, connectedness and control asymmetry, previously proposed to parameterize the conceptual system for language. The mapping is modeled as coupling between a neural node representing the lexical item and neural fields representing the conceptual dimensions. While lexical knowledge is modeled as a stable coupling pattern, real-time lexical meaning retrieval is modeled as the motion of neural activation patterns between metastable states corresponding to semantic interpretations or readings. Model simulations capture two previously reported empirical observations: (1) contextual modulation of lexical semantic interpretation, and (2) individual variation in the magnitude of this modulation. Simulations also generate a novel prediction that the by-trial relationship between sentence reading time and acceptability should be contextually modulated. An experiment combining self-paced reading and acceptability judgments replicates previous results and confirms the new model prediction. Altogether, results support a novel perspective on lexical polysemy: that the many related meanings of a word are metastable neural activation states that arise from the nonlinear dynamics of neural populations governing interpretation on continuous semantic dimensions.
Authors: Masahiro Suzuki, Hiroki Sakaji
Abstract: Many natural language processing (NLP) tasks in English or general domains are widely available and are often used to evaluate pre-trained language models. In contrast, there are fewer tasks available for languages other than English and for the financial domain. In particular, tasks in Japanese and the financial domain are limited. We construct two large datasets using materials published by a Japanese central government agency. The datasets provide three Japanese financial NLP tasks, which include a 3-class and 12-class classification for categorizing sentences, as well as a 5-class classification task for sentiment analysis. Our datasets are designed to be comprehensive and up-to-date, leveraging an automatic update framework that ensures the latest task datasets are publicly available anytime.
Authors: Cheng-Kuang Wu, Zhi Rui Tam, Chao-Chung Wu, Chieh-Yen Lin, Hung-yi Lee, Yun-Nung Chen
Abstract: In this study, we explore the proactive ability of LLMs to seek user support, using text-to-SQL generation as a case study. We propose metrics to evaluate the trade-off between performance improvements and user burden, and investigate whether LLMs can determine when to request help and examine their performance with varying levels of information availability. Our experiments reveal that without external feedback, many LLMs struggle to recognize their need for additional support. Our findings highlight the importance of external signals and provide insights for future research on improving support-seeking strategies.
Authors: Seyed Mohammad Sadegh Dashti, Amid Khatibi Bardsiri, Mehdi Jafari Shahbazzadeh
Abstract: This research introduces a state-of-the-art Persian spelling correction system that seamlessly integrates deep learning techniques with phonetic analysis, significantly enhancing the accuracy and efficiency of natural language processing (NLP) for Persian. Utilizing a fine-tuned language representation model, our methodology effectively combines deep contextual analysis with phonetic insights, adeptly correcting both non-word and real-word spelling errors. This strategy proves particularly effective in tackling the unique complexities of Persian spelling, including its elaborate morphology and the challenge of homophony. A thorough evaluation on a wide-ranging dataset confirms our system's superior performance compared to existing methods, with impressive F1-Scores of 0.890 for detecting real-word errors and 0.905 for correcting them. Additionally, the system demonstrates a strong capability in non-word error correction, achieving an F1-Score of 0.891. These results illustrate the significant benefits of incorporating phonetic insights into deep learning models for spelling correction. Our contributions not only advance Persian language processing by providing a versatile solution for a variety of NLP applications but also pave the way for future research in the field, emphasizing the critical role of phonetic analysis in developing effective spelling correction system.
Authors: Nemika Tyagi, Mihir Parmar, Mohith Kulkarni, Aswin RRV, Nisarg Patel, Mutsumi Nakamura, Arindam Mitra, Chitta Baral
Abstract: Solving grid puzzles involves a significant amount of logical reasoning. Hence, it is a good domain to evaluate the reasoning capability of a model which can then guide us to improve the reasoning ability of models. However, most existing works evaluate only the final predicted answer of a puzzle, without delving into an in-depth analysis of the LLMs' reasoning chains (such as where they falter) or providing any finer metrics to evaluate them. Since LLMs may rely on simple heuristics or artifacts to predict the final answer, it is crucial to evaluate the generated reasoning chain beyond overall correctness measures, for accurately evaluating the reasoning abilities of LLMs. To this end, we first develop GridPuzzle, an evaluation dataset comprising 274 grid-based puzzles with different complexities. Second, we propose a new error taxonomy derived from manual analysis of reasoning chains from LLMs including GPT-4, Claude-3, Gemini, Mistral, and Llama-2. Then, we develop an LLM-based framework for large-scale subjective evaluation (i.e., identifying errors) and an objective metric, PuzzleEval, to evaluate the correctness of reasoning chains. Evaluating reasoning chains from LLMs leads to several interesting findings. We further show that existing prompting methods used for enhancing models' reasoning abilities do not improve performance on GridPuzzle. This highlights the importance of understanding fine-grained errors and presents a challenge for future research to enhance LLMs' puzzle-solving abilities by developing methods that address these errors. Data and source code are available at https://github.com/Mihir3009/GridPuzzle.
Authors: Seyed Mohammad Sadegh Dashti, Amid Khatibi Bardsiri, Mehdi Jafari Shahbazzadeh
Abstract: Automatic spelling correction stands as a pivotal challenge within the ambit of natural language processing (NLP), demanding nuanced solutions. Traditional spelling correction techniques are typically only capable of detecting and correcting non-word errors, such as typos and misspellings. However, context-sensitive errors, also known as real-word errors, are more challenging to detect because they are valid words that are used incorrectly in a given context. The Persian language, characterized by its rich morphology and complex syntax, presents formidable challenges to automatic spelling correction systems. Furthermore, the limited availability of Persian language resources makes it difficult to train effective spelling correction models. This paper introduces a cutting-edge approach for precise and efficient real-word error correction in Persian text. Our methodology adopts a structured, multi-tiered approach, employing semantic analysis, feature selection, and advanced classifiers to enhance error detection and correction efficacy. The innovative architecture discovers and stores semantic similarities between words and phrases in Persian text. The classifiers accurately identify real-word errors, while the semantic ranking algorithm determines the most probable corrections for real-word errors, taking into account specific spelling correction and context properties such as context, semantic similarity, and edit-distance measures. Evaluations have demonstrated that our proposed method surpasses previous Persian real-word error correction models. Our method achieves an impressive F-measure of 96.6% in the detection phase and an accuracy of 99.1% in the correction phase. These results clearly indicate that our approach is a highly promising solution for automatic real-word error correction in Persian text.
Authors: Sourabrata Mukherjee, Ondrej Du\v{s}ek
Abstract: Text Style Transfer (TST) is a pivotal task in natural language generation to manipulate text style attributes while preserving style-independent content. The attributes targeted in TST can vary widely, including politeness, authorship, mitigation of offensive language, modification of feelings, and adjustment of text formality. TST has become a widely researched topic with substantial advancements in recent years. This paper provides an introductory overview of TST, addressing its challenges, existing approaches, datasets, evaluation measures, subtasks, and applications. This fundamental overview improves understanding of the background and fundamentals of text style transfer.
Authors: Jiayu Lin, Guanrong Chen, Bojun Jin, Chenyang Li, Shutong Jia, Wancong Lin, Yang Sun, Yuhang He, Caihua Yang, Jianzhu Bao, Jipeng Wu, Wen Su, Jinglu Chen, Xinyi Li, Tianyu Chen, Mingjie Han, Shuaiwen Du, Zijian Wang, Jiyin Li, Fuzhong Suo, Hao Wang, Nuanchen Lin, Rui Feng Xu, Long Zhang, Jiuxin Cao, Ting Jin, Zhongyu Wei
Abstract: In this paper we present the results of the AI-Debater 2023 Challenge held by the Chinese Conference on Affect Computing (CCAC 2023), and introduce the related datasets. We organize two tracks to handle the argumentative generation tasks in different scenarios, namely, Counter-Argument Generation (Track 1) and Claim-based Argument Generation (Track 2). Each track is equipped with its distinct dataset and baseline model respectively. In total, 32 competing teams register for the challenge, from which we received 11 successful submissions. In this paper, we will present the results of the challenge and a summary of the systems, highlighting commonalities and innovations among participating systems. Datasets and baseline models of the AI-Debater 2023 Challenge have been already released and can be accessed through the official website of the challenge.
Authors: Shuyu Lei, Lingen Liu, Jiaolong Yang, Yasen Jiao, Yuxiang Yang, Yushu Yang, Xiang Guo
Abstract: Existing auto-regressive language models have demonstrated a remarkable capability to perform a new task with just a few examples in prompt, without requiring any additional training. In order to extend this capability to a multi-modal setting (i.e. speech and language), this paper introduces the Seal model, an abbreviation for speech language model. It incorporates a novel alignment method, in which Kullback-Leibler divergence loss is performed to train a projector that bridges a frozen speech encoder with a frozen language model decoder. The resulting Seal model exhibits robust performance as a few-shot learner on two speech understanding tasks. Additionally, consistency experiments are conducted to validate its robustness on different pre-trained language models.
Authors: Yongxin Huang, Kexin Wang, Goran Glava\v{s}, Iryna Gurevych
Abstract: Multilingual sentence encoders are commonly obtained by training multilingual language models to map sentences from different languages into a shared semantic space. As such, they are subject to curse of multilinguality, a loss of monolingual representational accuracy due to parameter sharing. Another limitation of multilingual sentence encoders is the trade-off between monolingual and cross-lingual performance. Training for cross-lingual alignment of sentence embeddings distorts the optimal monolingual structure of semantic spaces of individual languages, harming the utility of sentence embeddings in monolingual tasks. In this work, we address both issues by modular training of sentence encoders, i.e., by separating monolingual specialization from cross-lingual alignment. We first efficiently train language-specific sentence encoders to avoid negative interference between languages (i.e., the curse). We then align all non-English monolingual encoders to the English encoder by training a cross-lingual alignment adapter on top of each, preventing interference with monolingual specialization from the first step. In both steps, we resort to contrastive learning on machine-translated paraphrase data. Monolingual and cross-lingual evaluations on semantic text similarity/relatedness and multiple-choice QA render our modular solution more effective than multilingual sentence encoders, especially benefiting low-resource languages.
Authors: Quentin Malartic, Nilabhra Roy Chowdhury, Ruxandra Cojocaru, Mugariya Farooq, Giulia Campesan, Yasser Abdelaziz Dahou Djilali, Sanath Narayan, Ankit Singh, Maksim Velikanov, Basma El Amel Boussaha, Mohammed Al-Yafeai, Hamza Alobeidli, Leen Al Qadi, Mohamed El Amine Seddik, Kirill Fedyanin, Reda Alami, Hakim Hacid
Abstract: We introduce Falcon2-11B, a foundation model trained on over five trillion tokens, and its multimodal counterpart, Falcon2-11B-vlm, which is a vision-to-text model. We report our findings during the training of the Falcon2-11B which follows a multi-stage approach where the early stages are distinguished by their context length and a final stage where we use a curated, high-quality dataset. Additionally, we report the effect of doubling the batch size mid-training and how training loss spikes are affected by the learning rate. The downstream performance of the foundation model is evaluated on established benchmarks, including multilingual and code datasets. The foundation model shows strong generalization across all the tasks which makes it suitable for downstream finetuning use cases. For the vision language model, we report the performance on several benchmarks and show that our model achieves a higher average score compared to open-source models of similar size. The model weights and code of both Falcon2-11B and Falcon2-11B-vlm are made available under a permissive license.
Authors: Silviu Pitis, Ziang Xiao, Nicolas Le Roux, Alessandro Sordoni
Abstract: While finetuning language models from pairwise preferences has proven remarkably effective, the underspecified nature of natural language presents critical challenges. Direct preference feedback is uninterpretable, difficult to provide where multidimensional criteria may apply, and often inconsistent, either because it is based on incomplete instructions or provided by diverse principals. To address these challenges, we consider the two-step preference modeling procedure that first resolves the under-specification by selecting a context, and then evaluates preference with respect to the chosen context. We decompose reward modeling error according to these two steps, which suggests that supervising context in addition to context-specific preference may be a viable approach to aligning models with diverse human preferences. For this to work, the ability of models to evaluate context-specific preference is critical. To this end, we contribute context-conditioned preference datasets and accompanying experiments that investigate the ability of language models to evaluate context-specific preference. We use our datasets to (1) show that existing preference models benefit from, but fail to fully consider, added context, (2) finetune a context-aware reward model with context-specific performance exceeding that of GPT-4 and Llama 3 70B on tested datasets, and (3) investigate the value of context-aware preference modeling.
Authors: Shayne Longpre, Robert Mahari, Ariel Lee, Campbell Lund, Hamidah Oderinwale, William Brannon, Nayan Saxena, Naana Obeng-Marnu, Tobin South, Cole Hunter, Kevin Klyman, Christopher Klamm, Hailey Schoelkopf, Nikhil Singh, Manuel Cherep, Ahmad Anis, An Dinh, Caroline Chitongo, Da Yin, Damien Sileo, Deividas Mataciunas, Diganta Misra, Emad Alghamdi, Enrico Shippole, Jianguo Zhang, Joanna Materzynska, Kun Qian, Kush Tiwary, Lester Miranda, Manan Dey, Minnie Liang, Mohammed Hamdy, Niklas Muennighoff, Seonghyeon Ye, Seungone Kim, Shrestha Mohanty, Vipul Gupta, Vivek Sharma, Vu Minh Chien, Xuhui Zhou, Yizhi Li, Caiming Xiong, Luis Villa, Stella Biderman, Hanlin Li, Daphne Ippolito, Sara Hooker, Jad Kabbara, Sandy Pentland
Abstract: General-purpose artificial intelligence (AI) systems are built on massive swathes of public web data, assembled into corpora such as C4, RefinedWeb, and Dolma. To our knowledge, we conduct the first, large-scale, longitudinal audit of the consent protocols for the web domains underlying AI training corpora. Our audit of 14,000 web domains provides an expansive view of crawlable web data and how consent preferences to use it are changing over time. We observe a proliferation of AI-specific clauses to limit use, acute differences in restrictions on AI developers, as well as general inconsistencies between websites' expressed intentions in their Terms of Service and their robots.txt. We diagnose these as symptoms of ineffective web protocols, not designed to cope with the widespread re-purposing of the internet for AI. Our longitudinal analyses show that in a single year (2023-2024) there has been a rapid crescendo of data restrictions from web sources, rendering ~5%+ of all tokens in C4, or 28%+ of the most actively maintained, critical sources in C4, fully restricted from use. For Terms of Service crawling restrictions, a full 45% of C4 is now restricted. If respected or enforced, these restrictions are rapidly biasing the diversity, freshness, and scaling laws for general-purpose AI systems. We hope to illustrate the emerging crisis in data consent, foreclosing much of the open web, not only for commercial AI, but non-commercial AI and academic purposes.
Authors: Apurv Verma, Satyapriya Krishna, Sebastian Gehrmann, Madhavan Seshadri, Anu Pradhan, Tom Ault, Leslie Barrett, David Rabinowitz, John Doucette, NhatHai Phan
Abstract: Creating secure and resilient applications with large language models (LLM) requires anticipating, adjusting to, and countering unforeseen threats. Red-teaming has emerged as a critical technique for identifying vulnerabilities in real-world LLM implementations. This paper presents a detailed threat model and provides a systematization of knowledge (SoK) of red-teaming attacks on LLMs. We develop a taxonomy of attacks based on the stages of the LLM development and deployment process and extract various insights from previous research. In addition, we compile methods for defense and practical red-teaming strategies for practitioners. By delineating prominent attack motifs and shedding light on various entry points, this paper provides a framework for improving the security and robustness of LLM-based systems.
Authors: Dmitrii Galimzianov, Viacheslav Vyshegorodtsev
Abstract: Interruption in a dialogue occurs when the listener begins their speech before the current speaker finishes speaking. Interruptions can be broadly divided into two groups: cooperative (when the listener wants to support the speaker), and competitive (when the listener tries to take control of the conversation against the speaker's will). A system that automatically classifies interruptions can be used in call centers, specifically in the tasks of customer satisfaction monitoring and agent monitoring. In this study, we developed a text-based interruption classification model by preparing an in-house dataset consisting of ASR-transcribed customer support telephone dialogues in Russian. We fine-tuned Conversational RuBERT on our dataset and optimized hyperparameters, and the model performed well. With further improvements, the proposed model can be applied to automatic monitoring systems.
Authors: Desta Haileselassie Hagos, Rick Battle, Danda B. Rawat
Abstract: The emergence of Generative Artificial Intelligence (AI) and Large Language Models (LLMs) has marked a new era of Natural Language Processing (NLP), introducing unprecedented capabilities that are revolutionizing various domains. This paper explores the current state of these cutting-edge technologies, demonstrating their remarkable advancements and wide-ranging applications. Our paper contributes to providing a holistic perspective on the technical foundations, practical applications, and emerging challenges within the evolving landscape of Generative AI and LLMs. We believe that understanding the generative capabilities of AI systems and the specific context of LLMs is crucial for researchers, practitioners, and policymakers to collaboratively shape the responsible and ethical integration of these technologies into various domains. Furthermore, we identify and address main research gaps, providing valuable insights to guide future research endeavors within the AI research community.
Authors: Antonis Antoniades, Xinyi Wang, Yanai Elazar, Alfonso Amayuelas, Alon Albalak, Kexun Zhang, William Yang Wang
Abstract: Despite the proven utility of large language models (LLMs) in real-world applications, there remains a lack of understanding regarding how they leverage their large-scale pretraining text corpora to achieve such capabilities. In this work, we investigate the interplay between generalization and memorization in pretrained LLMs at scale, through a comprehensive $n$-gram analysis of their training data. Our experiments focus on three general task types: translation, question-answering, and multiple-choice reasoning. With various sizes of open-source LLMs and their pretraining corpora, we observe that as the model size increases, the task-relevant $n$-gram pair data becomes increasingly important, leading to improved task performance, decreased memorization, stronger generalization, and emergent abilities. Our results support the hypothesis that LLMs' capabilities emerge from a delicate balance of memorization and generalization with sufficient task-related pretraining data, and point the way to larger-scale analyses that could further improve our understanding of these models.
Authors: Mengru Wang, Yunzhi Yao, Ziwen Xu, Shuofei Qiao, Shumin Deng, Peng Wang, Xiang Chen, Jia-Chen Gu, Yong Jiang, Pengjun Xie, Fei Huang, Huajun Chen, Ningyu Zhang
Abstract: Understanding knowledge mechanisms in Large Language Models (LLMs) is crucial for advancing towards trustworthy AGI. This paper reviews knowledge mechanism analysis from a novel taxonomy including knowledge utilization and evolution. Knowledge utilization delves into the mechanism of memorization, comprehension and application, and creation. Knowledge evolution focuses on the dynamic progression of knowledge within individual and group LLMs. Moreover, we discuss what knowledge LLMs have learned, the reasons for the fragility of parametric knowledge, and the potential dark knowledge (hypothesis) that will be challenging to address. We hope this work can help understand knowledge in LLMs and provide insights for future research.
Authors: Sarah Wiegreffe, Oyvind Tafjord, Yonatan Belinkov, Hannaneh Hajishirzi, Ashish Sabharwal
Abstract: Multiple-choice question answering (MCQA) is a key competence of performant transformer language models that is tested by mainstream benchmarks. However, recent evidence shows that models can have quite a range of performance, particularly when the task format is diversified slightly (such as by shuffling answer choice order). In this work we ask: how do successful models perform formatted MCQA? We employ vocabulary projection and activation patching methods to localize key hidden states that encode relevant information for predicting the correct answer. We find that prediction of a specific answer symbol is causally attributed to a single middle layer, and specifically its multi-head self-attention mechanism. We show that subsequent layers increase the probability of the predicted answer symbol in vocabulary space, and that this probability increase is associated with a sparse set of attention heads with unique roles. We additionally uncover differences in how different models adjust to alternative symbols. Finally, we demonstrate that a synthetic task can disentangle sources of model error to pinpoint when a model has learned formatted MCQA, and show that an inability to separate answer symbol tokens in vocabulary space is a property of models unable to perform formatted MCQA tasks.
Authors: EunJeong Hwang, Yichao Zhou, James Bradley Wendt, Beliz Gunel, Nguyen Vo, Jing Xie, Sandeep Tata
Abstract: Large language models (LLMs) often struggle with processing extensive input contexts, which can lead to redundant, inaccurate, or incoherent summaries. Recent methods have used unstructured memory to incrementally process these contexts, but they still suffer from information overload due to the volume of unstructured data handled. In our study, we introduce structured knowledge representations ($GU_{json}$), which significantly improve summarization performance by 40% and 14% across two public datasets. Most notably, we propose the Chain-of-Key strategy ($CoK_{json}$) that dynamically updates or augments these representations with new information, rather than recreating the structured memory for each new source. This method further enhances performance by 7% and 4% on the datasets.
Authors: Adib Mosharrof, A. B. Siddique
Abstract: Task-oriented dialog (TOD) systems play a crucial role in facilitating efficient interactions between users and machines by focusing on achieving specific goals through natural language communication. These systems traditionally rely on manually annotated metadata, such as dialog states and policy annotations, which is labor-intensive, expensive, inconsistent, and prone to errors, thereby limiting the potential to leverage the vast amounts of available conversational data. A critical aspect of TOD systems involves accessing and integrating information from external sources to effectively engage users. The process of determining when and how to query external resources represents a fundamental challenge in system design, however existing approaches expect this information to provided in the context. In this paper, we introduce Natural Language Task Oriented Dialog System (NL-ToD), a novel model that removes the dependency on manually annotated turn-wise data by utilizing dialog history and domain schemas to create a Zero Shot Generalizable TOD system. We also incorporate query generation as a core task of the system, where the output of the system could be a response to the user or an API query to communicate with an external resource. To achieve a more granular analysis of the system output, we classify the output into multiple categories: slot filling, retrieval, and query generation. Our analysis reveals that slot filling is the most challenging TOD task for all models. Experimental results on three popular TOD datasets (SGD, KETOD and BiToD) shows the effectiveness of our approach as NL-ToD outperforms state-of-the-art approaches, particularly with a \textbf{31.4\%} and \textbf{82.1\%} improvement in the BLEU-4 score on the SGD and KETOD dataset.
Authors: Jinfeng Wei, Xiaofeng Zhang
Abstract: In this work, we introduce DOPRA, a novel approach designed to mitigate hallucinations in multi-modal large language models (MLLMs). Unlike existing solutions that typically involve costly supplementary training data or the integration of external knowledge sources, DOPRA innovatively addresses hallucinations by decoding specific weighted layer penalties and redistribution, offering an economical and effective solution without additional resources. DOPRA is grounded in unique insights into the intrinsic mechanisms controlling hallucinations within MLLMs, especially the models' tendency to over-rely on a subset of summary tokens in the self-attention matrix, neglecting critical image-related information. This phenomenon is particularly pronounced in certain strata. To counteract this over-reliance, DOPRA employs a strategy of weighted overlay penalties and redistribution in specific layers, such as the 12th layer, during the decoding process. Furthermore, DOPRA includes a retrospective allocation process that re-examines the sequence of generated tokens, allowing the algorithm to reallocate token selection to better align with the actual image content, thereby reducing the incidence of hallucinatory descriptions in auto-generated captions. Overall, DOPRA represents a significant step forward in improving the output quality of MLLMs by systematically reducing hallucinations through targeted adjustments during the decoding process.
Authors: Gokcen Gokceoglu, Devrim Cavusoglu, Emre Akbas, \"Ozen Nergis Dolcerocca
Abstract: This paper introduces a multi-level, multi-label text classification dataset comprising over 3000 documents. The dataset features literary and critical texts from 19th-century Ottoman Turkish and Russian. It is the first study to apply large language models (LLMs) to this dataset, sourced from prominent literary periodicals of the era. The texts have been meticulously organized and labeled. This was done according to a taxonomic framework that takes into account both their structural and semantic attributes. Articles are categorized and tagged with bibliometric metadata by human experts. We present baseline classification results using a classical bag-of-words (BoW) naive Bayes model and three modern LLMs: multilingual BERT, Falcon, and Llama-v2. We found that in certain cases, Bag of Words (BoW) outperforms Large Language Models (LLMs), emphasizing the need for additional research, especially in low-resource language settings. This dataset is expected to be a valuable resource for researchers in natural language processing and machine learning, especially for historical and low-resource languages. The dataset is publicly available^1.
Authors: Minwoo Lee, Hyukhun Koh, Minsung Kim, Kyomin Jung
Abstract: In machine translation, the problem of ambiguously gendered input has been pointed out, where the gender of an entity is not available in the source sentence. To address this ambiguity issue, the task of controlled translation that takes the gender of the ambiguous entity as additional input have been proposed. However, most existing works have only considered a simplified setup of one target gender for input. In this paper, we tackle controlled translation in a more realistic setting of inputs with multiple entities and propose Gender-of-Entity (GoE) prompting method for LLMs. Our proposed method instructs the model with fine-grained entity-level gender information to translate with correct gender inflections. By utilizing four evaluation benchmarks, we investigate the controlled translation capability of LLMs in multiple dimensions and find that LLMs reach state-of-the-art performance in controlled translation. Furthermore, we discover an emergence of gender interference phenomenon when controlling the gender of multiple entities. Finally, we address the limitations of existing gender accuracy evaluation metrics and propose leveraging LLMs as an evaluator for gender inflection in machine translation.
Authors: Gilad Yehudai, Haim Kaplan, Asma Ghandeharioun, Mor Geva, Amir Globerson
Abstract: Large language models based on the transformer architectures can solve highly complex tasks. But are there simple tasks that such models cannot solve? Here we focus on very simple counting tasks, that involve counting how many times a token in the vocabulary have appeared in a string. We show that if the dimension of the transformer state is linear in the context length, this task can be solved. However, the solution we propose does not scale beyond this limit, and we provide theoretical arguments for why it is likely impossible for a size limited transformer to implement this task. Our empirical results demonstrate the same phase-transition in performance, as anticipated by the theoretical argument. Our results demonstrate the importance of understanding how transformers can solve simple tasks.
Authors: Xiaoran Liu, Qipeng Guo, Yuerong Song, Zhigeng Liu, Kai Lv, Hang Yan, Linlin Li, Qun Liu, Xipeng Qiu
Abstract: The maximum supported context length is a critical bottleneck limiting the practical application of the Large Language Model (LLM). Although existing length extrapolation methods can extend the context of LLMs to millions of tokens, these methods all have an explicit upper bound. In this work, we propose LongCache, a training-free approach that enables LLM to support an infinite context with finite context scope, through full-context cache selection and training-free integration. This effectively frees LLMs from the length extrapolation issue. We validate LongCache on the LongBench and L-Eval and demonstrate its performance is on par with traditional full-attention mechanisms. Furthermore, we have applied LongCache on mainstream LLMs, including LLaMA3 and Mistral-v0.3, enabling them to support context lengths of at least 400K in Needle-In-A-Haystack tests. We will improve the efficiency of LongCache by GPU-aware optimization soon.
Authors: Liang Shi, Zhengju Tang, Zhi Yang
Abstract: The increasing volume of data stored in relational databases has led to the need for efficient querying and utilization of this data in various sectors. However, writing SQL queries requires specialized knowledge, which poses a challenge for non-professional users trying to access and query databases. Text-to-SQL parsing solves this issue by converting natural language queries into SQL queries, thus making database access more accessible for non-expert users. To take advantage of the recent developments in Large Language Models (LLMs), a range of new methods have emerged, with a primary focus on prompt engineering and fine-tuning. This survey provides a comprehensive overview of LLMs in text-to-SQL tasks, discussing benchmark datasets, prompt engineering, fine-tuning methods, and future research directions. We hope this review will enable readers to gain a broader understanding of the recent advances in this field and offer some insights into its future trajectory.
Authors: Rylan Schaeffer, Dan Valentine, Luke Bailey, James Chua, Crist\'obal Eyzaguirre, Zane Durante, Joe Benton, Brando Miranda, Henry Sleight, John Hughes, Rajashree Agrawal, Mrinank Sharma, Scott Emmons, Sanmi Koyejo, Ethan Perez
Abstract: The integration of new modalities into frontier AI systems offers exciting capabilities, but also increases the possibility such systems can be adversarially manipulated in undesirable ways. In this work, we focus on a popular class of vision-language models (VLMs) that generate text outputs conditioned on visual and textual inputs. We conducted a large-scale empirical study to assess the transferability of gradient-based universal image "jailbreaks" using a diverse set of over 40 open-parameter VLMs, including 18 new VLMs that we publicly release. Overall, we find that transferable gradient-based image jailbreaks are extremely difficult to obtain. When an image jailbreak is optimized against a single VLM or against an ensemble of VLMs, the jailbreak successfully jailbreaks the attacked VLM(s), but exhibits little-to-no transfer to any other VLMs; transfer is not affected by whether the attacked and target VLMs possess matching vision backbones or language models, whether the language model underwent instruction-following and/or safety-alignment training, or many other factors. Only two settings display partially successful transfer: between identically-pretrained and identically-initialized VLMs with slightly different VLM training data, and between different training checkpoints of a single VLM. Leveraging these results, we then demonstrate that transfer can be significantly improved against a specific target VLM by attacking larger ensembles of "highly-similar" VLMs. These results stand in stark contrast to existing evidence of universal and transferable text jailbreaks against language models and transferable adversarial attacks against image classifiers, suggesting that VLMs may be more robust to gradient-based transfer attacks.
Authors: Gaurav Verma, Rynaa Grover, Jiawei Zhou, Binny Mathew, Jordan Kraemer, Munmun De Choudhury, Srijan Kumar
Abstract: Violence-provoking speech -- speech that implicitly or explicitly promotes violence against the members of the targeted community, contributed to a massive surge in anti-Asian crimes during the pandemic. While previous works have characterized and built tools for detecting other forms of harmful speech, like fear speech and hate speech, our work takes a community-centric approach to studying anti-Asian violence-provoking speech. Using data from ~420k Twitter posts spanning a 3-year duration (January 1, 2020 to February 1, 2023), we develop a codebook to characterize anti-Asian violence-provoking speech and collect a community-crowdsourced dataset to facilitate its large-scale detection using state-of-the-art classifiers. We contrast the capabilities of natural language processing classifiers, ranging from BERT-based to LLM-based classifiers, in detecting violence-provoking speech with their capabilities to detect anti-Asian hateful speech. In contrast to prior work that has demonstrated the effectiveness of such classifiers in detecting hateful speech ($F_1 = 0.89$), our work shows that accurate and reliable detection of violence-provoking speech is a challenging task ($F_1 = 0.69$). We discuss the implications of our findings, particularly the need for proactive interventions to support Asian communities during public health crises. The resources related to the study are available at https://claws-lab.github.io/violence-provoking-speech/.
URLs: https://claws-lab.github.io/violence-provoking-speech/.
Authors: Kian Ahrabian, Xihui Lin, Barun Patra, Vishrav Chaudhary, Alon Benhaim, Jay Pujara, Xia Song
Abstract: With the growing utilization of large language models (LLMs) across domains, alignment towards human preferences has become one of the most critical aspects of training models. At the forefront of state-of-the-art human alignment methods are preference optimization methods (*PO). However, prior research has often concentrated on identifying the best-performing method, typically involving a grid search over hyperparameters, which can be impractical for general practitioners. In this paper, we aim to identify the algorithm that, while being performant, is simultaneously more robust to varying hyperparameters, thereby increasing the likelihood of achieving better results. We focus on a realistic out-of-distribution (OOD) scenario that mirrors real-world applications of human alignment, offering practical insights into the strengths and weaknesses of these methods. Furthermore, to better understand the shortcomings of generations from the different methods, we analyze the model generations through the lens of KL divergence of the SFT model and the response length statistics. Our analysis reveals that the widely adopted DPO method consistently produces lengthy responses of inferior quality that are very close to the SFT responses. Motivated by these findings, we propose an embarrassingly simple extension to the DPO algorithm, LN-DPO, resulting in more concise responses without sacrificing quality compared to the policy obtained by vanilla DPO.
Authors: Jipeng Zhang, Yaxuan Qin, Renjie Pi, Weizhong Zhang, Rui Pan, Tong Zhang
Abstract: Instruction tuning has achieved unprecedented success in NLP, turning large language models into versatile chatbots. However, the increasing variety and volume of instruction datasets demand significant computational resources. To address this, it is essential to extract a small and highly informative subset (i.e., Coreset) that achieves comparable performance to the full dataset. Achieving this goal poses non-trivial challenges: 1) data selection requires accurate data representations that reflect the training samples' quality, 2) considering the diverse nature of instruction datasets, and 3) ensuring the efficiency of the coreset selection algorithm for large models. To address these challenges, we propose Task-Agnostic Gradient Clustered COreset Selection (TAGCOS). Specifically, we leverage sample gradients as the data representations, perform clustering to group similar data, and apply an efficient greedy algorithm for coreset selection. Experimental results show that our algorithm, selecting only 5% of the data, surpasses other unsupervised methods and achieves performance close to that of the full dataset.
Authors: Anisha Saha, Abhisek Tiwari, Sai Ruthvik, Sriparna Saha
Abstract: We often summarize a multi-party conversation in two stages: chunking with homogeneous units and summarizing the chunks. Thus, we hypothesize that there exists a correlation between homogeneous speaker chunking and overall summarization tasks. In this work, we investigate the effectiveness of a multi-faceted approach that simultaneously produces summaries of medical concerns, doctor impressions, and an overall view. We introduce a multi-modal, multi-tasking, knowledge-infused medical dialogue summary generation (MMK-Summation) model, which is incorporated with adapter-based fine-tuning through a gated mechanism for multi-modal information integration. The model, MMK-Summation, takes dialogues as input, extracts pertinent external knowledge based on the context, integrates the knowledge and visual cues from the dialogues into the textual content, and ultimately generates concise summaries encompassing medical concerns, doctor impressions, and a comprehensive overview. The introduced model surpasses multiple baselines and traditional summarization models across all evaluation metrics (including human evaluation), which firmly demonstrates the efficacy of the knowledge-guided multi-tasking, multimodal medical conversation summarization. The code is available at https://github.com/NLP-RL/MMK-Summation.
Authors: Erik Cambria, Lorenzo Malandri, Fabio Mercorio, Navid Nobani, Andrea Seveso
Abstract: In this survey, we address the key challenges in Large Language Models (LLM) research, focusing on the importance of interpretability. Driven by increasing interest from AI and business sectors, we highlight the need for transparency in LLMs. We examine the dual paths in current LLM research and eXplainable Artificial Intelligence (XAI): enhancing performance through XAI and the emerging focus on model interpretability. Our paper advocates for a balanced approach that values interpretability equally with functional advancements. Recognizing the rapid development in LLM research, our survey includes both peer-reviewed and preprint (arXiv) papers, offering a comprehensive overview of XAI's role in LLM research. We conclude by urging the research community to advance both LLM and XAI fields together.
Authors: Liwen Sun, James Zhao, Megan Han, Chenyan Xiong
Abstract: Multimodal foundation models hold significant potential for automating radiology report generation, thereby assisting clinicians in diagnosing cardiac diseases. However, generated reports often suffer from serious factual inaccuracy. In this paper, we introduce a fact-aware multimodal retrieval-augmented pipeline in generating accurate radiology reports (FactMM-RAG). We first leverage RadGraph to mine factual report pairs, then integrate factual knowledge to train a universal multimodal retriever. Given a radiology image, our retriever can identify high-quality reference reports to augment multimodal foundation models, thus enhancing the factual completeness and correctness of report generation. Experiments on two benchmark datasets show that our multimodal retriever outperforms state-of-the-art retrievers on both language generation and radiology-specific metrics, up to 6.5% and 2% score in F1CheXbert and F1RadGraph. Further analysis indicates that employing our factually-informed training strategy imposes an effective supervision signal, without relying on explicit diagnostic label guidance, and successfully propagates fact-aware capabilities from the multimodal retriever to the multimodal foundation model in radiology report generation.
Authors: Kuan-Yen Lin
Abstract: Understanding rich dialogues often requires NLP systems to access relevant commonsense persona knowledge, but retrieving this knowledge is challenging due to complex contexts and the implicit nature of commonsense. This paper presents our approach to the Commonsense Persona Knowledge Linking (CPKL) challenge, addressing the critical need for integrating persona and commonsense knowledge in open-domain dialogue systems. We introduce SynCPKL Pipeline, a pipeline that leverages Large Language Models to generate high-quality synthetic datasets for training commonsense persona knowledge linkers. To demonstrate the efficacy of our approach, we present SynCPKL, a new dataset specifically designed for this task. Our experiments validate the effectiveness of SynCPKL for training commonsense persona knowledge linkers. Additionally, our top-performing model, Derberta-SynCPKL, secured first place in the CPKL challenge by a 16% improvement in F1 score. We released both SynCPKL and Derberta-SynCPKL at https://github.com/irislin1006/CPKL.
Authors: Guangliang Liu, Haitao Mao, Jiliang Tang, Kristen Marie Johnson
Abstract: Large Language Models (LLMs) are capable of producing content that perpetuates stereotypes, discrimination, and toxicity. The recently proposed moral self-correction is a computationally efficient method for reducing harmful content in the responses of LLMs. However, the process of how injecting self-correction instructions can modify the behavior of LLMs remains under-explored. In this paper, we explore the effectiveness of moral self-correction by answering three research questions: (1) In what scenarios does moral self-correction work? (2) What are the internal mechanisms of LLMs, e.g., hidden states, that are influenced by moral self-correction instructions? (3) Is intrinsic moral self-correction actually superficial? We argue that self-correction can help LLMs find a shortcut to more morally correct output, rather than truly reducing the immorality stored in hidden states. Through empirical investigation with tasks of language generation and multi-choice question answering, we conclude: (i) LLMs exhibit good performance across both tasks, and self-correction instructions are particularly beneficial when the correct answer is already top-ranked; (ii) The morality levels in intermediate hidden states are strong indicators as to whether one instruction would be more effective than another; (iii) Based on our analysis of intermediate hidden states and task case studies of self-correction behaviors, we are first to propose the hypothesis that intrinsic moral self-correction is in fact superficial.
Authors: Senbin Zhu, Hanjie Zhao, Xingren Wang, Shanhong Liu, Yuxiang Jia, Hongying Zan
Abstract: The DimABSA task requires fine-grained sentiment intensity prediction for restaurant reviews, including scores for Valence and Arousal dimensions for each Aspect Term. In this study, we propose a Coarse-to-Fine In-context Learning(CFICL) method based on the Baichuan2-7B model for the DimABSA task in the SIGHAN 2024 workshop. Our method improves prediction accuracy through a two-stage optimization process. In the first stage, we use fixed in-context examples and prompt templates to enhance the model's sentiment recognition capability and provide initial predictions for the test data. In the second stage, we encode the Opinion field using BERT and select the most similar training data as new in-context examples based on similarity. These examples include the Opinion field and its scores, as well as related opinion words and their average scores. By filtering for sentiment polarity, we ensure that the examples are consistent with the test data. Our method significantly improves prediction accuracy and consistency by effectively utilizing training data and optimizing in-context examples, as validated by experimental results.
Authors: David Heineman, Yao Dou, Wei Xu
Abstract: While instruction fine-tuned LLMs are effective text generators, sensitivity to prompt construction makes performance unstable and sub-optimal in practice. Relying on a single "best" prompt cannot capture all differing approaches to a generation problem. Using this observation, we propose multi-prompt decoding, where many candidate generations are decoded from a prompt bank at inference-time. To ensemble candidates, we use Minimum Bayes Risk (MBR) decoding, which selects a final output using a trained value metric. We show multi-prompt improves MBR across a comprehensive set of conditional generation tasks, and show this is a result of estimating a more diverse and higher quality candidate space than that of a single prompt. Further experiments confirm multi-prompt improves generation across tasks, models and metrics.
Authors: Chunyang Li, Hao Peng, Xiaozhi Wang, Yunjia Qi, Lei Hou, Bin Xu, Juanzi Li
Abstract: Event Factuality Detection (EFD) task determines the factuality of textual events, i.e., classifying whether an event is a fact, possibility, or impossibility, which is essential for faithfully understanding and utilizing event knowledge. However, due to the lack of high-quality large-scale data, event factuality detection is under-explored in event understanding research, which limits the development of EFD community. To address these issues and provide faithful event understanding, we introduce MAVEN-Fact, a large-scale and high-quality EFD dataset based on the MAVEN dataset. MAVEN-Fact includes factuality annotations of 112,276 events, making it the largest EFD dataset. Extensive experiments demonstrate that MAVEN-Fact is challenging for both conventional fine-tuned models and large language models (LLMs). Thanks to the comprehensive annotations of event arguments and relations in MAVEN, MAVEN-Fact also supports some further analyses and we find that adopting event arguments and relations helps in event factuality detection for fine-tuned models but does not benefit LLMs. Furthermore, we preliminarily study an application case of event factuality detection and find it helps in mitigating event-related hallucination in LLMs. Our dataset and codes can be obtained from \url{https://github.com/lcy2723/MAVEN-FACT}
Authors: Yuan Pu, Zhuolun He, Tairu Qiu, Haoyuan Wu, Bei Yu
Abstract: Retrieval augmented generation (RAG) enhances the accuracy and reliability of generative AI models by sourcing factual information from external databases, which is extensively employed in document-grounded question-answering (QA) tasks. Off-the-shelf RAG flows are well pretrained on general-purpose documents, yet they encounter significant challenges when being applied to knowledge-intensive vertical domains, such as electronic design automation (EDA). This paper addresses such issue by proposing a customized RAG framework along with three domain-specific techniques for EDA tool documentation QA, including a contrastive learning scheme for text embedding model fine-tuning, a reranker distilled from proprietary LLM, and a generative LLM fine-tuned with high-quality domain corpus. Furthermore, we have developed and released a documentation QA evaluation benchmark, ORD-QA, for OpenROAD, an advanced RTL-to-GDSII design platform. Experimental results demonstrate that our proposed RAG flow and techniques have achieved superior performance on ORD-QA as well as on a commercial tool, compared with state-of-the-arts. The ORD-QA benchmark and the training dataset for our customized RAG flow are open-source at https://github.com/lesliepy99/RAG-EDA.
Authors: Mengxian Lyu, Cheng Peng, Daniel Paredes, Ziyi Chen, Aokun Chen, Jiang Bian, Yonghui Wu
Abstract: Automatic generation of discharge summaries presents significant challenges due to the length of clinical documentation, the dispersed nature of patient information, and the diverse terminology used in healthcare. This paper presents a hybrid solution for generating discharge summary sections as part of our participation in the "Discharge Me!" Challenge at the BioNLP 2024 Shared Task. We developed a two-stage generation method using both extractive and abstractive techniques, in which we first apply name entity recognition (NER) to extract key clinical concepts, which are then used as input for a prompt-tuning-based GatorTronGPT model to generate coherent text for two important sections including "Brief Hospital Course" and "Discharge Instructions". Our system was ranked 5th in this challenge, achieving an overall score of 0.284. The results demonstrate the effectiveness of our hybrid solution in improving the quality of automated discharge section generation.
Authors: Luyu Qiu, Jianing Li, Chi Su, Chen Jason Zhang, Lei Chen
Abstract: Transformer-based large language models have achieved remarkable performance across various natural language processing tasks. However, they often struggle with seemingly easy tasks like arithmetic despite their vast capabilities. This stark disparity raise human's concerns about their safe and ethical use, hinder their widespread adoption.In this paper, we focus on a typical arithmetic task, integer multiplication, to explore and explain the imperfection of transformers in this domain. We provide comprehensive analysis of a vanilla transformer trained to perform n-digit integer multiplication. Our observations indicate that the model decomposes multiplication task into multiple parallel subtasks, sequentially optimizing each subtask for each digit to complete the final multiplication. Based on observation and analysis, we infer the reasons of transformers deficiencies in multiplication tasks lies in their difficulty in calculating successive carryovers and caching intermediate results, and confirmed this inference through experiments. Guided by these findings, we propose improvements to enhance transformers performance on multiplication tasks. These enhancements are validated through rigorous testing and mathematical modeling, not only enhance transformer's interpretability, but also improve its performance, e.g., we achieve over 99.9% accuracy on 5-digit integer multiplication with a tiny transformer, outperform LLMs GPT-4. Our method contributes to the broader fields of model understanding and interpretability, paving the way for analyzing more complex tasks and Transformer models. This work underscores the importance of explainable AI, helping to build trust in large language models and promoting their adoption in critical applications.
Authors: Rongwu Xu, Zi'an Zhou, Tianwei Zhang, Zehan Qi, Su Yao, Ke Xu, Wei Xu, Han Qiu
Abstract: The common toxicity and societal bias in contents generated by large language models (LLMs) necessitate strategies to reduce harm. Present solutions often demand white-box access to the model or substantial training, which is impractical for cutting-edge commercial LLMs. Moreover, prevailing prompting methods depend on external tool feedback and fail to simultaneously lessen toxicity and bias. Motivated by social psychology principles, we propose a novel strategy named \textbf{perspective-taking prompting (\textsc{PeT})} that inspires LLMs to integrate diverse human perspectives and self-regulate their responses. This self-correction mechanism can significantly diminish toxicity (up to $89\%$) and bias (up to $73\%$) in LLMs' responses. Rigorous evaluations and ablation studies are conducted on two commercial LLMs (ChatGPT and GLM) and three open-source LLMs, revealing \textsc{PeT}'s superiority in producing less harmful responses, outperforming five strong baselines.
Authors: Simon Gonzalez
Abstract: Social Media platforms have offered invaluable opportunities for linguistic research. The availability of up-to-date data, coming from any part in the world, and coming from natural contexts, has allowed researchers to study language in real time. One of the fields that has made great use of social media platforms is Corpus Linguistics. There is currently a wide range of projects which have been able to successfully create corpora from social media. In this paper, we present the development and deployment of a linguistic corpus from Twitter posts in English, coming from 26 news agencies and 27 individuals. The main goal was to create a fully annotated English corpus for linguistic analysis. We include information on morphology and syntax, as well as NLP features such as tokenization, lemmas, and n- grams. The information is presented through a range of powerful visualisations for users to explore linguistic patterns in the corpus. With this tool, we aim to contribute to the area of language technologies applied to linguistic research.
Authors: Simon Gonzalez
Abstract: Pronunciation dictionaries are an important component in the process of speech forced alignment. The accuracy of these dictionaries has a strong effect on the aligned speech data since they help the mapping between orthographic transcriptions and acoustic signals. In this paper, I present the creation of a comprehensive pronunciation dictionary in Spanish (ESPADA) that can be used in most of the dialect variants of Spanish data. Current dictionaries focus on specific regional variants, but with the flexible nature of our tool, it can be readily applied to capture the most common phonetic differences across major dialectal variants. We propose improvements to current pronunciation dictionaries as well as mapping other relevant annotations such as morphological and lexical information. In terms of size, it is currently the most complete dictionary with more than 628,000 entries, representing words from 16 countries. All entries come with their corresponding pronunciations, morphological and lexical tagging, and other relevant information for phonetic analysis: stress patterns, phonotactics, IPA transcriptions, and more. This aims to equip socio-phonetic researchers with a complete open-source tool that enhances dialectal research within socio-phonetic frameworks in the Spanish language.
Authors: M Saiful Bari, Yazeed Alnumay, Norah A. Alzahrani, Nouf M. Alotaibi, Hisham A. Alyahya, Sultan AlRashed, Faisal A. Mirza, Shaykhah Z. Alsubaie, Hassan A. Alahmed, Ghadah Alabduljabbar, Raghad Alkhathran, Yousef Almushayqih, Raneem Alnajim, Salman Alsubaihi, Maryam Al Mansour, Majed Alrubaian, Ali Alammari, Zaki Alawami, Abdulmohsen Al-Thubaity, Ahmed Abdelali, Jeril Kuriakose, Abdalghani Abujabal, Nora Al-Twairesh, Areeb Alowisheq, Haidar Khan
Abstract: We present ALLaM: Arabic Large Language Model, a series of large language models to support the ecosystem of Arabic Language Technologies (ALT). ALLaM is carefully trained considering the values of language alignment and knowledge transfer at scale. Our autoregressive decoder-only architecture models demonstrate how second-language acquisition via vocabulary expansion and pretraining on a mixture of Arabic and English text can steer a model towards a new language (Arabic) without any catastrophic forgetting in the original language (English). Furthermore, we highlight the effectiveness of using parallel/translated data to aid the process of knowledge alignment between languages. Finally, we show that extensive alignment with human preferences can significantly enhance the performance of a language model compared to models of a larger scale with lower quality alignment. ALLaM achieves state-of-the-art performance in various Arabic benchmarks, including MMLU Arabic, ACVA, and Arabic Exams. Our aligned models improve both in Arabic and English from their base aligned models.
Authors: Xiao Liu, Liangzhi Li, Tong Xiang, Fuying Ye, Lu Wei, Wangyue Li, Noa Garcia
Abstract: With the development of large language models (LLMs) like ChatGPT, both their vast applications and potential vulnerabilities have come to the forefront. While developers have integrated multiple safety mechanisms to mitigate their misuse, a risk remains, particularly when models encounter adversarial inputs. This study unveils an attack mechanism that capitalizes on human conversation strategies to extract harmful information from LLMs. We delineate three pivotal strategies: (i) decomposing malicious questions into seemingly innocent sub-questions; (ii) rewriting overtly malicious questions into more covert, benign-sounding ones; (iii) enhancing the harmfulness of responses by prompting models for illustrative examples. Unlike conventional methods that target explicit malicious responses, our approach delves deeper into the nature of the information provided in responses. Through our experiments conducted on GPT-3.5-turbo, GPT-4, and Llama2, our method has demonstrated a marked efficacy compared to conventional attack methods. In summary, this work introduces a novel attack method that outperforms previous approaches, raising an important question: How to discern whether the ultimate intent in a dialogue is malicious?
Authors: Xi Chen, Songyang Zhang, Qibing Bai, Kai Chen, Satoshi Nakamura
Abstract: We introduces LLaST, a framework for building high-performance Large Language model based Speech-to-text Translation systems. We address the limitations of end-to-end speech translation(E2E ST) models by exploring model architecture design and optimization techniques tailored for LLMs. Our approach includes LLM-based speech translation architecture design, ASR-augmented training, multilingual data augmentation, and dual-LoRA optimization. Our approach demonstrates superior performance on the CoVoST-2 benchmark and showcases exceptional scaling capabilities powered by LLMs. We believe this effective method will serve as a strong baseline for speech translation and provide insights for future improvements of the LLM-based speech translation framework. We release the data, code and models in https://github.com/openaudiolab/LLaST.
Authors: Song Wang, Xun Wang, Jie Mei, Yujia Xie, Sean Muarray, Zhang Li, Lingfeng Wu, Si-Qing Chen, Wayne Xiong
Abstract: Hallucination, a phenomenon where large language models (LLMs) produce output that is factually incorrect or unrelated to the input, is a major challenge for LLM applications that require accuracy and dependability. In this paper, we introduce a reliable and high-speed production system aimed at detecting and rectifying the hallucination issue within LLMs. Our system encompasses named entity recognition (NER), natural language inference (NLI), span-based detection (SBD), and an intricate decision tree-based process to reliably detect a wide range of hallucinations in LLM responses. Furthermore, our team has crafted a rewriting mechanism that maintains an optimal mix of precision, response time, and cost-effectiveness. We detail the core elements of our framework and underscore the paramount challenges tied to response time, availability, and performance metrics, which are crucial for real-world deployment of these technologies. Our extensive evaluation, utilizing offline data and live production traffic, confirms the efficacy of our proposed framework and service.
Authors: Daeun Lee, Jaewoong Choi, Hiroshi Mizuseki, Byungju Lee
Abstract: Recent studies have increasingly applied natural language processing (NLP) to automatically extract experimental research data from the extensive battery materials literature. Despite the complex process involved in battery manufacturing -- from material synthesis to cell assembly -- there has been no comprehensive study systematically organizing this information. In response, we propose a language modeling-based protocol, Text-to-Battery Recipe (T2BR), for the automatic extraction of end-to-end battery recipes, validated using a case study on batteries containing LiFePO4 cathode material. We report machine learning-based paper filtering models, screening 2,174 relevant papers from the keyword-based search results, and unsupervised topic models to identify 2,876 paragraphs related to cathode synthesis and 2,958 paragraphs related to cell assembly. Then, focusing on the two topics, two deep learning-based named entity recognition models are developed to extract a total of 30 entities -- including precursors, active materials, and synthesis methods -- achieving F1 scores of 88.18% and 94.61%. The accurate extraction of entities enables the systematic generation of 165 end-toend recipes of LiFePO4 batteries. Our protocol and results offer valuable insights into specific trends, such as associations between precursor materials and synthesis methods, or combinations between different precursor materials. We anticipate that our findings will serve as a foundational knowledge base for facilitating battery-recipe information retrieval. The proposed protocol will significantly accelerate the review of battery material literature and catalyze innovations in battery design and development.
Authors: Zihao Li, Shaoxiong Ji, Timothee Mickus, Vincent Segonne, J\"org Tiedemann
Abstract: Pretrained language models (PLMs) display impressive performances and have captured the attention of the NLP community. Establishing the best practices in pretraining has therefore become a major point of focus for much of NLP research -- especially since the insights developed for monolingual English models need not carry to more complex multilingual. One significant caveat of the current state of the art is that different works are rarely comparable: they often discuss different parameter counts, training data, and evaluation methodology. This paper proposes a comparison of multilingual pretraining objectives in a controlled methodological environment. We ensure that training data and model architectures are comparable, and discuss the downstream performances across 6 languages that we observe in probing and fine-tuning scenarios. We make two key observations: (1) the architecture dictates which pretraining objective is optimal; (2) multilingual translation is a very effective pre-training objective under the right conditions. We make our code, data, and model weights available at \texttt{\url{https://github.com/Helsinki-NLP/lm-vs-mt}}.
Authors: Dingyao Yu, Yang An, Wei Ye, Xiongfeng Xiao, Shaoguang Mao, Tao Ge, Shikun Zhang
Abstract: Chinese Spelling Correction (CSC) commonly lacks large-scale high-quality corpora, due to the labor-intensive labeling of spelling errors in real-life human writing or typing scenarios. Two data augmentation methods are widely adopted: (1) \textit{Random Replacement} with the guidance of confusion sets and (2) \textit{OCR/ASR-based Generation} that simulates character misusing. However, both methods inevitably introduce noisy data (e.g., false spelling errors), potentially leading to over-correction. By carefully analyzing the two types of corpora, we find that though the latter achieves more robust generalization performance, the former yields better-calibrated CSC models. We then provide a theoretical analysis of this empirical observation, based on which a corpus refining strategy is proposed. Specifically, OCR/ASR-based data samples are fed into a well-calibrated CSC model trained on random replacement-based corpora and then filtered based on prediction confidence. By learning a simple BERT-based model on the refined OCR/ASR-based corpus, we set up impressive state-of-the-art performance on three widely-used benchmarks, while significantly alleviating over-correction (e.g., lowering false positive predictions).
Authors: Yifei Gao, Jie Ou, Lei Wang, Fanhua Shang, Jaji Wu, Jun Cheng
Abstract: Large Language Models (LLMs) showcase remarkable performance and robust deductive capabilities, yet their expansive size complicates deployment and raises environmental concerns due to substantial resource consumption. The recent development of a quantization technique known as Learnable Singular-value Increment (LSI) has addressed some of these quantization challenges. Leveraging insights from LSI and our extensive research, we have developed innovative methods that enhance the performance of quantized LLMs, particularly in low-bit settings. Our methods consistently deliver state-of-the-art results across various quantization scenarios and offer deep theoretical insights into the quantization process, elucidating the potential of quantized models for widespread application.
Authors: Chunzhen Jin, Yongfeng Huang, Yaqi Wang, Peng Cao, Osmar Zaiane
Abstract: Text style transfer, an important research direction in natural language processing, aims to adapt the text to various preferences but often faces challenges with limited resources. In this work, we introduce a novel method termed Style Extraction and Tunable Inference via Dual-level Transferable Prompt Learning (SETTP) for effective style transfer in low-resource scenarios. First, SETTP learns source style-level prompts containing fundamental style characteristics from high-resource style transfer. During training, the source style-level prompts are transferred through an attention module to derive a target style-level prompt for beneficial knowledge provision in low-resource style transfer. Additionally, we propose instance-level prompts obtained by clustering the target resources based on the semantic content to reduce semantic bias. We also propose an automated evaluation approach of style similarity based on alignment with human evaluations using ChatGPT-4. Our experiments across three resourceful styles show that SETTP requires only 1/20th of the data volume to achieve performance comparable to state-of-the-art methods. In tasks involving scarce data like writing style and role style, SETTP outperforms previous methods by 16.24\%.
Authors: Yuetong Zhao, Hongyu Cao, Xianyu Zhao, Zhijian Ou
Abstract: Since the launch of ChatGPT at the end of 2022, generative dialogue models represented by ChatGPT have quickly become essential tools in daily life. As user expectations increase, enhancing the capability of generative dialogue models to solve complex problems has become a focal point of current research. This paper delves into the effectiveness of the RAFT (Retrieval Augmented Fine-Tuning) method in improving the performance of Generative dialogue models. RAFT combines chain-of-thought with model supervised fine-tuning (SFT) and retrieval augmented generation (RAG), which significantly enhanced the model's information extraction and logical reasoning abilities. We evaluated the RAFT method across multiple datasets and analysed its performance in various reasoning tasks, including long-form QA and short-form QA tasks, tasks in both Chinese and English, and supportive and comparison reasoning tasks. Notably, it addresses the gaps in previous research regarding long-form QA tasks and Chinese datasets. Moreover, we also evaluate the benefit of the chain-of-thought (CoT) in the RAFT method. This work offers valuable insights for studies focused on enhancing the performance of generative dialogue models.
Authors: Soojin Yoon, Sungho Ko, Tongyoung Kim, SeongKu Kang, Jinyoung Yeo, Dongha Lee
Abstract: Cross-lingual entity alignment (EA) enables the integration of multiple knowledge graphs (KGs) across different languages, providing users with seamless access to diverse and comprehensive knowledge.Existing methods, mostly supervised, face challenges in obtaining labeled entity pairs. To address this, recent studies have shifted towards a self-supervised and unsupervised frameworks. Despite their effectiveness, these approaches have limitations: (1) they mainly focus on entity features, neglecting the semantic information of relations, (2) they assume isomorphism between source and target graphs, leading to noise and reduced alignment accuracy, and (3) they are susceptible to noise in the textual features, especially when encountering inconsistent translations or Out-Of-Vocabulary (OOV) problems. In this paper, we propose ERAlign, an unsupervised and robust cross-lingual EA framework that jointly performs Entity-level and Relation-level Alignment using semantic textual features of relations and entities. Its refinement process iteratively enhances results by fusing entity-level and relation-level alignments based on neighbor triple matching. The additional verification process examines the entities' neighbor triples as the linearized text. This \textit{Align-and-Verify} pipeline that rigorously assesses alignment results, achieving near-perfect alignment even in the presence of noisy textual features of entities. Our extensive experiments demonstrate that robustness and general applicability of \proposed improved the accuracy and effectiveness of EA tasks, contributing significantly to knowledge-oriented applications.
Authors: Nauman Riaz, Saifullah Saifullah, Stefan Agne, Andreas Dengel, Sheraz Ahmed
Abstract: In this study, we introduce StylusAI, a novel architecture leveraging diffusion models in the domain of handwriting style generation. StylusAI is specifically designed to adapt and integrate the stylistic nuances of one language's handwriting into another, particularly focusing on blending English handwriting styles into the context of the German writing system. This approach enables the generation of German text in English handwriting styles and German handwriting styles into English, enriching machine-generated handwriting diversity while ensuring that the generated text remains legible across both languages. To support the development and evaluation of StylusAI, we present the \lq{Deutscher Handschriften-Datensatz}\rq~(DHSD), a comprehensive dataset encompassing 37 distinct handwriting styles within the German language. This dataset provides a fundamental resource for training and benchmarking in the realm of handwritten text generation. Our results demonstrate that StylusAI not only introduces a new method for style adaptation in handwritten text generation but also surpasses existing models in generating handwriting samples that improve both text quality and stylistic fidelity, evidenced by its performance on the IAM database and our newly proposed DHSD. Thus, StylusAI represents a significant advancement in the field of handwriting style generation, offering promising avenues for future research and applications in cross-linguistic style adaptation for languages with similar scripts.
Authors: Danni Yu, Marina Bondi, Ken Hylannd
Abstract: One of the most powerful and enduring ideas in written discourse analysis is that genres can be described in terms of the moves which structure a writer's purpose. Considerable research has sought to identify these distinct communicative acts, but analyses have been beset by problems of subjectivity, reliability and the time-consuming need for multiple coders to confirm analyses. In this paper we employ the affordances of GPT-4 to automate the annotation process by using natural language prompts. Focusing on abstracts from articles in four applied linguistics journals, we devise prompts which enable the model to identify moves effectively. The annotated outputs of these prompts were evaluated by two assessors with a third addressing disagreements. The results show that an 8-shot prompt was more effective than one using two, confirming that the inclusion of examples illustrating areas of variability can enhance GPT-4's ability to recognize multiple moves in a single sentence and reduce bias related to textual position. We suggest that GPT-4 offers considerable potential in automating this annotation process, when human actors with domain specific linguistic expertise inform the prompting process.
Authors: Soroosh Tayebi Arasteh, Mahshad Lotfinia, Keno Bressem, Robert Siepmann, Dyke Ferber, Christiane Kuhl, Jakob Nikolas Kather, Sven Nebelung, Daniel Truhn
Abstract: Large language models (LLMs) have advanced the field of artificial intelligence (AI) in medicine. However LLMs often generate outdated or inaccurate information based on static training datasets. Retrieval augmented generation (RAG) mitigates this by integrating outside data sources. While previous RAG systems used pre-assembled, fixed databases with limited flexibility, we have developed Radiology RAG (RadioRAG) as an end-to-end framework that retrieves data from authoritative radiologic online sources in real-time. RadioRAG is evaluated using a dedicated radiologic question-and-answer dataset (RadioQA). We evaluate the diagnostic accuracy of various LLMs when answering radiology-specific questions with and without access to additional online information via RAG. Using 80 questions from RSNA Case Collection across radiologic subspecialties and 24 additional expert-curated questions, for which the correct gold-standard answers were available, LLMs (GPT-3.5-turbo, GPT-4, Mistral-7B, Mixtral-8x7B, and Llama3 [8B and 70B]) were prompted with and without RadioRAG. RadioRAG retrieved context-specific information from www.radiopaedia.org in real-time and incorporated them into its reply. RadioRAG consistently improved diagnostic accuracy across all LLMs, with relative improvements ranging from 2% to 54%. It matched or exceeded question answering without RAG across radiologic subspecialties, particularly in breast imaging and emergency radiology. However, degree of improvement varied among models; GPT-3.5-turbo and Mixtral-8x7B-instruct-v0.1 saw notable gains, while Mistral-7B-instruct-v0.2 showed no improvement, highlighting variability in its effectiveness. LLMs benefit when provided access to domain-specific data beyond their training data. For radiology, RadioRAG establishes a robust framework that substantially improves diagnostic accuracy and factuality in radiological question answering.
Authors: Joy He-Yueya, Wanjing Anya Ma, Kanishk Gandhi, Benjamin W. Domingue, Emma Brunskill, Noah D. Goodman
Abstract: Language models (LMs) are increasingly used to simulate human-like responses in scenarios where accurately mimicking a population's behavior can guide decision-making, such as in developing educational materials and designing public policies. The objective of these simulations is for LMs to capture the variations in human responses, rather than merely providing the expected correct answers. Prior work has shown that LMs often generate unrealistically accurate responses, but there are no established metrics to quantify how closely the knowledge distribution of LMs aligns with that of humans. To address this, we introduce "psychometric alignment," a metric that measures the extent to which LMs reflect human knowledge distribution. Assessing this alignment involves collecting responses from both LMs and humans to the same set of test items and using Item Response Theory to analyze the differences in item functioning between the groups. We demonstrate that our metric can capture important variations in populations that traditional metrics, like differences in accuracy, fail to capture. We apply this metric to assess existing LMs for their alignment with human knowledge distributions across three real-world domains. We find significant misalignment between LMs and human populations, though using persona-based prompts can improve alignment. Interestingly, smaller LMs tend to achieve greater psychometric alignment than larger LMs. Further, training LMs on human response data from the target distribution enhances their psychometric alignment on unseen test items, but the effectiveness of such training varies across domains.
Authors: Ishan Kavathekar, Anku Rani, Ashmit Chamoli, Ponnurangam Kumaraguru, Amit Sheth, Amitava Das
Abstract: The widespread adoption of large language models (LLMs) and awareness around multilingual LLMs have raised concerns regarding the potential risks and repercussions linked to the misapplication of AI-generated text, necessitating increased vigilance. While these models are primarily trained for English, their extensive training on vast datasets covering almost the entire web, equips them with capabilities to perform well in numerous other languages. AI-Generated Text Detection (AGTD) has emerged as a topic that has already received immediate attention in research, with some initial methods having been proposed, soon followed by the emergence of techniques to bypass detection. In this paper, we report our investigation on AGTD for an indic language Hindi. Our major contributions are in four folds: i) examined 26 LLMs to evaluate their proficiency in generating Hindi text, ii) introducing the AI-generated news article in Hindi ($AG_{hi}$) dataset, iii) evaluated the effectiveness of five recently proposed AGTD techniques: ConDA, J-Guard, RADAR, RAIDAR and Intrinsic Dimension Estimation for detecting AI-generated Hindi text, iv) proposed Hindi AI Detectability Index ($ADI_{hi}$) which shows a spectrum to understand the evolving landscape of eloquence of AI-generated text in Hindi. We will make the codes and datasets available to encourage further research.
Authors: Ori Yoran, Samuel Joseph Amouyal, Chaitanya Malaviya, Ben Bogin, Ofir Press, Jonathan Berant
Abstract: Language agents, built on top of language models (LMs), are systems that can interact with complex environments, such as the open web. In this work, we examine whether such agents can perform realistic and time-consuming tasks on the web, e.g., monitoring real-estate markets or locating relevant nearby businesses. We introduce AssistantBench, a challenging new benchmark consisting of 214 realistic tasks that can be automatically evaluated, covering different scenarios and domains. We find that AssistantBench exposes the limitations of current systems, including language models and retrieval-augmented language models, as no model reaches an accuracy of more than 25 points. While closed-book LMs perform well, they exhibit low precision since they tend to hallucinate facts. State-of-the-art web agents reach a score of near zero. Additionally, we introduce SeePlanAct (SPA), a new web agent that significantly outperforms previous agents, and an ensemble of SPA and closed-book models reaches the best overall performance. Moreover, we analyze failures of current systems and highlight that web navigation remains a major challenge.
Authors: Zhuoyan Xu, Zhenmei Shi, Yingyu Liang
Abstract: Large language models (LLMs) have emerged as powerful tools for many AI problems and exhibit remarkable in-context learning (ICL) capabilities. Compositional ability, solving unseen complex tasks that combine two or more simple tasks, is an essential reasoning ability for Artificial General Intelligence. Despite LLM's tremendous success, how they approach composite tasks, especially those not encountered during the pretraining phase, remains an open question and largely ununderstood. In this study, we delve into the ICL capabilities of LLMs on composite tasks, with only simple tasks as in-context examples. We develop a test suite of composite tasks that include linguistic and logical challenges and perform empirical studies across different LLM families. We observe that models exhibit divergent behaviors: (1) For simpler composite tasks that apply distinct mapping mechanisms to different input segments, the models demonstrate decent compositional ability, while scaling up the model enhances this ability; (2) for more complex composite tasks that involving reasoning multiple steps, where each step represent one task, models typically underperform, and scaling up generally provide no improvements. We offer theoretical analysis in a simplified setting, explaining that models exhibit compositional capability when the task handles different input parts separately. We believe our work sheds new light on the capabilities of LLMs in solving composite tasks regarding the nature of the tasks and model scale. Our dataset and code are available at {\url{https://github.com/OliverXUZY/LLM_Compose}}.
Authors: Zhi Hao Luo, Luis Lara, Ge Ya Luo, Florian Golemo, Christopher Beckham, Christopher Pal
Abstract: Text conditioned generative models for images have yielded impressive results. Text conditioned floorplan generation as a special type of raster image generation task also received particular attention. However there are many use cases in floorpla generation where numerical properties of the generated result are more important than the aesthetics. For instance, one might want to specify sizes for certain rooms in a floorplan and compare the generated floorplan with given specifications Current approaches, datasets and commonly used evaluations do not support these kinds of constraints. As such, an attractive strategy is to generate an intermediate data structure that contains numerical properties of a floorplan which can be used to generate the final floorplan image. To explore this setting we (1) construct a new dataset for this data-structure to data-structure formulation of floorplan generation using two popular image based floorplan datasets RPLAN and ProcTHOR-10k, and provide the tools to convert further procedurally generated ProcTHOR floorplan data into our format. (2) We explore the task of floorplan generation given a partial or complete set of constraints and we design a series of metrics and benchmarks to enable evaluating how well samples generated from models respect the constraints. (3) We create multiple baselines by finetuning a large language model (LLM), Llama3, and demonstrate the feasibility of using floorplan data structure conditioned LLMs for the problem of floorplan generation respecting numerical constraints. We hope that our new datasets and benchmarks will encourage further research on different ways to improve the performance of LLMs and other generative modelling techniques for generating designs where quantitative constraints are only partially specified, but must be respected.
Authors: Steffen Kleinle, Jakob Prange, Annemarie Friedrich
Abstract: When immigrating to a new country, it is easy to feel overwhelmed by the need to obtain information on financial support, housing, schooling, language courses, and other issues. If relocation is rushed or even forced, the necessity for high-quality answers to such questions is all the more urgent. Official immigration counselors are usually overbooked, and online systems could guide newcomers to the requested information or a suitable counseling service. To this end, we present OMoS-QA, a dataset of German and English questions paired with relevant trustworthy documents and manually annotated answers, specifically tailored to this scenario. Questions are automatically generated with an open-source large language model (LLM) and answer sentences are selected by crowd workers with high agreement. With our data, we conduct a comparison of 5 pretrained LLMs on the task of extractive question answering (QA) in German and English. Across all models and both languages, we find high precision and low-to-mid recall in selecting answer sentences, which is a favorable trade-off to avoid misleading users. This performance even holds up when the question language does not match the document language. When it comes to identifying unanswerable questions given a context, there are larger differences between the two languages.
Authors: Rian Dolphin, Joe Dursun, Jonathan Chow, Jarrett Blankenship, Katie Adams, Quinton Pike
Abstract: Financial news plays a crucial role in decision-making processes across the financial sector, yet the efficient processing of this information into a structured format remains challenging. This paper presents a novel approach to financial news processing that leverages Large Language Models (LLMs) to overcome limitations that previously prevented the extraction of structured data from unstructured financial news. We introduce a system that extracts relevant company tickers from raw news article content, performs sentiment analysis at the company level, and generates summaries, all without relying on pre-structured data feeds. Our methodology combines the generative capabilities of LLMs, and recent prompting techniques, with a robust validation framework that uses a tailored string similarity approach. Evaluation on a dataset of 5530 financial news articles demonstrates the effectiveness of our approach, with 90% of articles not missing any tickers compared with current data providers, and 22% of articles having additional relevant tickers. In addition to this paper, the methodology has been implemented at scale with the resulting processed data made available through a live API endpoint, which is updated in real-time with the latest news. To the best of our knowledge, we are the first data provider to offer granular, per-company sentiment analysis from news articles, enhancing the depth of information available to market participants. We also release the evaluation dataset of 5530 processed articles as a static file, which we hope will facilitate further research leveraging financial news.
Authors: Catarina G Belem, Markelle Kelly, Mark Steyvers, Sameer Singh, Padhraic Smyth
Abstract: Uncertainty expressions such as ``probably'' or ``highly unlikely'' are pervasive in human language. While prior work has established that there is population-level agreement in terms of how humans interpret these expressions, there has been little inquiry into the abilities of language models to interpret such expressions. In this paper, we investigate how language models map linguistic expressions of uncertainty to numerical responses. Our approach assesses whether language models can employ theory of mind in this setting: understanding the uncertainty of another agent about a particular statement, independently of the model's own certainty about that statement. We evaluate both humans and 10 popular language models on a task created to assess these abilities. Unexpectedly, we find that 8 out of 10 models are able to map uncertainty expressions to probabilistic responses in a human-like manner. However, we observe systematically different behavior depending on whether a statement is actually true or false. This sensitivity indicates that language models are substantially more susceptible to bias based on their prior knowledge (as compared to humans). These findings raise important questions and have broad implications for human-AI alignment and AI-AI communication.
Authors: Wataru Nakata, Kentaro Seki, Hitomi Yanaka, Yuki Saito, Shinnosuke Takamichi, Hiroshi Saruwatari
Abstract: Spoken dialogue plays a crucial role in human-AI interactions, necessitating dialogue-oriented spoken language models (SLMs). To develop versatile SLMs, large-scale and diverse speech datasets are essential. Additionally, to ensure hiqh-quality speech generation, the data must be spontaneous like in-wild data and must be acoustically clean with noise removed. Despite the critical need, no open-source corpus meeting all these criteria has been available. This study addresses this gap by constructing and releasing a large-scale spoken dialogue corpus, named Japanese Corpus for Human-AI Talks (J-CHAT), which is publicly accessible. Furthermore, this paper presents a language-independent method for corpus construction and describes experiments on dialogue generation using SLMs trained on J-CHAT. Experimental results indicate that the collected data from multiple domains by our method improve the naturalness and meaningfulness of dialogue generation.
Authors: He Bai, Tatiana Likhomanenko, Ruixiang Zhang, Zijin Gu, Zakaria Aldeneh, Navdeep Jaitly
Abstract: Large language models have revolutionized natural language processing by leveraging self-supervised pretraining on vast textual data. Inspired by this success, researchers have investigated complicated speech tokenization methods to discretize continuous speech signals so that language modeling techniques can be applied to speech data. However, existing approaches either model semantic tokens, potentially losing acoustic information, or model acoustic tokens, risking the loss of semantic information. Having multiple token types also complicates the architecture and requires additional pretraining. Here we show that discretizing mel-filterbank channels into discrete intensity bins produces a simple representation (dMel), that performs better than other existing speech tokenization methods. Using a transformer decoder-only architecture for speech-text modeling, we comprehensively evaluate different speech tokenization methods on speech recognition (ASR), speech synthesis (TTS). Our results demonstrate the effectiveness of dMel in achieving high performance on both tasks within a unified framework, paving the way for efficient and effective joint modeling of speech and text.
Authors: Mahdi Buali, Robert Hoehndorf
Abstract: Automated Theorem Proving (ATP) faces challenges due to its complexity and computational demands. Recent work has explored using Large Language Models (LLMs) for ATP action selection, but these methods can be resource-intensive. This study introduces FEAS, an agent that enhances the COPRA in-context learning framework within Lean. FEAS refines prompt generation, response parsing, and incorporates domain-specific heuristics for functional equations. It introduces FunEq, a curated dataset of functional equation problems with varying difficulty. FEAS outperforms baselines on FunEq, particularly with the integration of domain-specific heuristics. The results demonstrate FEAS's effectiveness in generating and formalizing high-level proof strategies into Lean proofs, showcasing the potential of tailored approaches for specific ATP challenges.
Authors: Ritabrata Roy Choudhury
Abstract: The proposed model aims to develop a speech recognition technology for hearing, speech, or cognitively disabled people. All the available technology in the field of speech recognition doesn't come with an interface for communication for people with hearing, speech, or cognitive disabilities. The proposed model proposes the speech from the user, is transmitted to the speech recognition layer where it is converted into text and then that text is then transmitted to the morse code conversion layer where the morse code of the corresponding speech is given as the output. The accuracy of the model is completely dependent on speech recognition, as the morse code conversion is a process. The model is tested with recorded audio files with different parameters. The proposed model's WER and accuracy are both determined to be 10.18% and 89.82%, respectively.
Authors: Ralph P. Lano
Abstract: This study presents a theoretical model for a self-replicating mechanical system inspired by biological processes within living cells and supported by computer simulations. The model decomposes self-replication into core components, each of which is executed by a single machine constructed from a set of basic block types. Key functionalities such as sorting, copying, and building, are demonstrated. The model provides valuable insights into the constraints of self-replicating systems. The discussion also addresses the spatial and timing behavior of the system, as well as its efficiency and complexity. This work provides a foundational framework for future studies on self-replicating mechanisms and their information-processing applications.
Authors: Xiaoyu Tan (INF Technology), Yongxin Deng (Shanghai University of Engineering Science), Xihe Qiu (Shanghai University of Engineering Science), Weidi Xu (INF Technology), Chao Qu (INF Technology), Wei Chu (INF Technology), Yinghui Xu (Fudan University), Yuan Qi (Fudan University)
Abstract: Large language models (LLMs) have shown exceptional performance as general-purpose assistants, excelling across a variety of reasoning tasks. This achievement represents a significant step toward achieving artificial general intelligence (AGI). Despite these advancements, the effectiveness of LLMs often hinges on the specific prompting strategies employed, and there remains a lack of a robust framework to facilitate learning and generalization across diverse reasoning tasks. To address these challenges, we introduce a novel learning framework, THOUGHT-LIKE-PRO In this framework, we utilize imitation learning to imitate the Chain-of-Thought (CoT) process which is verified and translated from reasoning trajectories generated by a symbolic Prolog logic engine. This framework proceeds in a self-driven manner, that enables LLMs to formulate rules and statements from given instructions and leverage the symbolic Prolog engine to derive results. Subsequently, LLMs convert Prolog-derived successive reasoning trajectories into natural language CoT for imitation learning. Our empirical findings indicate that our proposed approach substantially enhances the reasoning abilities of LLMs and demonstrates robust generalization across out-of-distribution reasoning tasks.
Authors: Andr\'e F. Cruz, Moritz Hardt, Celestine Mendler-D\"unner
Abstract: Current question-answering benchmarks predominantly focus on accuracy in realizable prediction tasks. Conditioned on a question and answer-key, does the most likely token match the ground truth? Such benchmarks necessarily fail to evaluate language models' ability to quantify outcome uncertainty. In this work, we focus on the use of language models as risk scores for unrealizable prediction tasks. We introduce folktexts, a software package to systematically generate risk scores using large language models, and evaluate them against benchmark prediction tasks. Specifically, the package derives natural language tasks from US Census data products, inspired by popular tabular data benchmarks. A flexible API allows for any task to be constructed out of 28 census features whose values are mapped to prompt-completion pairs. We demonstrate the utility of folktexts through a sweep of empirical insights on 16 recent large language models, inspecting risk scores, calibration curves, and diverse evaluation metrics. We find that zero-shot risk sores have high predictive signal while being widely miscalibrated: base models overestimate outcome uncertainty, while instruction-tuned models underestimate uncertainty and generate over-confident risk scores.
Authors: Pier Giuseppe Sessa, Robert Dadashi, L\'eonard Hussenot, Johan Ferret, Nino Vieillard, Alexandre Ram\'e, Bobak Shariari, Sarah Perrin, Abe Friesen, Geoffrey Cideron, Sertan Girgin, Piotr Stanczyk, Andrea Michi, Danila Sinopalnikov, Sabela Ramos, Am\'elie H\'eliou, Aliaksei Severyn, Matt Hoffman, Nikola Momchev, Olivier Bachem
Abstract: Reinforcement learning from human feedback (RLHF) is a key driver of quality and safety in state-of-the-art large language models. Yet, a surprisingly simple and strong inference-time strategy is Best-of-N sampling that selects the best generation among N candidates. In this paper, we propose Best-of-N Distillation (BOND), a novel RLHF algorithm that seeks to emulate Best-of-N but without its significant computational overhead at inference time. Specifically, BOND is a distribution matching algorithm that forces the distribution of generations from the policy to get closer to the Best-of-N distribution. We use the Jeffreys divergence (a linear combination of forward and backward KL) to balance between mode-covering and mode-seeking behavior, and derive an iterative formulation that utilizes a moving anchor for efficiency. We demonstrate the effectiveness of our approach and several design choices through experiments on abstractive summarization and Gemma models. Aligning Gemma policies with BOND outperforms other RLHF algorithms by improving results on several benchmarks.
Authors: Yunseon Choi, Sangmin Bae, Seonghyun Ban, Minchan Jeong, Chuheng Zhang, Lei Song, Li Zhao, Jiang Bian, Kee-Eung Kim
Abstract: With the advent of foundation models, prompt tuning has positioned itself as an important technique for directing model behaviors and eliciting desired responses. Prompt tuning regards selecting appropriate keywords included into the input, thereby adapting to the downstream task without adjusting or fine-tuning the model parameters. There is a wide range of work in prompt tuning, from approaches that directly harness the backpropagated gradient signals from the model, to those employing black-box optimization such as reinforcement learning (RL) methods. Our primary focus is on RLPrompt, which aims to find optimal prompt tokens leveraging soft Q-learning. While the results show promise, we have observed that the prompts frequently appear unnatural, which impedes their interpretability. We address this limitation by using sparse Tsallis entropy regularization, a principled approach to filtering out unlikely tokens from consideration. We extensively evaluate our approach across various tasks, including few-shot text classification, unsupervised text style transfer, and textual inversion from images. The results indicate a notable improvement over baselines, highlighting the efficacy of our approach in addressing the challenges of prompt tuning. Moreover, we show that the prompts discovered using our method are more natural and interpretable compared to those from other baselines.
Authors: Yanxi Chen, Yaliang Li, Bolin Ding, Jingren Zhou
Abstract: We initiate a formal investigation into the design and analysis of LLM-based algorithms, i.e. algorithms that contain one or multiple calls of large language models (LLMs) as sub-routines and critically rely on the capabilities of LLMs. While LLM-based algorithms, ranging from basic LLM calls with prompt engineering to complicated LLM-powered agent systems and compound AI systems, have achieved remarkable empirical success, the design and optimization of them have mostly relied on heuristics and trial-and-errors, which is largely due to a lack of formal and analytical study for these algorithms. To fill this gap, we start by identifying the computational-graph representation of LLM-based algorithms, the design principle of task decomposition, and some key abstractions, which then facilitate our formal analysis for the accuracy and efficiency of LLM-based algorithms, despite the black-box nature of LLMs. We further consider parallel decomposition for a case study, providing extensive analytical and empirical study for four concrete examples of this pattern. Our proposed framework holds promise for advancing LLM-based algorithms, by revealing the reasons behind curious empirical phenomena, guiding the choices of hyperparameters, predicting the empirical performance of algorithms, and inspiring new algorithm design. To promote further study of LLM-based algorithms, we release our source code at https://github.com/modelscope/agentscope/tree/main/examples/paper_llm_based_algorithm.
URLs: https://github.com/modelscope/agentscope/tree/main/examples/paper_llm_based_algorithm.
Authors: Ze Yu Zhang, Arun Verma, Finale Doshi-Velez, Bryan Kian Hsiang Low
Abstract: Large language models (LLMs) are widely used in decision-making, but their reliability, especially in critical tasks like healthcare, is not well-established. Therefore, understanding how LLMs reason and make decisions is crucial for their safe deployment. This paper investigates how the uncertainty of responses generated by LLMs relates to the information provided in the input prompt. Leveraging the insight that LLMs learn to infer latent concepts during pretraining, we propose a prompt-response concept model that explains how LLMs generate responses and helps understand the relationship between prompts and response uncertainty. We show that the uncertainty decreases as the prompt's informativeness increases, similar to epistemic uncertainty. Our detailed experimental results on real datasets validate our proposed model.
Authors: Chen Shen, Chunfeng Lian, Wanqing Zhang, Fan Wang, Jianhua Zhang, Shuanliang Fan, Xin Wei, Gongji Wang, Kehan Li, Hongshu Mu, Hao Wu, Xinggong Liang, Jianhua Ma, Zhenyuan Wang
Abstract: Forensic pathology is critical in determining the cause and manner of death through post-mortem examinations, both macroscopic and microscopic. The field, however, grapples with issues such as outcome variability, laborious processes, and a scarcity of trained professionals. This paper presents SongCi, an innovative visual-language model (VLM) designed specifically for forensic pathology. SongCi utilizes advanced prototypical cross-modal self-supervised contrastive learning to enhance the accuracy, efficiency, and generalizability of forensic analyses. It was pre-trained and evaluated on a comprehensive multi-center dataset, which includes over 16 million high-resolution image patches, 2,228 vision-language pairs of post-mortem whole slide images (WSIs), and corresponding gross key findings, along with 471 distinct diagnostic outcomes. Our findings indicate that SongCi surpasses existing multi-modal AI models in many forensic pathology tasks, performs comparably to experienced forensic pathologists and significantly better than less experienced ones, and provides detailed multi-modal explainability, offering critical assistance in forensic investigations. To the best of our knowledge, SongCi is the first VLM specifically developed for forensic pathological analysis and the first large-vocabulary computational pathology (CPath) model that directly processes gigapixel WSIs in forensic science.
Authors: Md Zarif Hossain, Ahmed Imteaj
Abstract: Vision-language models (VLMs) have achieved significant strides in recent times specially in multimodal tasks, yet they remain susceptible to adversarial attacks on their vision components. To address this, we propose Sim-CLIP, an unsupervised adversarial fine-tuning method that enhances the robustness of the widely-used CLIP vision encoder against such attacks while maintaining semantic richness and specificity. By employing a Siamese architecture with cosine similarity loss, Sim-CLIP learns semantically meaningful and attack-resilient visual representations without requiring large batch sizes or momentum encoders. Our results demonstrate that VLMs enhanced with Sim-CLIP's fine-tuned CLIP encoder exhibit significantly enhanced robustness against adversarial attacks, while preserving semantic meaning of the perturbed images. Notably, Sim-CLIP does not require additional training or fine-tuning of the VLM itself; replacing the original vision encoder with our fine-tuned Sim-CLIP suffices to provide robustness. This work underscores the significance of reinforcing foundational models like CLIP to safeguard the reliability of downstream VLM applications, paving the way for more secure and effective multimodal systems.
Authors: Biswadip Mandal, Xiangci Li, Jessica Ouyang
Abstract: A key challenge in citation text generation is that the length of generated text often differs from the length of the target, lowering the quality of the generation. While prior works have investigated length-controlled generation, their effectiveness depends on knowing the appropriate generation length. In this work, we present an in-depth study of the limitations of predicting scientific citation text length and explore the use of heuristic estimates of desired length.
Authors: Alessio Buscemi, Daniele Proverbio
Abstract: The ethical implications and potentials for misuse of Generative Artificial Intelligence are increasingly worrying topics. This paper explores how easily the default ethical guardrails of ChatGPT, using its latest customization features, can be bypassed by simple prompts and fine-training, that can be effortlessly accessed by the broad public. This malevolently altered version of ChatGPT, nicknamed "RogueGPT", responded with worrying behaviours, beyond those triggered by jailbreak prompts. We conduct an empirical study of RogueGPT responses, assessing its flexibility in answering questions pertaining to what should be disallowed usage. Our findings raise significant concerns about the model's knowledge about topics like illegal drug production, torture methods and terrorism. The ease of driving ChatGPT astray, coupled with its global accessibility, highlights severe issues regarding the data quality used for training the foundational model and the implementation of ethical safeguards. We thus underline the responsibilities and dangers of user-driven modifications, and the broader effects that these may have on the design of safeguarding and ethical modules implemented by AI programmers.
Authors: Shivprasad Sagare, Hemachandran S, Kinshuk Sarabhai, Prashant Ullegaddi, Rajeshkumar SA
Abstract: Recent advances in multimodal LLMs, have led to several video-text models being proposed for critical video-related tasks. However, most of the previous works support visual input only, essentially muting the audio signal in the video. Few models that support both audio and visual input, are not explicitly trained on audio data. Hence, the effect of audio towards video understanding is largely unexplored. To this end, we propose a model architecture that handles audio-visual inputs explicitly. We train our model with both audio and visual data from a video instruction-tuning dataset. Comparison with vision-only baselines, and other audio-visual models showcase that training on audio data indeed leads to improved grounding of responses. For better evaluation of audio-visual models, we also release a human-annotated benchmark dataset, with audio-aware question-answer pairs.
Authors: Jianxin Liang, Xiaojun Meng, Yueqian Wang, Chang Liu, Qun Liu, Dongyan Zhao
Abstract: Video Question Answering (VideoQA) has emerged as a challenging frontier in the field of multimedia processing, requiring intricate interactions between visual and textual modalities. Simply uniformly sampling frames or indiscriminately aggregating frame-level visual features often falls short in capturing the nuanced and relevant contexts of videos to well perform VideoQA. To mitigate these issues, we propose VidF4, a novel VideoQA framework equipped with tailored frame selection strategy for effective and efficient VideoQA. We propose three frame-scoring mechanisms that consider both question relevance and inter-frame similarity to evaluate the importance of each frame for a given question on the video. Furthermore, we design a differentiable adaptive frame sampling mechanism to facilitate end-to-end training for the frame selector and answer generator. The experimental results across three widely adopted benchmarks demonstrate that our model consistently outperforms existing VideoQA methods, establishing a new SOTA across NExT-QA (+0.3%), STAR (+0.9%), and TVQA (+1.0%). Furthermore, through both quantitative and qualitative analyses, we validate the effectiveness of each design choice.
Authors: Zongyue Qin, Chen Luo, Zhengyang Wang, Haoming Jiang, Yizhou Sun
Abstract: Large language models (LLMs) excel in many natural language processing (NLP) tasks. However, since LLMs can only incorporate new knowledge through training or supervised fine-tuning processes, they are unsuitable for applications that demand precise, up-to-date, and private information not available in the training corpora. This precise, up-to-date, and private information is typically stored in relational databases. Thus, a promising solution is to augment LLMs with the inclusion of relational databases as external memory. This can ensure the timeliness, correctness, and consistency of data, and assist LLMs in performing complex arithmetic operations beyond their inherent capabilities. However, bridging the gap between LLMs and relational databases is challenging. It requires the awareness of databases and data values stored in databases to select correct databases and issue correct SQL queries. Besides, it is necessary for the external memory to be independent of the LLM to meet the needs of real-world applications. We introduce a novel LLM-agnostic memory architecture comprising a database selection memory, a data value memory, and relational databases. And we design an elegant pipeline to retrieve information from it. Besides, we carefully design the prompts to instruct the LLM to maximize the framework's potential. To evaluate our method, we compose a new dataset with various types of questions. Experimental results show that our framework enables LLMs to effectively answer database-related questions, which is beyond their direct ability.
Authors: Hao Duong Le, Xin Xia, Zhang Chen
Abstract: Large Language Models (LLMs) have demonstrated significant potential in causal discovery tasks by utilizing their vast expert knowledge from extensive text corpora. However, the multi-agent capabilities of LLMs in causal discovery remain underexplored. This paper introduces a general framework to investigate this potential. The first is the Meta Agents Model, which relies exclusively on reasoning and discussions among LLM agents to conduct causal discovery. The second is the Coding Agents Model, which leverages the agents' ability to plan, write, and execute code, utilizing advanced statistical libraries for causal discovery. The third is the Hybrid Model, which integrates both the Meta Agents Model and CodingAgents Model approaches, combining the statistical analysis and reasoning skills of multiple agents. Our proposed framework shows promising results by effectively utilizing LLMs expert knowledge, reasoning capabilities, multi-agent cooperation, and statistical causal methods. By exploring the multi-agent potential of LLMs, we aim to establish a foundation for further research in utilizing LLMs multi-agent for solving causal-related problems.
Authors: Karina Vida, Fabian Damken, Anne Lauscher
Abstract: Large language models (LLMs) increasingly find their way into the most diverse areas of our everyday lives. They indirectly influence people's decisions or opinions through their daily use. Therefore, understanding how and which moral judgements these LLMs make is crucial. However, morality is not universal and depends on the cultural background. This raises the question of whether these cultural preferences are also reflected in LLMs when prompted in different languages or whether moral decision-making is consistent across different languages. So far, most research has focused on investigating the inherent values of LLMs in English. While a few works conduct multilingual analyses of moral bias in LLMs in a multilingual setting, these analyses do not go beyond atomic actions. To the best of our knowledge, a multilingual analysis of moral bias in dilemmas has not yet been conducted. To address this, our paper builds on the moral machine experiment (MME) to investigate the moral preferences of five LLMs, Falcon, Gemini, Llama, GPT, and MPT, in a multilingual setting and compares them with the preferences collected from humans belonging to different cultures. To accomplish this, we generate 6500 scenarios of the MME and prompt the models in ten languages on which action to take. Our analysis reveals that all LLMs inhibit different moral biases to some degree and that they not only differ from the human preferences but also across multiple languages within the models themselves. Moreover, we find that almost all models, particularly Llama 3, divert greatly from human values and, for instance, prefer saving fewer people over saving more.
Authors: Kwanyong Park, Kuniaki Saito, Donghyun Kim
Abstract: Vision-language (VL) models often exhibit a limited understanding of complex expressions of visual objects (e.g., attributes, shapes, and their relations), given complex and diverse language queries. Traditional approaches attempt to improve VL models using hard negative synthetic text, but their effectiveness is limited. In this paper, we harness the exceptional compositional understanding capabilities of generative foundational models. We introduce a novel method for structured synthetic data generation aimed at enhancing the compositional understanding of VL models in language-based object detection. Our framework generates densely paired positive and negative triplets (image, text descriptions, and bounding boxes) in both image and text domains. By leveraging these synthetic triplets, we transform 'weaker' VL models into 'stronger' models in terms of compositional understanding, a process we call "Weak-to-Strong Compositional Learning" (WSCL). To achieve this, we propose a new compositional contrastive learning formulation that discovers semantics and structures in complex descriptions from synthetic triplets. As a result, VL models trained with our synthetic data generation exhibit a significant performance boost in the Omnilabel benchmark by up to +5AP and the D3 benchmark by +6.9AP upon existing baselines.
Authors: Melissa Dell
Abstract: Deep learning provides powerful methods to impute structured information from large-scale, unstructured text and image datasets. For example, economists might wish to detect the presence of economic activity in satellite images, or to measure the topics or entities mentioned in social media, the congressional record, or firm filings. This review introduces deep neural networks, covering methods such as classifiers, regression models, generative AI, and embedding models. Applications include classification, document digitization, record linkage, and methods for data exploration in massive scale text and image corpora. When suitable methods are used, deep learning models can be cheap to tune and can scale affordably to problems involving millions or billions of data points.. The review is accompanied by a companion website, EconDL, with user-friendly demo notebooks, software resources, and a knowledge base that provides technical details and additional applications.
Authors: Wenbin An, Feng Tian, Jiahao Nie, Wenkai Shi, Haonan Lin, Yan Chen, QianYing Wang, Yaqiang Wu, Guang Dai, Ping Chen
Abstract: Knowledge-based Visual Question Answering (KVQA) requires both image and world knowledge to answer questions. Current methods first retrieve knowledge from the image and external knowledge base with the original complex question, then generate answers with Large Language Models (LLMs). However, since the original question contains complex elements that require knowledge from different sources, acquiring different kinds of knowledge in a coupled manner may confuse models and hinder them from retrieving precise knowledge. Furthermore, the ``forward-only'' answering process fails to explicitly capture the knowledge needs of LLMs, which can further hurt answering quality. To cope with the above limitations, we propose DKA: Disentangled Knowledge Acquisition from LLM feedback, a training-free framework that disentangles knowledge acquisition to avoid confusion and uses LLM's feedback to specify the required knowledge. Specifically, DKA requires LLMs to specify what knowledge they need to answer the question and decompose the original complex question into two simple sub-questions: Image-based sub-question and Knowledge-based sub-question. Then we use the two sub-questions to retrieve knowledge from the image and knowledge base, respectively. In this way, two knowledge acquisition models can focus on the content that corresponds to them and avoid disturbance of irrelevant elements in the original complex question, which can help to provide more precise knowledge and better align the knowledge needs of LLMs to yield correct answers. Experiments on benchmark datasets show that DKA significantly outperforms SOTA models. To facilitate future research, our data and code are available at \url{https://github.com/Lackel/DKA}.
Authors: Jiaxing Zhang, Jiayi Liu, Dongsheng Luo, Jennifer Neville, Hua Wei
Abstract: Recent studies seek to provide Graph Neural Network (GNN) interpretability via multiple unsupervised learning models. Due to the scarcity of datasets, current methods easily suffer from learning bias. To solve this problem, we embed a Large Language Model (LLM) as knowledge into the GNN explanation network to avoid the learning bias problem. We inject LLM as a Bayesian Inference (BI) module to mitigate learning bias. The efficacy of the BI module has been proven both theoretically and experimentally. We conduct experiments on both synthetic and real-world datasets. The innovation of our work lies in two parts: 1. We provide a novel view of the possibility of an LLM functioning as a Bayesian inference to improve the performance of existing algorithms; 2. We are the first to discuss the learning bias issues in the GNN explanation problem.
Authors: Simon Gonzalez
Abstract: The field of conlang has evidenced an important growth in the last decades. This has been the product of a wide interest in the use and study of conlangs for artistic purposes. However, one important question is what it is happening with conlang in the academic world. This paper aims to have an overall understanding of the literature on conlang research. With this we aim to give a realistic picture of the field in present days. We have implemented a computational linguistic approach, combining bibliometrics and network analysis to examine all publications available in the Scopus database. Analysing over 2300 academic publications since 1927 until 2022, we have found that Esperanto is by far the most documented conlang. Three main authors have contributed to this: Garv\'ia R., Fiedler S., and Blanke D. The 1970s and 1980s have been the decades where the foundations of current research have been built. In terms of methodologies, language learning and experimental linguistics are the ones contributing to most to the preferred approaches of study in the field. We present the results and discuss our limitations and future work.
Authors: Aki H\"arm\"a, Marcin Pietrasik, Anna Wilbik
Abstract: Large pretrained self-attention neural networks, or transformers, have been very successful in various tasks recently. The performance of a model on a given task depends on its ability to memorize and generalize the training data. Large transformer models, which may have billions of parameters, in theory have a huge capacity to memorize content. However, the current algorithms for the optimization fall short of the theoretical capacity, and the capacity is also highly dependent on the content. In this paper, we focus on the memory capacity of these models obtained using common training algorithms and synthetic training data. Based on the results, we derive an empirical capacity model (ECM) for a generic transformer. The ECM can be used to design task-specific transformer models with an optimal number of parameters in cases where the target memorization capability of the task can be defined.
Authors: Adway Girish, Alliot Nagle, Marco Bondaschi, Michael Gastpar, Ashok Vardhan Makkuva, Hyeji Kim
Abstract: We formalize the problem of prompt compression for large language models (LLMs) and present a framework to unify token-level prompt compression methods which create hard prompts for black-box models. We derive the distortion-rate function for this setup as a linear program, and provide an efficient algorithm to compute this fundamental limit via the dual of the linear program. Using the distortion-rate function as the baseline, we study the performance of existing compression schemes on a synthetic dataset consisting of prompts generated from a Markov chain, natural language queries, and their respective answers. Our empirical analysis demonstrates the criticality of query-aware prompt compression, where the compressor has knowledge of the downstream task/query for the black-box LLM. We show that there is a large gap between the performance of current prompt compression methods and the optimal strategy, and propose a query-aware, variable-rate adaptation of a prior work to close the gap. We extend our experiments to a small natural language dataset to further confirm our findings on our synthetic dataset.
Authors: Georgy Tyukin, Gbetondji J-S Dovonon, Jean Kaddour, Pasquale Minervini
Abstract: The inference demand for LLMs has skyrocketed in recent months, and serving models with low latencies remains challenging due to the quadratic input length complexity of the attention layers. In this work, we investigate the effect of dropping MLP and attention layers at inference time on the performance of Llama-v2 models. We find that dropping dreeper attention layers only marginally decreases performance but leads to the best speedups alongside dropping entire layers. For example, removing 33\% of attention layers in a 13B Llama2 model results in a 1.8\% drop in average performance over the OpenLLM benchmark. We also observe that skipping layers except the latter layers reduces performances for more layers skipped, except for skipping the attention layers.
Authors: Abhay Sheshadri, Aidan Ewart, Phillip Guo, Aengus Lynch, Cindy Wu, Vivek Hebbar, Henry Sleight, Asa Cooper Stickland, Ethan Perez, Dylan Hadfield-Menell, Stephen Casper
Abstract: Large language models (LLMs) can often be made to behave in undesirable ways that they are explicitly fine-tuned not to. For example, the LLM red-teaming literature has produced a wide variety of `jailbreaking' techniques to elicit harmful text from models that were fine-tuned to be harmless. Recent work on red-teaming, model editing, and interpretability suggests that this challenge stems from how (adversarial) fine-tuning largely serves to suppress rather than remove undesirable capabilities from LLMs. Prior work has introduced latent adversarial training (LAT) as a way to improve robustness to broad classes of failures. These prior works have considered untargeted latent space attacks where the adversary perturbs latent activations to maximize loss on examples of desirable behavior. Untargeted LAT can provide a generic type of robustness but does not leverage information about specific failure modes. Here, we experiment with targeted LAT where the adversary seeks to minimize loss on a specific competing task. We find that it can augment a wide variety of state-of-the-art methods. First, we use targeted LAT to improve robustness to jailbreaks, outperforming a strong R2D2 baseline with orders of magnitude less compute. Second, we use it to more effectively remove backdoors with no knowledge of the trigger. Finally, we use it to more effectively unlearn knowledge for specific undesirable tasks in a way that is also more robust to re-learning. Overall, our results suggest that targeted LAT can be an effective tool for defending against harmful behaviors from LLMs.
Authors: Jesse Roberts, Lindsey Roberts, Alice Reed
Abstract: The internet offers tremendous access to services, social connections, and needed products. However, to those without sufficient experience, engaging with businesses and friends across the internet can be daunting due to the ever present danger of scammers and thieves, to say nothing of the myriad of potential computer viruses. Like a forest rich with both edible and poisonous plants, those familiar with the norms inhabit it safely with ease while newcomers need a guide. However, reliance on a human digital guide can be taxing and often impractical. We propose and pilot a simple but unexplored idea: could an LLM provide the necessary support to help the elderly who are separated by the digital divide safely achieve digital autonomy?
Authors: Haoning Wu, Dongxu Li, Bei Chen, Junnan Li
Abstract: Large multimodal models (LMMs) are processing increasingly longer and richer inputs. Albeit the progress, few public benchmark is available to measure such development. To mitigate this gap, we introduce LongVideoBench, a question-answering benchmark that features video-language interleaved inputs up to an hour long. Our benchmark includes 3,763 varying-length web-collected videos with their subtitles across diverse themes, designed to comprehensively evaluate LMMs on long-term multimodal understanding. To achieve this, we interpret the primary challenge as to accurately retrieve and reason over detailed multimodal information from long inputs. As such, we formulate a novel video question-answering task termed referring reasoning. Specifically, as part of the question, it contains a referring query that references related video contexts, called referred context. The model is then required to reason over relevant video details from the referred context. Following the paradigm of referring reasoning, we curate 6,678 human-annotated multiple-choice questions in 17 fine-grained categories, establishing one of the most comprehensive benchmarks for long-form video understanding. Evaluations suggest that the LongVideoBench presents significant challenges even for the most advanced proprietary models (e.g. GPT-4o, Gemini-1.5-Pro, GPT-4-Turbo), while their open-source counterparts show an even larger performance gap. In addition, our results indicate that model performance on the benchmark improves only when they are capable of processing more frames, positioning LongVideoBench as a valuable benchmark for evaluating future-generation long-context LMMs.
Authors: Kaiwen Wang, Rahul Kidambi, Ryan Sullivan, Alekh Agarwal, Christoph Dann, Andrea Michi, Marco Gelmi, Yunxuan Li, Raghav Gupta, Avinava Dubey, Alexandre Ram\'e, Johan Ferret, Geoffrey Cideron, Le Hou, Hongkun Yu, Amr Ahmed, Aranyak Mehta, L\'eonard Hussenot, Olivier Bachem, Edouard Leurent
Abstract: Reward-based finetuning is crucial for aligning language policies with intended behaviors (e.g., creativity and safety). A key challenge here is to develop steerable language models that trade-off multiple (conflicting) objectives in a flexible and efficient manner. This paper presents Conditioned Language Policy (CLP), a general framework for finetuning language models on multiple objectives. Building on techniques from multi-task training and parameter-efficient finetuning, CLP can learn steerable models that effectively trade-off conflicting objectives at inference time. Notably, this does not require training or maintaining multiple models to achieve different trade-offs between the objectives. Through an extensive set of experiments and ablations, we show that the CLP framework learns steerable models that outperform and Pareto-dominate the current state-of-the-art approaches for multi-objective finetuning.
Authors: Manfred Georg, Garrett Tanzer, Saad Hassan, Maximus Shengelia, Esha Uboweja, Sam Sepah, Sean Forbes, Thad Starner
Abstract: Progress in machine understanding of sign languages has been slow and hampered by limited data. In this paper, we present FSboard, an American Sign Language fingerspelling dataset situated in a mobile text entry use case, collected from 147 paid and consenting Deaf signers using Pixel 4A selfie cameras in a variety of environments. Fingerspelling recognition is an incomplete solution that is only one small part of sign language translation, but it could provide some immediate benefit to Deaf/Hard of Hearing signers as more broadly capable technology develops. At >3 million characters in length and >250 hours in duration, FSboard is the largest fingerspelling recognition dataset to date by a factor of >10x. As a simple baseline, we finetune 30 Hz MediaPipe Holistic landmark inputs into ByT5-Small and achieve 11.1% Character Error Rate (CER) on a test set with unique phrases and signers. This quality degrades gracefully when decreasing frame rate and excluding face/body landmarks: plausible optimizations to help models run on device in real time.
Authors: Yulin Xu, Zhen Yang, Fandong Meng, JieZhou
Abstract: Complete Multi-lingual Neural Machine Translation (C-MNMT) achieves superior performance against the conventional MNMT by constructing multi-way aligned corpus, i.e., aligning bilingual training examples from different language pairs when either their source or target sides are identical. However, since exactly identical sentences from different language pairs are scarce, the power of the multi-way aligned corpus is limited by its scale. To handle this problem, this paper proposes "Extract and Generate" (EAG), a two-step approach to construct large-scale and high-quality multi-way aligned corpus from bilingual data. Specifically, we first extract candidate aligned examples by pairing the bilingual examples from different language pairs with highly similar source or target sentences; and then generate the final aligned examples from the candidates with a well-trained generation model. With this two-step pipeline, EAG can construct a large-scale and multi-way aligned corpus whose diversity is almost identical to the original bilingual corpus. Experiments on two publicly available datasets i.e., WMT-5 and OPUS-100, show that the proposed method achieves significant improvements over strong baselines, with +1.1 and +1.4 BLEU points improvements on the two datasets respectively.
Authors: Fenglin Liu, Bang Yang, Chenyu You, Xian Wu, Shen Ge, Zhangdaihong Liu, Xu Sun, Yang Yang, David A. Clifton
Abstract: Language models (LMs), including large language models (such as ChatGPT), have the potential to assist clinicians in generating various clinical notes. However, LMs are prone to produce ``hallucinations'', i.e., generated content that is not aligned with facts and knowledge. In this paper, we propose the Re$^3$Writer method with retrieval-augmented generation and knowledge-grounded reasoning to enable LMs to generate faithful clinical texts. We demonstrate the effectiveness of our method in generating patient discharge instructions. It requires the LMs not to only understand the patients' long clinical documents, i.e., the health records during hospitalization, but also to generate critical instructional information provided both to carers and to the patient at the time of discharge. The proposed Re$^3$Writer imitates the working patterns of physicians to first \textbf{re}trieve related working experience from historical instructions written by physicians, then \textbf{re}ason related medical knowledge. Finally, it \textbf{re}fines the retrieved working experience and reasoned medical knowledge to extract useful information, which is used to generate the discharge instructions for previously-unseen patients. Our experiments show that, using our method, the performance of five representative LMs can be substantially boosted across all metrics. Meanwhile, we show results from human evaluations to measure the effectiveness in terms of fluency, faithfulness, and comprehensiveness.
Authors: Bruce W. Lee, Jason Hyung-Jong Lee
Abstract: Traditional English readability formulas, or equations, were largely developed in the 20th century. Nonetheless, many researchers still rely on them for various NLP applications. This phenomenon is presumably due to the convenience and straightforwardness of readability formulas. In this work, we contribute to the NLP community by 1. introducing New English Readability Formula (NERF), 2. recalibrating the coefficients of old readability formulas (Flesch-Kincaid Grade Level, Fog Index, SMOG Index, Coleman-Liau Index, and Automated Readability Index), 3. evaluating the readability formulas, for use in text simplification studies and medical texts, and 4. developing a Python-based program for the wide application to various NLP projects.
Authors: Shizhe Diao, Pengcheng Wang, Yong Lin, Rui Pan, Xiang Liu, Tong Zhang
Abstract: The increasing scale of large language models (LLMs) brings emergent abilities to various complex tasks requiring reasoning, such as arithmetic and commonsense reasoning. It is known that the effective design of task-specific prompts is critical for LLMs' ability to produce high-quality answers. In particular, an effective approach for complex question-and-answer tasks is example-based prompting with chain-of-thought (CoT) reasoning, which significantly improves the performance of LLMs. However, current CoT methods rely on a fixed set of human-annotated exemplars, which are not necessarily the most effective examples for different tasks. This paper proposes a new method, Active-Prompt, to adapt LLMs to different tasks with task-specific example prompts (annotated with human-designed CoT reasoning). For this purpose, we propose a solution to the key problem of determining which questions are the most important and helpful ones to annotate from a pool of task-specific queries. By borrowing ideas from the related problem of uncertainty-based active learning, we introduce several metrics to characterize the uncertainty so as to select the most uncertain questions for annotation. Experimental results demonstrate the superiority of our proposed method, achieving state-of-the-art on eight complex reasoning tasks. Further analyses of different uncertainty metrics, pool sizes, zero-shot learning, and accuracy-uncertainty relationship demonstrate the effectiveness of our method. Our code will be available at https://github.com/shizhediao/active-prompt.
Authors: Kai Zhang, Rong Zhou, Eashan Adhikarla, Zhiling Yan, Yixin Liu, Jun Yu, Zhengliang Liu, Xun Chen, Brian D. Davison, Hui Ren, Jing Huang, Chen Chen, Yuyin Zhou, Sunyang Fu, Wei Liu, Tianming Liu, Xiang Li, Yong Chen, Lifang He, James Zou, Quanzheng Li, Hongfang Liu, Lichao Sun
Abstract: Traditional biomedical artificial intelligence (AI) models, designed for specific tasks or modalities, often exhibit limited flexibility in real-world deployment and struggle to utilize holistic information. Generalist AI holds the potential to address these limitations due to its versatility in interpreting different data types and generating tailored outputs for diverse needs. However, existing biomedical generalist AI solutions are typically heavyweight and closed source to researchers, practitioners, and patients. Here, we propose BiomedGPT, the first open-source and lightweight vision-language foundation model, designed as a generalist capable of performing various biomedical tasks. BiomedGPT achieved state-of-the-art results in 16 out of 25 experiments while maintaining a computing-friendly model scale. We also conducted human evaluations to assess the capabilities of BiomedGPT in radiology visual question answering, report generation, and summarization. BiomedGPT exhibits robust prediction ability with a low error rate of 3.8% in question answering, satisfactory performance with an error rate of 8.3% in writing complex radiology reports, and competitive summarization ability with a nearly equivalent preference score to human experts. Our method demonstrates that effective training with diverse data can lead to more practical biomedical AI for improving diagnosis and workflow efficiency.
Authors: Ohad Rubin, Jonathan Berant
Abstract: Retrieval-augmented language models (LMs) have received much attention recently. However, typically the retriever is not trained jointly as a native component of the LM, but added post-hoc to an already-pretrained LM, which limits the ability of the LM and the retriever to adapt to one another. In this work, we propose the Retrieval-Pretrained Transformer (RPT), an architecture and training procedure for jointly training a retrieval-augmented LM from scratch and apply it to the task of modeling long texts. Given a recently generated text chunk in a long document, the LM computes query representations, which are then used to retrieve earlier chunks in the document, located potentially tens of thousands of tokens before. Information from retrieved chunks is fused into the LM representations to predict the next target chunk. We train the retriever component with a semantic objective, where the goal is to retrieve chunks that increase the probability of the next chunk, according to a reference LM. We evaluate RPT on four long-range language modeling tasks, spanning books, code, and mathematical writing, and demonstrate that RPT improves retrieval quality and subsequently perplexity across the board compared to strong baselines.
Authors: Harini S I, Somesh Singh, Yaman K Singla, Aanisha Bhattacharyya, Veeky Baths, Changyou Chen, Rajiv Ratn Shah, Balaji Krishnamurthy
Abstract: Marketers spend billions of dollars on advertisements, but to what end? At purchase time, if customers cannot recognize the brand for which they saw an ad, the money spent on the ad is essentially wasted. Despite its importance in marketing, until now, there has been no large-scale study on the memorability of ads. All previous memorability studies have been conducted on short-term recall on specific content types like action videos. On the other hand, the advertising industry only cares about long-term memorability, and ads are almost always highly multimodal. Therefore, we release the first memorability dataset, LAMBDA, consisting of 1749 participants and 2205 ads covering 276 brands. Running statistical tests over different participant subpopulations and ad types, we find many interesting insights into what makes an ad memorable, e.g., fast-moving ads are more memorable than those with slower scenes; people who use ad-blockers remember a lower number of ads than those who don't. Next, we present a model, Henry, to predict the memorability of a content. Henry achieves state-of-the-art performance across all prominent literature memorability datasets. It shows strong generalization performance with better results in 0-shot on unseen datasets. Finally, with the intent of memorable ad generation, we present a scalable method to build a high-quality memorable ad generation model by leveraging automatically annotated data. Our approach, SEED (Self rEwarding mEmorability Modeling), starts with a language model trained on LAMBDA as seed data and progressively trains an LLM to generate more memorable ads. We show that the generated advertisements have 44% higher memorability scores than the original ads. We release this large-scale ad dataset, UltraLAMBDA, consisting of 5 million ads. Our code and datasets are available at https://behavior-in-the-wild.github.io/memorability.
Authors: Hiromu Yakura
Abstract: Metaphors and sarcasm are precious fruits of our highly evolved social communication skills. However, children with the condition then known as Asperger syndrome are known to have difficulties in comprehending sarcasm, even if they possess adequate verbal IQs for understanding metaphors. Accordingly, researchers had employed a screening test that assesses metaphor and sarcasm comprehension to distinguish Asperger syndrome from other conditions with similar external behaviors (e.g., attention-deficit/hyperactivity disorder). This study employs a standardized test to evaluate recent large language models' (LLMs) understanding of nuanced human communication. The results indicate improved metaphor comprehension with increased model parameters; however, no similar improvement was observed for sarcasm comprehension. Considering that a human's ability to grasp sarcasm has been associated with the amygdala, a pivotal cerebral region for emotional learning, a distinctive strategy for training LLMs would be imperative to imbue them with the ability in a cognitively grounded manner.
Authors: Vernon Toh Yan Han, Ratish Puduppully, Nancy F. Chen
Abstract: Large Language Models (LLMs), combined with program-based solving techniques, are increasingly demonstrating proficiency in mathematical reasoning. For example, closed-source models such as OpenAI GPT-4 and Claude show excellent results in solving math word problems. However, progress in math word problem-solving for open-source LLMs is limited, and the challenges these models face are not well-studied. In this paper, we study the performance of strong open-source LLMs, including Llama 2 (7B), Code Llama (7B), and Mistral (7B) on math word problems using program-based solving techniques. Specifically, we analyze the outputs of these models when applied to math word problems and identify a category of problems that pose a significant challenge, particularly those involving quantities spanning multiple units. To address this issue, we propose a systematic approach by defining the units for each quantity and ensuring the consistency of these units during mathematical operations. We developed Unit Consistency Programs (UCPs), an annotated dataset of math word problems, each paired with programs containing unit specifications and unit verification routines. We fine-tuned Llama 2 (7B), Code Llama (7B), and Mistral (7B) models with UCPs to produce theirVerityMath variants. Our findings indicate that our approach, which incorporates unit consistency, currently slightly underperforms compared to an approach that does not. To understand the reasons behind this, we conduct an in-depth error analysis and suggest options for future improvements. Our code and dataset are available at https://github.com/vernontoh/VerityMath.
Authors: Chengshu Li, Jacky Liang, Andy Zeng, Xinyun Chen, Karol Hausman, Dorsa Sadigh, Sergey Levine, Li Fei-Fei, Fei Xia, Brian Ichter
Abstract: Code provides a general syntactic structure to build complex programs and perform precise computations when paired with a code interpreter - we hypothesize that language models (LMs) can leverage code-writing to improve Chain of Thought reasoning not only for logic and arithmetic tasks, but also for semantic ones (and in particular, those that are a mix of both). For example, consider prompting an LM to write code that counts the number of times it detects sarcasm in an essay: the LM may struggle to write an implementation for "detect_sarcasm(string)" that can be executed by the interpreter (handling the edge cases would be insurmountable). However, LMs may still produce a valid solution if they not only write code, but also selectively "emulate" the interpreter by generating the expected output of "detect_sarcasm(string)". In this work, we propose Chain of Code (CoC), a simple yet surprisingly effective extension that improves LM code-driven reasoning. The key idea is to encourage LMs to format semantic sub-tasks in a program as flexible pseudocode that the interpreter can explicitly catch undefined behaviors and hand off to simulate with an LM (as an "LMulator"). Experiments demonstrate that Chain of Code outperforms Chain of Thought and other baselines across a variety of benchmarks; on BIG-Bench Hard, Chain of Code achieves 84%, a gain of 12% over Chain of Thought. In a nutshell, CoC broadens the scope of reasoning questions that LMs can answer by "thinking in code".
Authors: Min-Han Shih, Ho-Lam Chung, Yu-Chi Pai, Ming-Hao Hsu, Guan-Ting Lin, Shang-Wen Li, Hung-yi Lee
Abstract: In recent advancements in spoken question answering (QA), end-to-end models have made significant strides. However, previous research has primarily focused on extractive span selection. While this extractive-based approach is effective when answers are present directly within the input, it falls short in addressing abstractive questions, where answers are not directly extracted but inferred from the given information. To bridge this gap, we introduce the first end-to-end Generative Spoken Question Answering (GSQA) model that empowers the system to engage in abstractive reasoning. The challenge in training our GSQA model lies in the absence of a spoken abstractive QA dataset. We propose using text models for initialization and leveraging the extractive QA dataset to transfer knowledge from the text generative model to the spoken generative model. Experimental results indicate that our model surpasses the previous extractive model by 3% on extractive QA datasets. Furthermore, the GSQA model has only been fine-tuned on the spoken extractive QA dataset. Despite not having seen any spoken abstractive QA data, it can still closely match the performance of the cascade model. In conclusion, our GSQA model shows the potential to generalize to a broad spectrum of questions, thus further expanding the spoken question answering capabilities of abstractive QA. Our code is available at https://voidful.github.io/GSQA
Authors: Vladim\'ir Bo\v{z}a
Abstract: Pruning large language models (LLMs) is a challenging task due to their enormous size. The primary difficulty is fine-tuning the model after pruning, which is needed to recover the lost performance caused by dropping weights. Recent approaches have either ignored fine-tuning entirely, focusing on efficient pruning criteria, or attempted layer-wise weight updates, preserving the behavior of each layer. However, even layer-wise weight updates can be costly for LLMs, and previous works have resorted to various approximations. In our paper, we propose a fast and effective weight update algorithm for pruned layers based on the Alternating Direction Method of Multipliers (ADMM). We further extend it with a simple gradual pruning mask selection and achieve state-of-the-art pruning performance across a wide range of LLMs. Code is available at https://github.com/fmfi-compbio/admm-pruning.
Authors: Israa Jaradat, Haiqi Zhang, Chengkai Li
Abstract: Cherry-picking refers to the deliberate selection of evidence or facts that favor a particular viewpoint while ignoring or distorting evidence that supports an opposing perspective. Manually identifying cherry-picked statements in news stories can be challenging. In this study, we introduce a novel approach to detecting cherry-picked statements by identifying missing important statements in a target news story using language models and contextual information from other news sources. Furthermore, this research introduces a novel dataset specifically designed for training and evaluating cherry-picking detection models. Our best performing model achieves an F-1 score of about 89% in detecting important statements. Moreover, results show the effectiveness of incorporating external knowledge from alternative narratives when assessing statement importance.
Authors: Zhuoyuan Mao, Yen Yu
Abstract: This article introduces contrastive alignment instructions (AlignInstruct) to address two challenges in machine translation (MT) on large language models (LLMs). One is the expansion of supported languages to previously unseen ones. The second relates to the lack of data in low-resource languages. Model fine-tuning through MT instructions (MTInstruct) is a straightforward approach to the first challenge. However, MTInstruct is limited by weak cross-lingual signals inherent in the second challenge. AlignInstruct emphasizes cross-lingual supervision via a cross-lingual discriminator built using statistical word alignments. Our results based on fine-tuning the BLOOMZ models (1b1, 3b, and 7b1) in up to 24 unseen languages showed that: (1) LLMs can effectively translate unseen languages using MTInstruct; (2) AlignInstruct led to consistent improvements in translation quality across 48 translation directions involving English; (3) Discriminator-based instructions outperformed their generative counterparts as cross-lingual instructions; (4) AlignInstruct improved performance in 30 zero-shot directions.
Authors: Divyanshu Aggarwal, Ashutosh Sathe, Ishaan Watts, Sunayana Sitaram
Abstract: Parameter Efficient Finetuning (PEFT) has emerged as a viable solution for improving the performance of Large Language Models (LLMs) without requiring massive resources and compute. Prior work on multilingual evaluation has shown that there is a large gap between the performance of LLMs on English and other languages. Further, there is also a large gap between the performance of smaller open-source models and larger LLMs. Finetuning can be an effective way to bridge this gap and make language models more equitable. In this work, we finetune the LLama-2-7B and Mistral-7B models on two synthetic multilingual instruction tuning datasets to determine its effect on model performance on six downstream tasks covering forty languages in all. Additionally, we experiment with various parameters, such as rank for low-rank adaptation and values of quantisation to determine their effects on downstream performance and find that higher rank and higher quantisation values benefit low-resource languages. We find that PEFT of smaller open-source models sometimes bridges the gap between the performance of these models and the larger ones, however, English performance can take a hit. We also find that finetuning sometimes improves performance on low-resource languages, while degrading performance on high-resource languages.
Authors: Saibo Geng, Berkay D\"oner, Chris Wendler, Martin Josifoski, Robert West
Abstract: Constrained decoding, a technique for enforcing constraints on language model outputs, offers a way to control text generation without retraining or architectural modifications. Its application is, however, typically restricted to models that give users access to next-token distributions (usually via softmax logits), which poses a limitation with blackbox large language models (LLMs). This paper introduces sketch-guided constrained decoding (SGCD), a novel approach to constrained decoding for blackbox LLMs, which operates without access to the logits of the blackbox LLM. SGCD utilizes a locally hosted auxiliary model to refine the output of an unconstrained blackbox LLM, effectively treating this initial output as a "sketch" for further elaboration. This approach is complementary to traditional logit-based techniques and enables the application of constrained decoding in settings where full model transparency is unavailable. We demonstrate the efficacy of SGCD through experiments in closed information extraction and constituency parsing, showing how it enhances the utility and flexibility of blackbox LLMs for complex NLP tasks.
Authors: Kuang-Huei Lee, Xinyun Chen, Hiroki Furuta, John Canny, Ian Fischer
Abstract: Current Large Language Models (LLMs) are not only limited to some maximum context length, but also are not able to robustly consume long inputs. To address these limitations, we propose ReadAgent, an LLM agent system that increases effective context length up to 20x in our experiments. Inspired by how humans interactively read long documents, we implement ReadAgent as a simple prompting system that uses the advanced language capabilities of LLMs to (1) decide what content to store together in a memory episode, (2) compress those memory episodes into short episodic memories called gist memories, and (3) take actions to look up passages in the original text if ReadAgent needs to remind itself of relevant details to complete a task. We evaluate ReadAgent against baselines using retrieval methods, using the original long contexts, and using the gist memories. These evaluations are performed on three long-document reading comprehension tasks: QuALITY, NarrativeQA, and QMSum. ReadAgent outperforms the baselines on all three tasks while extending the effective context window by 3.5-20x.
Authors: Alexis Chevalier, Jiayi Geng, Alexander Wettig, Howard Chen, Sebastian Mizera, Toni Annala, Max Jameson Aragon, Arturo Rodr\'iguez Fanlo, Simon Frieder, Simon Machado, Akshara Prabhakar, Ellie Thieu, Jiachen T. Wang, Zirui Wang, Xindi Wu, Mengzhou Xia, Wenhan Xia, Jiatong Yu, Jun-Jie Zhu, Zhiyong Jason Ren, Sanjeev Arora, Danqi Chen
Abstract: NLP has recently made exciting progress toward training language models (LMs) with strong scientific problem-solving skills. However, model development has not focused on real-life use-cases of LMs for science, including applications in education that require processing long scientific documents. To address this, we introduce TutorEval and TutorChat. TutorEval is a diverse question-answering benchmark consisting of questions about long chapters from STEM textbooks, written by experts. TutorEval helps measure real-life usability of LMs as scientific assistants, and it is the first benchmark combining long contexts, free-form generation, and multi-disciplinary scientific knowledge. Moreover, we show that fine-tuning base models with existing dialogue datasets leads to poor performance on TutorEval. Therefore, we create TutorChat, a dataset of 80,000 long synthetic dialogues about textbooks. We use TutorChat to fine-tune Llemma models with 7B and 34B parameters. These LM tutors specialized in math have a 32K-token context window, and they excel at TutorEval while performing strongly on GSM8K and MATH. Our datasets build on open-source materials, and we release our models, data, and evaluations.
Authors: Mikayel Samvelyan, Sharath Chandra Raparthy, Andrei Lupu, Eric Hambro, Aram H. Markosyan, Manish Bhatt, Yuning Mao, Minqi Jiang, Jack Parker-Holder, Jakob Foerster, Tim Rockt\"aschel, Roberta Raileanu
Abstract: As large language models (LLMs) become increasingly prevalent across many real-world applications, understanding and enhancing their robustness to adversarial attacks is of paramount importance. Existing methods for identifying adversarial prompts tend to focus on specific domains, lack diversity, or require extensive human annotations. To address these limitations, we present Rainbow Teaming, a novel black-box approach for producing a diverse collection of adversarial prompts. Rainbow Teaming casts adversarial prompt generation as a quality-diversity problem, and uses open-ended search to generate prompts that are both effective and diverse. Focusing on the safety domain, we use Rainbow Teaming to target various state-of-the-art LLMs, including the Llama 2 and Llama 3 models. Our approach reveals hundreds of effective adversarial prompts, with an attack success rate exceeding 90% across all tested models. Furthermore, we demonstrate that fine-tuning models with synthetic data generated by the Rainbow Teaming method significantly enhances their safety without sacrificing general performance or helpfulness. We additionally explore the versatility of Rainbow Teaming by applying it to question answering and cybersecurity, showcasing its potential to drive robust open-ended self-improvement in a wide range of applications.
Authors: Gaurav Verma, Minje Choi, Kartik Sharma, Jamelle Watson-Daniels, Sejoon Oh, Srijan Kumar
Abstract: Multimodal large language models (MLLMs) like LLaVA and GPT-4(V) enable general-purpose conversations about images with the language modality. As off-the-shelf MLLMs may have limited capabilities on images from domains like dermatology and agriculture, they must be fine-tuned to unlock domain-specific applications. The prevalent architecture of current open-source MLLMs comprises two major modules: an image-language (cross-modal) projection network and a large language model. It is desirable to understand the roles of these two modules in modeling domain-specific visual attributes to inform the design of future models and streamline the interpretability efforts on the current models. To this end, via experiments on 4 datasets and under 2 fine-tuning settings, we find that as the MLLM is fine-tuned, it indeed gains domain-specific visual capabilities, but the updates do not lead to the projection extracting relevant domain-specific visual attributes. Our results indicate that the domain-specific visual attributes are modeled by the LLM, even when only the projection is fine-tuned. Through this study, we offer a potential reinterpretation of the role of cross-modal projections in MLLM architectures. Project webpage: https://claws-lab.github.io/projection-in-MLLMs/
Authors: Yibin Lei, Di Wu, Tianyi Zhou, Tao Shen, Yu Cao, Chongyang Tao, Andrew Yates
Abstract: We introduce a new unsupervised text embedding method, Meta-Task Prompting with Explicit One-Word Limitation (MetaEOL), for generating high-quality sentence embeddings from Large Language Models (LLMs) without the need for model fine-tuning. Leveraging meta-task prompting, MetaEOL guides LLMs to produce embeddings through a series of carefully designed prompts that address multiple representational aspects. Our comprehensive experiments demonstrate that embeddings averaged from various meta-tasks are versatile embeddings that yield competitive performance on Semantic Textual Similarity (STS) benchmarks and excel in downstream tasks, surpassing contrastive-trained models. Our findings suggest a new scaling law, offering a versatile and resource-efficient approach for embedding generation across diverse scenarios.
Authors: Md Tahmid Rahman Laskar, Elena Khasanova, Xue-Yong Fu, Cheng Chen, Shashi Bhushan TN
Abstract: This work focuses on the task of query-based meeting summarization in which the summary of a context (meeting transcript) is generated in response to a specific query. When using Large Language Models (LLMs) for this task, usually a new call to the LLM inference endpoint/API is triggered for each new query, even if the context stays the same. However, repeated calls to the LLM inference endpoints would significantly increase the costs of using them in production, making LLMs impractical for many real-world use cases. To address this problem, in this paper, we investigate whether combining the queries for the same input context in a single prompt to minimize repeated calls can be successfully used in meeting summarization. In this regard, we conduct extensive experiments by comparing the performance of various popular LLMs: GPT-4, Gemini, Claude-3, LLaMA-2, Mistral, Phi-3, and Qwen-2 in single-query and multi-query settings. We observe that 100% reliability in generating the response in the expected format is usually limited to certain closed-source LLMs, with most open-source LLMs lagging behind (except a few 7B parameters LLMs like Mistral and Phi-3). We conclude that multi-query prompting could be useful to significantly optimize the inference costs in meeting summarization.
Authors: Tongyao Zhu, Qian Liu, Liang Pang, Zhengbao Jiang, Min-Yen Kan, Min Lin
Abstract: Recent developments in Language Models (LMs) have shown their effectiveness in NLP tasks, particularly in knowledge-intensive tasks. However, the mechanisms underlying knowledge storage and memory access within their parameters remain elusive. In this paper, we investigate whether a generative LM (e.g., GPT-2) is able to access its memory sequentially or randomly. Through carefully-designed synthetic tasks, covering the scenarios of full recitation, selective recitation and grounded question answering, we reveal that LMs manage to sequentially access their memory while encountering challenges in randomly accessing memorized content. We find that techniques including recitation and permutation improve the random memory access capability of LMs. Furthermore, by applying this intervention to realistic scenarios of open-domain question answering, we validate that enhancing random access by recitation leads to notable improvements in question answering. The code to reproduce our experiments can be found at https://github.com/sail-sg/lm-random-memory-access.
Authors: Yusheng Liao, Yutong Meng, Yuhao Wang, Hongcheng Liu, Yanfeng Wang, Yu Wang
Abstract: Large Language Models (LLMs) have demonstrated remarkable proficiency in human interactions, yet their application within the medical field remains insufficiently explored. Previous works mainly focus on the performance of medical knowledge with examinations, which is far from the realistic scenarios, falling short in assessing the abilities of LLMs on clinical tasks. In the quest to enhance the application of Large Language Models (LLMs) in healthcare, this paper introduces the Automated Interactive Evaluation (AIE) framework and the State-Aware Patient Simulator (SAPS), targeting the gap between traditional LLM evaluations and the nuanced demands of clinical practice. Unlike prior methods that rely on static medical knowledge assessments, AIE and SAPS provide a dynamic, realistic platform for assessing LLMs through multi-turn doctor-patient simulations. This approach offers a closer approximation to real clinical scenarios and allows for a detailed analysis of LLM behaviors in response to complex patient interactions. Our extensive experimental validation demonstrates the effectiveness of the AIE framework, with outcomes that align well with human evaluations, underscoring its potential to revolutionize medical LLM testing for improved healthcare delivery.
Authors: Mahdi Rezapour
Abstract: In this study, we explore the application of transformer-based models for emotion classification on text data. We train and evaluate several pre-trained transformer models, on the Emotion dataset using different variants of transformers. The paper also analyzes some factors that in-fluence the performance of the model, such as the fine-tuning of the transformer layer, the trainability of the layer, and the preprocessing of the text data. Our analysis reveals that commonly applied techniques like removing punctuation and stop words can hinder model performance. This might be because transformers strength lies in understanding contextual relationships within text. Elements like punctuation and stop words can still convey sentiment or emphasis and removing them might disrupt this context.
Authors: Huaiwen Zhang, Yu Chen, Ming Wang, Shi Feng
Abstract: Emotional Support Conversation (ESC) is a typical dialogue that can effectively assist the user in mitigating emotional pressures. However, owing to the inherent subjectivity involved in analyzing emotions, current non-artificial methodologies face challenges in effectively appraising the emotional support capability. These metrics exhibit a low correlation with human judgments. Concurrently, manual evaluation methods extremely will cause high costs. To solve these problems, we propose a novel model FEEL (Framework for Evaluating Emotional Support Capability with Large Lan-guage Models), employing Large Language Models (LLMs) as evaluators to assess emotional support capabilities. The model meticulously considers various evaluative aspects of ESC to apply a more comprehensive and accurate evaluation method for ESC. Additionally, it employs a probability distribution approach for a more stable result and integrates an ensemble learning strategy, leveraging multiple LLMs with assigned weights to enhance evaluation accuracy. To appraise the performance of FEEL, we conduct extensive experiments on existing ESC model dialogues. Experimental results demonstrate our model exhibits a substantial enhancement in alignment with human evaluations compared to the baselines. Our source code is available at https://github.com/Ansisy/FEEL.
Authors: Rushang Karia, Daniel Bramblett, Daksh Dobhal, Pulkit Verma, Siddharth Srivastava
Abstract: This paper presents $\forall$uto$\exists$val, a new approach for scaling LLM assessment in translating formal syntax -- such as first-order logic, regular expressions, etc -- to natural language (interpretation) or vice versa (compilation), thereby facilitating their use in applications such as generating/explaining logic and control flow for programs etc. Existing approaches for LLM assessment in these areas require labor-intensive ground-truth creation, the availability of which undermines the separation of training and test sets. Furthermore, such datasets typically include relatively few hand-coded test cases over which LLM accuracy is determined, thus making them inadequate for determining the safety or correctness of their generated outputs. We introduce a new approach that utilizes context-free grammars (CFGs) to generate out-of-distribution datasets on the fly and perform closed-loop testing of LLM capabilities using formal verifiers to guarantee the correctness of LLM outputs without any human intervention. We release our dataset and benchmark as open-source code at \url{https://github.com/AAIR-lab/auto-llm-assessment}. We also conduct an assessment of several SOTA closed and open-source LLMs to showcase the feasibility and scalability of this paradigm. Our experiments reveal that SOTA LLMs are unable to solve the formal translation task adequately.
Authors: Thibaut Thonet, Jos Rozen, Laurent Besacier
Abstract: Research on Large Language Models (LLMs) has recently witnessed an increasing interest in extending models' context size to better capture dependencies within long documents. While benchmarks have been proposed to assess long-range abilities, existing efforts primarily considered generic tasks that are not necessarily aligned with real-world applications. In contrast, our work proposes a new benchmark for long-context LLMs focused on a practical meeting assistant scenario. In this scenario, the long contexts consist of transcripts obtained by automatic speech recognition, presenting unique challenges for LLMs due to the inherent noisiness and oral nature of such data. Our benchmark, named ELITR-Bench, augments the existing ELITR corpus' transcripts with 271 manually crafted questions and their ground-truth answers. Our experiments with recent long-context LLMs on ELITR-Bench highlight a gap between open-source and proprietary models, especially when questions are asked sequentially within a conversation. We also provide a thorough analysis of our GPT-4-based evaluation method, encompassing insights from a crowdsourcing study. Our findings suggest that while GPT-4's evaluation scores are correlated with human judges', its ability to differentiate among more than three score levels may be limited.
Authors: Pengda Wang, Zilin Xiao, Hanjie Chen, Frederick L. Oswald
Abstract: Although large language models (LLMs) have demonstrated remarkable proficiency in modeling text and generating human-like text, they may exhibit biases acquired from training data in doing so. Specifically, LLMs may be susceptible to a common cognitive trap in human decision-making called the representativeness heuristic. This is a concept in psychology that refers to judging the likelihood of an event based on how closely it resembles a well-known prototype or typical example, versus considering broader facts or statistical evidence. This research investigates the impact of the representativeness heuristic on LLM reasoning. We created ReHeAT (Representativeness Heuristic AI Testing), a dataset containing a series of problems spanning six common types of representativeness heuristics. Experiments reveal that four LLMs applied to ReHeAT all exhibited representativeness heuristic biases. We further identify that the model's reasoning steps are often incorrectly based on a stereotype rather than on the problem's description. Interestingly, the performance improves when adding a hint in the prompt to remind the model to use its knowledge. This suggests the uniqueness of the representativeness heuristic compared to traditional biases. It can occur even when LLMs possess the correct knowledge while falling into a cognitive trap. This highlights the importance of future research focusing on the representativeness heuristic in model reasoning and decision-making and on developing solutions to address it.
Authors: Chaoqun He, Renjie Luo, Shengding Hu, Yuanqian Zhao, Jie Zhou, Hanghao Wu, Jiajie Zhang, Xu Han, Zhiyuan Liu, Maosong Sun
Abstract: Evaluation is pivotal for refining Large Language Models (LLMs), pinpointing their capabilities, and guiding enhancements. The rapid development of LLMs calls for a lightweight and easy-to-use framework for swift evaluation deployment. However, considering various implementation details, developing a comprehensive evaluation platform is never easy. Existing platforms are often complex and poorly modularized, hindering seamless incorporation into research workflows. This paper introduces UltraEval, a user-friendly evaluation framework characterized by its lightweight nature, comprehensiveness, modularity, and efficiency. We identify and reimplement three core components of model evaluation (models, data, and metrics). The resulting composability allows for the free combination of different models, tasks, prompts, benchmarks, and metrics within a unified evaluation workflow. Additionally, UltraEval supports diverse models owing to a unified HTTP service and provides sufficient inference acceleration. UltraEval is now available for researchers publicly.
Authors: Kota Tanabe, Masahiro Suzuki, Hiroki Sakaji, Itsuki Noda
Abstract: We construct an instruction dataset for the large language model (LLM) in the Japanese finance domain. Domain adaptation of language models, including LLMs, is receiving more attention as language models become more popular. This study demonstrates the effectiveness of domain adaptation through instruction tuning. To achieve this, we propose an instruction tuning data in Japanese called JaFIn, the Japanese Financial Instruction Dataset. JaFIn is manually constructed based on multiple data sources, including Japanese government websites, which provide extensive financial knowledge. We then utilize JaFIn to apply instruction tuning for several LLMs, demonstrating that our models specialized in finance have better domain adaptability than the original models. The financial-specialized LLMs created were evaluated using a quantitative Japanese financial benchmark and qualitative response comparisons, showing improved performance over the originals.
Authors: Ling Yue, Sixue Xing, Jintai Chen, Tianfan Fu
Abstract: Large Language Models (LLMs) and multi-agent systems have shown impressive capabilities in natural language tasks but face challenges in clinical trial applications, primarily due to limited access to external knowledge. Recognizing the potential of advanced clinical trial tools that aggregate and predict based on the latest medical data, we propose an integrated solution to enhance their accessibility and utility. We introduce Clinical Agent System (ClinicalAgent), a clinical multi-agent system designed for clinical trial tasks, leveraging GPT-4, multi-agent architectures, LEAST-TO-MOST, and ReAct reasoning technology. This integration not only boosts LLM performance in clinical contexts but also introduces novel functionalities. The proposed method achieves competitive predictive performance in clinical trial outcome prediction (0.7908 PR-AUC), obtaining a 0.3326 improvement over the standard prompt Method. Publicly available code can be found at https://anonymous.4open.science/r/ClinicalAgent-6671.
Authors: Dengchun Li, Yingzi Ma, Naizheng Wang, Zhengmao Ye, Zhiyuan Cheng, Yinghao Tang, Yan Zhang, Lei Duan, Jie Zuo, Cal Yang, Mingjie Tang
Abstract: Fine-tuning Large Language Models (LLMs) is a common practice to adapt pre-trained models for specific applications. While methods like LoRA have effectively addressed GPU memory constraints during fine-tuning, their performance often falls short, especially in multi-task scenarios. In contrast, Mixture-of-Expert (MoE) models, such as Mixtral 8x7B, demonstrate remarkable performance in multi-task learning scenarios while maintaining a reduced parameter count. However, the resource requirements of these MoEs remain challenging, particularly for consumer-grade GPUs with less than 24GB memory. To tackle these challenges, we propose MixLoRA, an approach to construct a resource-efficient sparse MoE model based on LoRA. MixLoRA inserts multiple LoRA-based experts within the feed-forward network block of a frozen pre-trained dense model and employs a commonly used top-k router. Unlike other LoRA-based MoE methods, MixLoRA enhances model performance by utilizing independent attention-layer LoRA adapters. Additionally, an auxiliary load balance loss is employed to address the imbalance problem of the router. Our evaluations show that MixLoRA improves about 9% accuracy compared to state-of-the-art PEFT methods in multi-task learning scenarios. We also propose a new high-throughput framework to alleviate the computation and memory bottlenecks during the training and inference of MOE models. This framework reduces GPU memory consumption by 40% and token computation latency by 30% during both training and inference.
Authors: Zihan Yan, Yaohong Xiang, Yun Huang
Abstract: Non-cognitive skills are crucial for personal and social life well-being, and such skill development can be supported by narrative-based (e.g., storytelling) technologies. While generative AI enables interactive and role-playing storytelling, little is known about how users engage with and perceive the use of AI in social life simulation for non-cognitive skills learning. Additionally, the benefits of AI mentorship on self-reflection awareness and ability in this context remain largely underexplored. To this end, we introduced Simulife++, an interactive platform enabled by a large language model (LLM). The system allows users to act as protagonists, creating stories with one or multiple AI-based characters in diverse social scenarios. In particular, we expanded the Human-AI interaction to a Human-AI-AI collaboration by including a Sage Agent, who acts as a bystander, providing users with some perspectives and guidance on their choices and conversations in terms of non-cognitive skills to promote reflection. In a within-subject user study, our quantitative results reveal that, when accompanied by Sage Agent, users exhibit significantly higher levels of reflection on motivation, self-perceptions, and resilience & coping, along with an enhanced experience of narrative transportation. Additionally, our qualitative findings suggest that Sage Agent plays a crucial role in promoting reflection on non-cognitive skills, enhancing social communication and decision-making performance, and improving overall user experience within Simulife++. Multiple supportive relationships between Sage Agent and users were also reported. We offer design implications for the application of generative AI in narrative solutions and the future potential of Sage Agent for non-cognitive skill development in broader social contexts.
Authors: Emre Onal, Klemens Fl\"oge, Emma Caldwell, Arsen Sheverdin, Vincent Fortuin
Abstract: Fine-tuned Large Language Models (LLMs) often suffer from overconfidence and poor calibration, particularly when fine-tuned on small datasets. To address these challenges, we propose a simple combination of Low-Rank Adaptation (LoRA) with Gaussian Stochastic Weight Averaging (SWAG), facilitating approximate Bayesian inference in LLMs. Through extensive testing across several Natural Language Processing (NLP) benchmarks, we demonstrate that our straightforward and computationally efficient approach improves model generalization and calibration competitively with comparable, more sophisticated methods for Bayesian inference in LLMs. We further show that our method exhibits greater robustness against distribution shift, as reflected in its improved performance on out-of-distribution tasks.
Authors: Zijin Hong, Jian Liu
Abstract: Event Extraction (EE) is an essential information extraction task that aims to extract event-related information from unstructured texts. The paradigm of this task has shifted from conventional classification-based methods to more contemporary question-answering-based (QA-based) approaches. However, in QA-based EE, the quality of the questions dramatically affects the extraction accuracy, and how to generate high-quality questions for QA-based EE remains a challenge. In this work, to tackle this challenge, we suggest four criteria to evaluate the quality of a question and propose a reinforcement learning method, RLQG, for QA-based EE that can generate generalizable, high-quality, and context-dependent questions and provides clear guidance to QA models. The extensive experiments conducted on ACE and RAMS datasets have strongly validated our approach's effectiveness, which also demonstrates its robustness in scenarios with limited training data. The corresponding code of RLQG is released for further research.
Authors: Manav Chaudhary, Harshit Gupta, Vasudeva Varma
Abstract: The proliferation of LLMs in various NLP tasks has sparked debates regarding their reliability, particularly in annotation tasks where biases and hallucinations may arise. In this shared task, we address the challenge of distinguishing annotations made by LLMs from those made by human domain experts in the context of COVID-19 symptom detection from tweets in Latin American Spanish. This paper presents BrainStorm @ iRELs approach to the SMM4H 2024 Shared Task, leveraging the inherent topical information in tweets, we propose a novel approach to identify and classify annotations, aiming to enhance the trustworthiness of annotated data.
Authors: Ying Zhang, Xiaofeng Li, Zhaoyang Liu, Haipeng Zhang
Abstract: The life trajectories of notable people have been studied to pinpoint the times and places of significant events such as birth, death, education, marriage, competition, work, speeches, scientific discoveries, artistic achievements, and battles. Understanding how these individuals interact with others provides valuable insights for broader research into human dynamics. However, the scarcity of trajectory data in terms of volume, density, and inter-person interactions, limits relevant studies from being comprehensive and interactive. We mine millions of biography pages from Wikipedia and tackle the generalization problem stemming from the variety and heterogeneity of the trajectory descriptions. Our ensemble model COSMOS, which combines the idea of semi-supervised learning and contrastive learning, achieves an F1 score of 85.95%. For this task, we also create a hand-curated dataset, WikiLifeTrajectory, consisting of 8,852 (person, time, location) triplets as ground truth. Besides, we perform an empirical analysis on the trajectories of 8,272 historians to demonstrate the validity of the extracted results. To facilitate the research on trajectory extractions and help the analytical studies to construct grand narratives, we make our code, the million-level extracted trajectories, and the WikiLifeTrajectory dataset publicly available.
Authors: Phoebe J. Wang, Max Kreminski
Abstract: Instruction-tuned large language models (LLMs) are capable of generating stories in response to open-ended user requests, but the resulting stories tend to be limited in their diversity. Older, symbolic approaches to story generation (such as planning) can generate substantially more diverse plot outlines, but are limited to producing stories that recombine a fixed set of hand-engineered character action templates. Can we combine the strengths of these approaches while mitigating their weaknesses? We propose to do so by using a higher-level and more abstract symbolic specification of high-level story structure -- implemented via answer set programming (ASP) -- to guide and diversify LLM-based story generation. Via semantic similarity analysis, we demonstrate that our approach produces more diverse stories than an unguided LLM, and via code excerpts, we demonstrate the improved compactness and flexibility of ASP-based outline generation over full-fledged narrative planning.
Authors: Anthony Costarelli, Mat Allen, Roman Hauksson, Grace Sodunke, Suhas Hariharan, Carlson Cheng, Wenjie Li, Joshua Clymer, Arjun Yadav
Abstract: Large language models have demonstrated remarkable few-shot performance on many natural language understanding tasks. Despite several demonstrations of using large language models in complex, strategic scenarios, there lacks a comprehensive framework for evaluating agents' performance across various types of reasoning found in games. To address this gap, we introduce GameBench, a cross-domain benchmark for evaluating strategic reasoning abilities of LLM agents. We focus on 9 different game environments, where each covers at least one axis of key reasoning skill identified in strategy games, and select games for which strategy explanations are unlikely to form a significant portion of models' pretraining corpuses. Our evaluations use GPT-3 and GPT-4 in their base form along with two scaffolding frameworks designed to enhance strategic reasoning ability: Chain-of-Thought (CoT) prompting and Reasoning Via Planning (RAP). Our results show that none of the tested models match human performance, and at worst GPT-4 performs worse than random action. CoT and RAP both improve scores but not comparable to human levels.
Authors: Sungwon Park, Sungwon Han, Meeyoung Cha
Abstract: The spread of fake news negatively impacts individuals and is regarded as a significant social challenge that needs to be addressed. A number of algorithmic and insightful features have been identified for detecting fake news. However, with the recent LLMs and their advanced generation capabilities, many of the detectable features (e.g., style-conversion attacks) can be altered, making it more challenging to distinguish from real news. This study proposes adversarial style augmentation, AdStyle, to train a fake news detector that remains robust against various style-conversion attacks. Our model's key mechanism is the careful use of LLMs to automatically generate a diverse yet coherent range of style-conversion attack prompts. This improves the generation of prompts that are particularly difficult for the detector to handle. Experiments show that our augmentation strategy improves robustness and detection performance when tested on fake news benchmark datasets.
Authors: Jingze Shi, Ting Xie, Bingheng Wu, Chunjun Zheng, Kai Wang
Abstract: Recent research has shown that combining Mamba with Transformer architecture, which has selective state space and quadratic self-attention mechanism, outperforms using Mamba or Transformer architecture alone in language modeling tasks. The quadratic self-attention mechanism effectively alleviates the shortcomings of selective state space in handling long-term dependencies of any element in the sequence. We propose a position information injection method that connects the selective state space model with the quadratic attention, and integrates these two architectures with hybrid experts with cross-sharing domains, so that we can enjoy the advantages of both. We design a new architecture with a more biomimetic idea: Observer-Thinker-Conceiver-Expresser (OTCE), which can compete with well-known medium-scale open-source language models on a small scale in language modeling tasks.
Authors: Erik Henriksson, Amanda Myntti, Anni Eskelinen, Selcen Erten-Johansson, Saara Hellstr\"om, Veronika Laippala
Abstract: This article explores deep learning models for the automatic identification of registers - text varieties such as news reports and discussion forums - in web-based datasets across 16 languages. Identifying web registers, or genres, is crucial for understanding the content of web-scale datasets, which have become essential in corpus and computational linguistics. Despite recent advances, the full potential of register classifiers in the noisy, unrestricted web remains largely unexplored, particularly in multilingual settings. We experiment with various deep learning models using the Multilingual CORE corpora, newly introduced in this article, which includes 16 languages annotated with a detailed, hierarchical taxonomy of 25 registers designed to cover the entire web. Our classifiers achieve state-of-the-art results using a multi-label approach, demonstrating that competitive performance is possible using a relatively complex register taxonomy. However, all models hit a performance ceiling at approximately 80% F1 score, which we attribute to the non-discrete nature of web registers and the inherent uncertainty in labeling some documents. By pruning ambiguous examples, we enhance model performance to over 90%. Additionally, multilingual models consistently outperform monolingual ones, especially benefiting languages with fewer training examples and smaller registers. Although a zero-shot setting reduces performance by an average of 7%, these drops are not correlated with specific registers or languages. Instead, we find that registers are surprisingly similar across languages.
Authors: Kibeom Nam
Abstract: Investigations into Aspect-Based Sentiment Analysis (ABSA) for Korean restaurant reviews are notably lacking in the existing literature. Our research proposes an intuitive and effective framework for ABSA in low-resource languages such as Korean. It optimizes prediction labels by integrating translated benchmark and unlabeled Korean data. Using a model fine-tuned on translated data, we pseudo-labeled the actual Korean NLI set. Subsequently, we applied LaBSE and MSP-based filtering to this pseudo-NLI set as implicit feature, enhancing Aspect Category Detection and Polarity determination through additional training. Incorporating dual filtering, this model bridged dataset gaps, achieving positive results in Korean ABSA with minimal resources. Through additional data injection pipelines, our approach aims to utilize high-resource data and construct effective models within communities, whether corporate or individual, in low-resource language countries. Compared to English ABSA, our framework showed an approximately 3% difference in F1 scores and accuracy. We release the dataset and our code for Korean ABSA, at this link.
Authors: Hwanjun Song, Hang Su, Igor Shalyminov, Jason Cai, Saab Mansour
Abstract: Automated evaluation is crucial for streamlining text summarization benchmarking and model development, given the costly and time-consuming nature of human evaluation. Traditional methods like ROUGE do not correlate well with human judgment, while recently proposed LLM-based metrics provide only summary-level assessment using Likert-scale scores. This limits deeper model analysis, e.g., we can only assign one hallucination score at the summary level, while at the sentence level, we can count sentences containing hallucinations. To remedy those limitations, we propose FineSurE, a fine-grained evaluator specifically tailored for the summarization task using large language models (LLMs). It also employs completeness and conciseness criteria, in addition to faithfulness, enabling multi-dimensional assessment. We compare various open-source and proprietary LLMs as backbones for FineSurE. In addition, we conduct extensive benchmarking of FineSurE against SOTA methods including NLI-, QA-, and LLM-based methods, showing improved performance especially on the completeness and conciseness dimensions. The code is available at https://github.com/DISL-Lab/FineSurE-ACL24.
Authors: Zheng Wang, Boxiao Jin, Zhongzhi Yu, Minjia Zhang
Abstract: How to efficiently serve Large Language Models (LLMs) has become a pressing issue because of their huge computational cost in their autoregressive generation process. To mitigate computational costs, LLMs often employ the KV Cache technique to improve the generation speed. While improving the computational efficiency, the storage requirements of the KV cache are substantial, particularly in long-context scenarios, leading to significant memory consumption. Existing KV cache eviction methods often degrade the performance of LLMs in long-context scenarios due to the information loss introduced by eviction. In this paper, we propose a novel KV cache merging approach, called KVMerger, to achieve adaptive KV cache compression for long-context tasks without significant performance degradation under constrained memory budgets. Our approach is inspired by the intriguing observation that key states exhibit high similarity at the token level within a single sequence. To facilitate merging, we develop an effective yet straightforward merging set identification algorithm to identify suitable KV states for merging. Our merging set identification algorithm stimulates the second observation that KV cache sparsity, from similarity perspective, is independent of the dataset and remains persistent at the model level. Subsequently, we propose a Gaussian kernel weighted merging algorithm to selectively merge all states within each merging set. We conduct extensive experiments to demonstrate the effectiveness of KVMerger for long-context tasks under constrained memory budgets, applying it to models including Llama2-7B-chat and Llama2-13B-chat. Using the LongBench and ZeroScroll benchmarks, we compare our method with other KV cache compression techniques, including H2O and CaM, showing that our method achieves superior performance across tasks with both 50% and 35% KV cache budgets.
Authors: Weihang Su, Yichen Tang, Qingyao Ai, Changyue Wang, Zhijing Wu, Yiqun Liu
Abstract: The emergence of Large Language Models (LLMs) has revolutionized how users access information, shifting from traditional search engines to direct question-and-answer interactions with LLMs. However, the widespread adoption of LLMs has revealed a significant challenge known as hallucination, wherein LLMs generate coherent yet factually inaccurate responses. This hallucination phenomenon has led to users' distrust in information retrieval systems based on LLMs. To tackle this challenge, this paper proposes Dynamic Retrieval Augmentation based on hallucination Detection (DRAD) as a novel method to detect and mitigate hallucinations in LLMs. DRAD improves upon traditional retrieval augmentation by dynamically adapting the retrieval process based on real-time hallucination detection. It features two main components: Real-time Hallucination Detection (RHD) for identifying potential hallucinations without external models, and Self-correction based on External Knowledge (SEK) for correcting these errors using external knowledge. Experiment results show that DRAD demonstrates superior performance in both detecting and mitigating hallucinations in LLMs. All of our code and data are open-sourced at https://github.com/oneal2000/EntityHallucination.
Authors: Roni Goldshmidt, Miriam Horovicz
Abstract: As large language models (LLMs) become increasingly prevalent in critical applications, the need for interpretable AI has grown. We introduce TokenSHAP, a novel method for interpreting LLMs by attributing importance to individual tokens or substrings within input prompts. This approach adapts Shapley values from cooperative game theory to natural language processing, offering a rigorous framework for understanding how different parts of an input contribute to a model's response. TokenSHAP leverages Monte Carlo sampling for computational efficiency, providing interpretable, quantitative measures of token importance. We demonstrate its efficacy across diverse prompts and LLM architectures, showing consistent improvements over existing baselines in alignment with human judgments, faithfulness to model behavior, and consistency. Our method's ability to capture nuanced interactions between tokens provides valuable insights into LLM behavior, enhancing model transparency, improving prompt engineering, and aiding in the development of more reliable AI systems. TokenSHAP represents a significant step towards the necessary interpretability for responsible AI deployment, contributing to the broader goal of creating more transparent, accountable, and trustworthy AI systems.
Authors: Xunyu Zhu, Jian Li, Can Ma, Weiping Wang
Abstract: Large Language Models (LLMs) have demonstrated exceptional proficiency in mathematical reasoning tasks due to their extensive parameter counts and training on vast datasets. Despite these capabilities, deploying LLMs is hindered by their computational demands. Distilling LLM mathematical reasoning into Smaller Language Models (SLMs) has emerged as a solution to this challenge, although these smaller models often suffer from errors in calculation and semantic understanding. Prior work has proposed Program-of-Thought Distillation (PoTD) to avoid calculation error. To further address semantic understanding errors, we propose Key-Point-Driven Mathematical Reasoning Distillation (KPDD). KPDD enhances the reasoning performance of SLMs by breaking down the problem-solving process into three stages: Core Question Extraction, Problem-Solving Information Extraction, and Step-by-Step Solution. This method is further divided into KPDD-CoT, which generates Chain-of-Thought rationales, and KPDD-PoT, which creates Program-of-Thought rationales. The experiment results show that KPDD-CoT significantly improves reasoning abilities, while KPDD-PoT achieves state-of-the-art performance in mathematical reasoning tasks. Our approach effectively mitigates misunderstanding errors, advancing the deployment of efficient and capable SLMs.
Authors: Zhiting Fan, Ruizhe Chen, Ruiling Xu, Zuozhu Liu
Abstract: Evaluating the bias in Large Language Models (LLMs) becomes increasingly crucial with their rapid development. However, existing evaluation methods rely on fixed-form outputs and cannot adapt to the flexible open-text generation scenarios of LLMs (e.g., sentence completion and question answering). To address this, we introduce BiasAlert, a plug-and-play tool designed to detect social bias in open-text generations of LLMs. BiasAlert integrates external human knowledge with inherent reasoning capabilities to detect bias reliably. Extensive experiments demonstrate that BiasAlert significantly outperforms existing state-of-the-art methods like GPT4-as-A-Judge in detecting bias. Furthermore, through application studies, we demonstrate the utility of BiasAlert in reliable LLM bias evaluation and bias mitigation across various scenarios. Model and code will be publicly released.
Authors: Hongyu Wang, Shuming Ma, Ruiping Wang, Furu Wei
Abstract: We introduce, Q-Sparse, a simple yet effective approach to training sparsely-activated large language models (LLMs). Q-Sparse enables full sparsity of activations in LLMs which can bring significant efficiency gains in inference. This is achieved by applying top-K sparsification to the activations and the straight-through-estimator to the training. We also introduce Block Q-Sparse for batch training and inference. The key results from this work are, (1) Q-Sparse can achieve results comparable to those of baseline LLMs while being much more efficient at inference time; (2) We present an inference-optimal scaling law for sparsely-activated LLMs; (3) Q-Sparse is effective in different settings, including training-from-scratch, continue-training of off-the-shelf LLMs, and finetuning; (4) Q-Sparse works for both full-precision and 1-bit LLMs (e.g., BitNet b1.58). Particularly, the synergy of BitNet b1.58 and Q-Sparse (can be equipped with MoE) provides the cornerstone and a clear path to revolutionize the efficiency, including cost and energy consumption, of future LLMs.
Authors: Nasim Borazjanizadeh, Steven T. Piantadosi
Abstract: Despite their linguistic competence, Large Language models (LLMs) often exhibit limitations in their ability to reason reliably and flexibly. To address this, we propose a neurosymbolic approach that prompts LLMs to extract and encode all relevant information from a problem statement as logical code statements, and then use a logic programming language (Prolog) to conduct the iterative computations of explicit deductive reasoning. Our approach significantly enhances the performance of LLMs on the standard mathematical reasoning benchmark, GSM8k, and the Navigate dataset from the BIG-bench dataset. Additionally, we introduce a novel dataset, the Non-Linear Reasoning (NLR) dataset, consisting of 55 unique word problems that target the shortcomings of the next token prediction paradigm of LLMs and require complex non-linear reasoning but only basic arithmetic skills to solve. Our findings demonstrate that the integration of Prolog enables LLMs to achieve high performance on the NLR dataset, which even the most advanced language models (including GPT4) fail to solve using text only.
Authors: Niloofar Mireshghallah, Maria Antoniak, Yash More, Yejin Choi, Golnoosh Farnadi
Abstract: Measuring personal disclosures made in human-chatbot interactions can provide a better understanding of users' AI literacy and facilitate privacy research for large language models (LLMs). We run an extensive, fine-grained analysis on the personal disclosures made by real users to commercial GPT models, investigating the leakage of personally identifiable and sensitive information. To understand the contexts in which users disclose to chatbots, we develop a taxonomy of tasks and sensitive topics, based on qualitative and quantitative analysis of naturally occurring conversations. We discuss these potential privacy harms and observe that: (1) personally identifiable information (PII) appears in unexpected contexts such as in translation or code editing (48% and 16% of the time, respectively) and (2) PII detection alone is insufficient to capture the sensitive topics that are common in human-chatbot interactions, such as detailed sexual preferences or specific drug use habits. We believe that these high disclosure rates are of significant importance for researchers and data curators, and we call for the design of appropriate nudging mechanisms to help users moderate their interactions.
Authors: Yuan Feng, Junlin Lv, Yukun Cao, Xike Xie, S. Kevin Zhou
Abstract: Large Language Models have excelled in various fields but encounter efficiency limitations due to the substantial Key-Value (KV) cache required for long-sequence inference. Recent efforts try to evict non-critical cache elements during runtime, thereby reducing cache size within given memory budgets while preserving generation quality. Our reexamination of foundational principles reveals that prevailing methods aim to minimize an upper bound of eviction loss, quantified as the L1 distance between the pre- and post-eviction outputs of multi-head self-attention mechanisms. Moreover, our analysis indicates that the common practices of uniformly assigning budgets across different attention heads during cache eviction hinder their budget utilization, negatively impacting generation quality. In light of these findings, we propose a simple yet effective adaptive budget allocation algorithm. This algorithm not only optimizes the loss upper bound in theory but also reduces the eviction loss in practice by aligning with the intrinsic patterns of self-attention mechanisms. Integrating this algorithm into two advanced methods, we develop Ada-SnapKV and Ada-Pyramid. Extensive evaluations on 16 datasets and the Needle-in-a-Haystack test confirm that they both significantly boost performance across various tasks.
Authors: Anum Afzal, Ribin Chalumattu, Florian Matthes, Laura Mascarell
Abstract: Despite the advances in the abstractive summarization task using Large Language Models (LLM), there is a lack of research that asses their abilities to easily adapt to different domains. We evaluate the domain adaptation abilities of a wide range of LLMs on the summarization task across various domains in both fine-tuning and in-context learning settings. We also present AdaptEval, the first domain adaptation evaluation suite. AdaptEval includes a domain benchmark and a set of metrics to facilitate the analysis of domain adaptation. Our results demonstrate that LLMs exhibit comparable performance in the in-context learning setting, regardless of their parameter scale.
Authors: Yinheng Li, Rogerio Bonatti, Sara Abdali, Justin Wagle, Kazuhito Koishida
Abstract: Using Large Language Models (LLMs) to generate synthetic data for model training has become increasingly popular in recent years. While LLMs are capable of producing realistic training data, the effectiveness of data generation is influenced by various factors, including the choice of prompt, task complexity, and the quality, quantity, and diversity of the generated data. In this work, we focus exclusively on using synthetic data for text classification tasks. Specifically, we use natural language understanding (NLU) models trained on synthetic data to assess the quality of synthetic data from different generation approaches. This work provides an empirical analysis of the impact of these factors and offers recommendations for better data generation practices.
Authors: Aniket Deroy, Kripabandhu Ghosh, Saptarshi Ghosh
Abstract: Automatic summarization of legal case judgements, which are known to be long and complex, has traditionally been tried via extractive summarization models. In recent years, generative models including abstractive summarization models and Large language models (LLMs) have gained huge popularity. In this paper, we explore the applicability of such models for legal case judgement summarization. We applied various domain specific abstractive summarization models and general domain LLMs as well as extractive summarization models over two sets of legal case judgements from the United Kingdom (UK) Supreme Court and the Indian (IN) Supreme Court and evaluated the quality of the generated summaries. We also perform experiments on a third dataset of legal documents of a different type, Government reports from the United States (US). Results show that abstractive summarization models and LLMs generally perform better than the extractive methods as per traditional metrics for evaluating summary quality. However, detailed investigation shows the presence of inconsistencies and hallucinations in the outputs of the generative models, and we explore ways to reduce the hallucinations and inconsistencies in the summaries. Overall, the investigation suggests that further improvements are needed to enhance the reliability of abstractive models and LLMs for legal case judgement summarization. At present, a human-in-the-loop technique is more suitable for performing manual checks to identify inconsistencies in the generated summaries.
Authors: Irene Siragusa, Roberto Pirrone
Abstract: This paper illustrates the architecture and training of Unipa-GPT, a chatbot relying on a Large Language Model, developed for assisting students in choosing a bachelor/master degree course at the University of Palermo. Unipa-GPT relies on gpt-3.5-turbo, it was presented in the context of the European Researchers' Night (SHARPER night). In our experiments we adopted both the Retrieval Augmented Generation (RAG) approach and fine-tuning to develop the system. The whole architecture of Unipa-GPT is presented, both the RAG and the fine-tuned systems are compared, and a brief discussion on their performance is reported. Further comparison with other Large Language Models and the experimental results during the SHARPER night are illustrated.
Authors: Verena Schoenmueller, Simon J. Blanchard, Gita V. Johar
Abstract: We propose that social-media users' own post histories are an underused yet valuable resource for studying fake-news sharing. By extracting textual cues from their prior posts, and contrasting their prevalence against random social-media users and others (e.g., those with similar socio-demographics, political news-sharers, and fact-check sharers), researchers can identify cues that distinguish fake-news sharers, predict those most likely to share fake news, and identify promising constructs to build interventions. Our research includes studies along these lines. In Study 1, we explore the distinctive language patterns of fake-news sharers, highlighting elements such as their higher use of anger and power-related words. In Study 2, we show that adding textual cues into predictive models enhances their accuracy in predicting fake-news sharers. In Study 3, we explore the contrasting role of trait and situational anger, and show trait anger is associated with a greater propensity to share both true and fake news. In Study 4, we introduce a way to authenticate Twitter accounts in surveys, before using it to explore how crafting an ad copy that resonates with users' sense of power encourages the adoption of fact-checking tools. We hope to encourage the use of novel research methods for marketers and misinformation researchers.
Authors: Florinel-Alin Croitoru, Nicolae-Catalin Ristea, Radu Tudor Ionescu, Nicu Sebe
Abstract: Most curriculum learning methods require an approach to sort the data samples by difficulty, which is often cumbersome to perform. In this work, we propose a novel curriculum learning approach termed Learning Rate Curriculum (LeRaC), which leverages the use of a different learning rate for each layer of a neural network to create a data-agnostic curriculum during the initial training epochs. More specifically, LeRaC assigns higher learning rates to neural layers closer to the input, gradually decreasing the learning rates as the layers are placed farther away from the input. The learning rates increase at various paces during the first training iterations, until they all reach the same value. From this point on, the neural model is trained as usual. This creates a model-level curriculum learning strategy that does not require sorting the examples by difficulty and is compatible with any neural network, generating higher performance levels regardless of the architecture. We conduct comprehensive experiments on 12 data sets from the computer vision (CIFAR-10, CIFAR-100, Tiny ImageNet, ImageNet-200, Food-101, UTKFace, PASCAL VOC), language (BoolQ, QNLI, RTE) and audio (ESC-50, CREMA-D) domains, considering various convolutional (ResNet-18, Wide-ResNet-50, DenseNet-121, YOLOv5), recurrent (LSTM) and transformer (CvT, BERT, SepTr) architectures. We compare our approach with the conventional training regime, as well as with Curriculum by Smoothing (CBS), a state-of-the-art data-agnostic curriculum learning approach. Unlike CBS, our performance improvements over the standard training regime are consistent across all data sets and models. Furthermore, we significantly surpass CBS in terms of training time (there is no additional cost over the standard training regime for LeRaC). Our code is freely available at: https://github.com/CroitoruAlin/LeRaC.
Authors: Van Bach Nguyen, J\"org Schl\"otterer, Christin Seifert
Abstract: The goal of Explainable AI (XAI) is to design methods to provide insights into the reasoning process of black-box models, such as deep neural networks, in order to explain them to humans. Social science research states that such explanations should be conversational, similar to human-to-human explanations. In this work, we show how to incorporate XAI in a conversational agent, using a standard design for the agent comprising natural language understanding and generation components. We build upon an XAI question bank, which we extend by quality-controlled paraphrases, to understand the user's information needs. We further systematically survey the literature for suitable explanation methods that provide the information to answer those questions, and present a comprehensive list of suggestions. Our work is the first step towards truly natural conversations about machine learning models with an explanation agent. The comprehensive list of XAI questions and the corresponding explanation methods may support other researchers in providing the necessary information to address users' demands. To facilitate future work, we release our source code and data.
Authors: Yue Wu, Yewen Fan, Paul Pu Liang, Amos Azaria, Yuanzhi Li, Tom M. Mitchell
Abstract: High sample complexity has long been a challenge for RL. On the other hand, humans learn to perform tasks not only from interaction or demonstrations, but also by reading unstructured text documents, e.g., instruction manuals. Instruction manuals and wiki pages are among the most abundant data that could inform agents of valuable features and policies or task-specific environmental dynamics and reward structures. Therefore, we hypothesize that the ability to utilize human-written instruction manuals to assist learning policies for specific tasks should lead to a more efficient and better-performing agent. We propose the Read and Reward framework. Read and Reward speeds up RL algorithms on Atari games by reading manuals released by the Atari game developers. Our framework consists of a QA Extraction module that extracts and summarizes relevant information from the manual and a Reasoning module that evaluates object-agent interactions based on information from the manual. An auxiliary reward is then provided to a standard A2C RL agent, when interaction is detected. Experimentally, various RL algorithms obtain significant improvement in performance and training speed when assisted by our design.
Authors: Jaemin Cho, Linjie Li, Zhengyuan Yang, Zhe Gan, Lijuan Wang, Mohit Bansal
Abstract: Spatial control is a core capability in controllable image generation. Advancements in layout-guided image generation have shown promising results on in-distribution (ID) datasets with similar spatial configurations. However, it is unclear how these models perform when facing out-of-distribution (OOD) samples with arbitrary, unseen layouts. In this paper, we propose LayoutBench, a diagnostic benchmark for layout-guided image generation that examines four categories of spatial control skills: number, position, size, and shape. We benchmark two recent representative layout-guided image generation methods and observe that the good ID layout control may not generalize well to arbitrary layouts in the wild (e.g., objects at the boundary). Next, we propose IterInpaint, a new baseline that generates foreground and background regions step-by-step via inpainting, demonstrating stronger generalizability than existing models on OOD layouts in LayoutBench. We perform quantitative and qualitative evaluation and fine-grained analysis on the four LayoutBench skills to pinpoint the weaknesses of existing models. We show comprehensive ablation studies on IterInpaint, including training task ratio, crop&paste vs. repaint, and generation order. Lastly, we evaluate the zero-shot performance of different pretrained layout-guided image generation models on LayoutBench-COCO, our new benchmark for OOD layouts with real objects, where our IterInpaint consistently outperforms SOTA baselines in all four splits. Project website: https://layoutbench.github.io
Authors: Sam van der Poel, Dakotah Lambert, Kalina Kostyszyn, Tiantian Gao, Rahul Verma, Derek Andersen, Joanne Chau, Emily Peterson, Cody St. Clair, Paul Fodor, Chihiro Shibata, Jeffrey Heinz
Abstract: Synthetic datasets constructed from formal languages allow fine-grained examination of the learning and generalization capabilities of machine learning systems for sequence classification. This article presents a new benchmark for machine learning systems on sequence classification called MLRegTest, which contains training, development, and test sets from 1,800 regular languages. Different kinds of formal languages represent different kinds of long-distance dependencies, and correctly identifying long-distance dependencies in sequences is a known challenge for ML systems to generalize successfully. MLRegTest organizes its languages according to their logical complexity (monadic second order, first order, propositional, or monomial expressions) and the kind of logical literals (string, tier-string, subsequence, or combinations thereof). The logical complexity and choice of literal provides a systematic way to understand different kinds of long-distance dependencies in regular languages, and therefore to understand the capacities of different ML systems to learn such long-distance dependencies. Finally, the performance of different neural networks (simple RNN, LSTM, GRU, transformer) on MLRegTest is examined. The main conclusion is that performance depends significantly on the kind of test set, the class of language, and the neural network architecture.
Authors: Siyu Zhang, Yeming Chen, Yaoru Sun, Fang Wang, Jun Yang, Lizhi Bai, Shangce Gao
Abstract: The key to integrating visual language tasks is to establish a good alignment strategy. Recently, visual semantic representation has achieved fine-grained visual understanding by dividing grids or image patches. However, the coarse-grained semantic interactions in image space should not be ignored, which hinders the extraction of complex contextual semantic relations at the scene boundaries. This paper proposes superpixels as comprehensive and robust visual primitives, which mine coarse-grained semantic interactions by clustering perceptually similar pixels, speeding up the subsequent processing of primitives. To capture superpixel-level semantic features, we propose a Multiscale Difference Graph Convolutional Network (MDGCN). It allows parsing the entire image as a fine-to-coarse visual hierarchy. To reason actual semantic relations, we reduce potential noise interference by aggregating difference information between adjacent graph nodes. Finally, we propose a multi-level fusion rule in a bottom-up manner to avoid understanding deviation by mining complementary spatial information at different levels. Experiments show that the proposed method can effectively promote the learning of multiple downstream tasks. Encouragingly, our method outperforms previous methods on all metrics. Our code will be released upon publication.
Authors: Junu Kim, Chaeeun Shim, Bosco Seong Kyu Yang, Chami Im, Sung Yoon Lim, Han-Gil Jeong, Edward Choi
Abstract: Machine learning (ML) has recently shown promising results in medical predictions using electronic health records (EHRs). However, since ML models typically have a limited capability in terms of input sizes, selecting specific medical events from EHRs for use as input is necessary. This selection process, often relying on expert opinion, can cause bottlenecks in development. We propose Retrieval-Enhanced Medical prediction model (REMed) to address such challenges. REMed can essentially evaluate unlimited medical events, select the relevant ones, and make predictions. This allows for an unrestricted input size, eliminating the need for manual event selection. We verified these properties through experiments involving 27 clinical prediction tasks across four independent cohorts, where REMed outperformed the baselines. Notably, we found that the preferences of REMed align closely with those of medical experts. We expect our approach to significantly expedite the development of EHR prediction models by minimizing clinicians' need for manual involvement.
Authors: Hanlin Zhang, Benjamin L. Edelman, Danilo Francati, Daniele Venturi, Giuseppe Ateniese, Boaz Barak
Abstract: Watermarking generative models consists of planting a statistical signal (watermark) in a model's output so that it can be later verified that the output was generated by the given model. A strong watermarking scheme satisfies the property that a computationally bounded attacker cannot erase the watermark without causing significant quality degradation. In this paper, we study the (im)possibility of strong watermarking schemes. We prove that, under well-specified and natural assumptions, strong watermarking is impossible to achieve. This holds even in the private detection algorithm setting, where the watermark insertion and detection algorithms share a secret key, unknown to the attacker. To prove this result, we introduce a generic efficient watermark attack; the attacker is not required to know the private key of the scheme or even which scheme is used. Our attack is based on two assumptions: (1) The attacker has access to a "quality oracle" that can evaluate whether a candidate output is a high-quality response to a prompt, and (2) The attacker has access to a "perturbation oracle" which can modify an output with a nontrivial probability of maintaining quality, and which induces an efficiently mixing random walk on high-quality outputs. We argue that both assumptions can be satisfied in practice by an attacker with weaker computational capabilities than the watermarked model itself, to which the attacker has only black-box access. Furthermore, our assumptions will likely only be easier to satisfy over time as models grow in capabilities and modalities. We demonstrate the feasibility of our attack by instantiating it to attack three existing watermarking schemes for large language models: Kirchenbauer et al. (2023), Kuditipudi et al. (2023), and Zhao et al. (2023). The same attack successfully removes the watermarks planted by all three schemes, with only minor quality degradation.
Authors: Anil Batra, Davide Moltisanti, Laura Sevilla-Lara, Marcus Rohrbach, Frank Keller
Abstract: Procedural videos, exemplified by recipe demonstrations, are instrumental in conveying step-by-step instructions. However, understanding such videos is challenging as it involves the precise localization of steps and the generation of textual instructions. Manually annotating steps and writing instructions is costly, which limits the size of current datasets and hinders effective learning. Leveraging large but noisy video-transcript datasets for pre-training can boost performance but demands significant computational resources. Furthermore, transcripts contain irrelevant content and differ in style from human-written instructions. To mitigate these issues, we propose a novel technique, Sieve-&-Swap, to automatically generate high-quality training data for the recipe domain: (i) Sieve: filters irrelevant transcripts and (ii) Swap: acquires high-quality text by replacing transcripts with human-written instruction from a text-only recipe dataset. The resulting dataset is three orders of magnitude smaller than current web-scale datasets but enables efficient training of large-scale models. Alongside Sieve-&-Swap, we propose Procedure Transformer (ProcX), a model for end-to-end step localization and instruction generation for procedural videos. When pre-trained on our curated dataset, this model achieves state-of-the-art performance on YouCook2 and Tasty while using a fraction of the training data. We have released code and dataset.
Authors: Yunlong Tang, Jing Bi, Siting Xu, Luchuan Song, Susan Liang, Teng Wang, Daoan Zhang, Jie An, Jingyang Lin, Rongyi Zhu, Ali Vosoughi, Chao Huang, Zeliang Zhang, Pinxin Liu, Mingqian Feng, Feng Zheng, Jianguo Zhang, Ping Luo, Jiebo Luo, Chenliang Xu
Abstract: With the burgeoning growth of online video platforms and the escalating volume of video content, the demand for proficient video understanding tools has intensified markedly. Given the remarkable capabilities of large language models (LLMs) in language and multimodal tasks, this survey provides a detailed overview of recent advancements in video understanding that harness the power of LLMs (Vid-LLMs). The emergent capabilities of Vid-LLMs are surprisingly advanced, particularly their ability for open-ended multi-granularity (general, temporal, and spatiotemporal) reasoning combined with commonsense knowledge, suggesting a promising path for future video understanding. We examine the unique characteristics and capabilities of Vid-LLMs, categorizing the approaches into three main types: Video Analyzer x LLM, Video Embedder x LLM, and (Analyzer + Embedder) x LLM. Furthermore, we identify five sub-types based on the functions of LLMs in Vid-LLMs: LLM as Summarizer, LLM as Manager, LLM as Text Decoder, LLM as Regressor, and LLM as Hidden Layer. Furthermore, this survey presents a comprehensive study of the tasks, datasets, benchmarks, and evaluation methodologies for Vid-LLMs. Additionally, it explores the expansive applications of Vid-LLMs across various domains, highlighting their remarkable scalability and versatility in real-world video understanding challenges. Finally, it summarizes the limitations of existing Vid-LLMs and outlines directions for future research. For more information, readers are recommended to visit the repository at https://github.com/yunlong10/Awesome-LLMs-for-Video-Understanding.
URLs: https://github.com/yunlong10/Awesome-LLMs-for-Video-Understanding.
Authors: Jiawen Kang, Lingwei Meng, Mingyu Cui, Haohan Guo, Xixin Wu, Xunying Liu, Helen Meng
Abstract: End-to-end multi-talker speech recognition has garnered great interest as an effective approach to directly transcribe overlapped speech from multiple speakers. Current methods typically adopt either 1) single-input multiple-output (SIMO) models with a branched encoder, or 2) single-input single-output (SISO) models based on attention-based encoder-decoder architecture with serialized output training (SOT). In this work, we propose a Cross-Speaker Encoding (CSE) network to address the limitations of SIMO models by aggregating cross-speaker representations. Furthermore, the CSE model is integrated with SOT to leverage both the advantages of SIMO and SISO while mitigating their drawbacks. To the best of our knowledge, this work represents an early effort to integrate SIMO and SISO for multi-talker speech recognition. Experiments on the two-speaker LibrispeechMix dataset show that the CES model reduces word error rate (WER) by 8% over the SIMO baseline. The CSE-SOT model reduces WER by 10% overall and by 16% on high-overlap speech compared to the SOT model. Code is available at https://github.com/kjw11/CSEnet-ASR.
Authors: Yuchen Zeng, Wonjun Kang, Yicong Chen, Hyung Il Koo, Kangwook Lee
Abstract: The evolution from Large Language Models (LLMs) to Multimodal Large Language Models (MLLMs) has spurred research into extending In-Context Learning (ICL) to its multimodal counterpart. Existing such studies have primarily concentrated on image-to-text ICL. However, the Text-to-Image ICL (T2I-ICL), with its unique characteristics and potential applications, remains underexplored. To address this gap, we formally define the task of T2I-ICL and present CoBSAT, the first T2I-ICL benchmark dataset, encompassing ten tasks. Utilizing our dataset to benchmark six state-of-the-art MLLMs, we uncover considerable difficulties MLLMs encounter in solving T2I-ICL. We identify the primary challenges as the inherent complexity of multimodality and image generation, and show that strategies such as fine-tuning and Chain-of-Thought prompting help to mitigate these difficulties, leading to notable improvements in performance. Our code and dataset are available at https://github.com/UW-Madison-Lee-Lab/CoBSAT.
Authors: Raghav Kapoor, Yash Parag Butala, Melisa Russak, Jing Yu Koh, Kiran Kamble, Waseem Alshikh, Ruslan Salakhutdinov
Abstract: For decades, human-computer interaction has fundamentally been manual. Even today, almost all productive work done on the computer necessitates human input at every step. Autonomous virtual agents represent an exciting step in automating many of these menial tasks. Virtual agents would empower users with limited technical proficiency to harness the full possibilities of computer systems. They could also enable the efficient streamlining of numerous computer tasks, ranging from calendar management to complex travel bookings, with minimal human intervention. In this paper, we introduce OmniACT, the first-of-a-kind dataset and benchmark for assessing an agent's capability to generate executable programs to accomplish computer tasks. Our scope extends beyond traditional web automation, covering a diverse range of desktop applications. The dataset consists of fundamental tasks such as "Play the next song", as well as longer horizon tasks such as "Send an email to John Doe mentioning the time and place to meet". Specifically, given a pair of screen image and a visually-grounded natural language task, the goal is to generate a script capable of fully executing the task. We run several strong baseline language model agents on our benchmark. The strongest baseline, GPT-4, performs the best on our benchmark However, its performance level still reaches only 15% of the human proficiency in generating executable scripts capable of completing the task, demonstrating the challenge of our task for conventional web agents. Our benchmark provides a platform to measure and evaluate the progress of language model agents in automating computer tasks and motivates future work towards building multimodal models that bridge large language models and the visual grounding of computer screens.
Authors: Hang Hua, Jing Shi, Kushal Kafle, Simon Jenni, Daoan Zhang, John Collomosse, Scott Cohen, Jiebo Luo
Abstract: Recent progress in large-scale pre-training has led to the development of advanced vision-language models (VLMs) with remarkable proficiency in comprehending and generating multimodal content. Despite the impressive ability to perform complex reasoning for VLMs, current models often struggle to effectively and precisely capture the compositional information on both the image and text sides. To address this, we propose FineMatch, a new aspect-based fine-grained text and image matching benchmark, focusing on text and image mismatch detection and correction. This benchmark introduces a novel task for boosting and evaluating the VLMs' compositionality for aspect-based fine-grained text and image matching. In this task, models are required to identify mismatched aspect phrases within a caption, determine the aspect's class, and propose corrections for an image-text pair that may contain between 0 and 3 mismatches. To evaluate the models' performance on this new task, we propose a new evaluation metric named ITM-IoU for which our experiments show a high correlation to human evaluation. In addition, we also provide a comprehensive experimental analysis of existing mainstream VLMs, including fully supervised learning and in-context learning settings. We have found that models trained on FineMatch demonstrate enhanced proficiency in detecting fine-grained text and image mismatches. Moreover, models (e.g., GPT-4V, Gemini Pro Vision) with strong abilities to perform multimodal in-context learning are not as skilled at fine-grained compositional image and text matching analysis. With FineMatch, we are able to build a system for text-to-image generation hallucination detection and correction.
Authors: Henry Peng Zou, Vinay Samuel, Yue Zhou, Weizhi Zhang, Liancheng Fang, Zihe Song, Philip S. Yu, Cornelia Caragea
Abstract: Existing datasets for attribute value extraction (AVE) predominantly focus on explicit attribute values while neglecting the implicit ones, lack product images, are often not publicly available, and lack an in-depth human inspection across diverse domains. To address these limitations, we present ImplicitAVE, the first, publicly available multimodal dataset for implicit attribute value extraction. ImplicitAVE, sourced from the MAVE dataset, is carefully curated and expanded to include implicit AVE and multimodality, resulting in a refined dataset of 68k training and 1.6k testing data across five domains. We also explore the application of multimodal large language models (MLLMs) to implicit AVE, establishing a comprehensive benchmark for MLLMs on the ImplicitAVE dataset. Six recent MLLMs with eleven variants are evaluated across diverse settings, revealing that implicit value extraction remains a challenging task for MLLMs. The contributions of this work include the development and release of ImplicitAVE, and the exploration and benchmarking of various MLLMs for implicit AVE, providing valuable insights and potential future research directions. Dataset and code are available at https://github.com/HenryPengZou/ImplicitAVE
Authors: Han Zhong, Guhao Feng, Wei Xiong, Xinle Cheng, Li Zhao, Di He, Jiang Bian, Liwei Wang
Abstract: In the classical Reinforcement Learning from Human Feedback (RLHF) framework, Proximal Policy Optimization (PPO) is employed to learn from sparse, sentence-level rewards -- a challenging scenario in traditional deep reinforcement learning. Despite the great successes of PPO in the alignment of state-of-the-art closed-source large language models (LLMs), its open-source implementation is still largely sub-optimal, as widely reported by numerous research studies. To address these issues, we introduce a framework that models RLHF problems as a Markov decision process (MDP), enabling the capture of fine-grained token-wise information. Furthermore, we provide theoretical insights that demonstrate the superiority of our MDP framework over the previous sentence-level bandit formulation. Under this framework, we introduce an algorithm, dubbed as Reinforced Token Optimization (\texttt{RTO}), which learns the token-wise reward function from preference data and performs policy optimization based on this learned token-wise reward signal. Theoretically, \texttt{RTO} is proven to have the capability of finding the near-optimal policy sample-efficiently. For its practical implementation, \texttt{RTO} innovatively integrates Direct Preference Optimization (DPO) and PPO. DPO, originally derived from sparse sentence rewards, surprisingly provides us with a token-wise characterization of response quality, which is seamlessly incorporated into our subsequent PPO training stage. Extensive real-world alignment experiments verify the effectiveness of the proposed approach.
Authors: Ruihao Gong, Yang Yong, Shiqiao Gu, Yushi Huang, Chentao Lv, Yunchen Zhang, Xianglong Liu, Dacheng Tao
Abstract: Recent advancements in large language models (LLMs) are propelling us toward artificial general intelligence with their remarkable emergent abilities and reasoning capabilities. However, the substantial computational and memory requirements limit the widespread adoption. Quantization, a key compression technique, can effectively mitigate these demands by compressing and accelerating LLMs, albeit with potential risks to accuracy. Numerous studies have aimed to minimize the accuracy loss associated with quantization. However, their quantization configurations vary from each other and cannot be fairly compared. In this paper, we present LLMC, a plug-and-play compression toolkit, to fairly and systematically explore the impact of quantization. LLMC integrates dozens of algorithms, models, and hardwares, offering high extensibility from integer to floating-point quantization, from LLM to vision-language (VLM) model, from fixed-bit to mixed precision, and from quantization to sparsification. Powered by this versatile toolkit, our benchmark covers three key aspects: calibration data, algorithms (three strategies), and data formats, providing novel insights and detailed analyses for further research and practical guidance for users. Our toolkit is available at \href{LLMC}{https://github.com/ModelTC/llmc}.
Authors: Leyi Pan, Aiwei Liu, Zhiwei He, Zitian Gao, Xuandong Zhao, Yijian Lu, Binglin Zhou, Shuliang Liu, Hanlin Zhang, Xuming Hu, Lijie Wen, Irwin King
Abstract: LLM watermarking, which embeds imperceptible yet algorithmically detectable signals in model outputs to identify LLM-generated text, has become crucial in mitigating the potential misuse of large language models. However, the abundance of LLM watermarking algorithms, their intricate mechanisms, and the complex evaluation procedures and perspectives pose challenges for researchers and the community to easily experiment with, understand, and assess the latest advancements. To address these issues, we introduce MarkLLM, an open-source toolkit for LLM watermarking. MarkLLM offers a unified and extensible framework for implementing LLM watermarking algorithms, while providing user-friendly interfaces to ensure ease of access. Furthermore, it enhances understanding by supporting automatic visualization of the underlying mechanisms of these algorithms. For evaluation, MarkLLM offers a comprehensive suite of 12 tools spanning three perspectives, along with two types of automated evaluation pipelines. Through MarkLLM, we aim to support researchers while improving the comprehension and involvement of the general public in LLM watermarking technology, fostering consensus and driving further advancements in research and application. Our code is available at https://github.com/THU-BPM/MarkLLM.
Authors: Xuyang Ge, Fukang Zhu, Wentao Shu, Junxuan Wang, Zhengfu He, Xipeng Qiu
Abstract: Circuit analysis of any certain model behavior is a central task in mechanistic interpretability. We introduce our circuit discovery pipeline with Sparse Autoencoders (SAEs) and a variant called Transcoders. With these two modules inserted into the model, the model's computation graph with respect to OV and MLP circuits becomes strictly linear. Our methods do not require linear approximation to compute the causal effect of each node. This fine-grained graph identifies both end-to-end and local circuits accounting for either logits or intermediate features. We can scalably apply this pipeline with a technique called Hierarchical Attribution. We analyze three kinds of circuits in GPT-2 Small: bracket, induction, and Indirect Object Identification circuits. Our results reveal new findings underlying existing discoveries.
Authors: Jianliang He, Siyu Chen, Fengzhuo Zhang, Zhuoran Yang
Abstract: In this work, from a theoretical lens, we aim to understand why large language model (LLM) empowered agents are able to solve decision-making problems in the physical world. To this end, consider a hierarchical reinforcement learning (RL) model where the LLM Planner and the Actor perform high-level task planning and low-level execution, respectively. Under this model, the LLM Planner navigates a partially observable Markov decision process (POMDP) by iteratively generating language-based subgoals via prompting. Under proper assumptions on the pretraining data, we prove that the pretrained LLM Planner effectively performs Bayesian aggregated imitation learning (BAIL) through in-context learning. Additionally, we highlight the necessity for exploration beyond the subgoals derived from BAIL by proving that naively executing the subgoals returned by LLM leads to a linear regret. As a remedy, we introduce an $\epsilon$-greedy exploration strategy to BAIL, which is proven to incur sublinear regret when the pretraining error is small. Finally, we extend our theoretical framework to include scenarios where the LLM Planner serves as a world model for inferring the transition model of the environment and to multi-agent settings, enabling coordination among multiple Actors.
Authors: Khai Le-Duc, David Thulke, Hung-Phong Tran, Long Vo-Dang, Khai-Nguyen Nguyen, Truong-Son Hy, Ralf Schl\"uter
Abstract: Spoken Named Entity Recognition (NER) aims to extracting named entities from speech and categorizing them into types like person, location, organization, etc. In this work, we present VietMed-NER - the first spoken NER dataset in the medical domain. To our best knowledge, our real-world dataset is the largest spoken NER dataset in the world in terms of the number of entity types, featuring 18 distinct types. Secondly, we present baseline results using various state-of-the-art pre-trained models: encoder-only and sequence-to-sequence. We found that pre-trained multilingual models XLM-R outperformed all monolingual models on both reference text and ASR output. Also in general, encoders perform better than sequence-to-sequence models for the NER task. By simply translating, the transcript is applicable not just to Vietnamese but to other languages as well. All code, data and models are made publicly available here: https://github.com/leduckhai/MultiMed
Authors: Shwai He, Guoheng Sun, Zheyu Shen, Ang Li
Abstract: Scaling Transformer-based large language models (LLMs) has demonstrated promising performance across various tasks. However, it also introduces redundant structures, posing challenges for real-world deployment. Despite some recognition of redundancy in LLMs, the variability of redundancy across different modules, such as MLP and Attention layers, is under-explored. In this work, we investigate the varying redundancy across different modules within Transformers, including Blocks, MLP, and Attention layers, using a similarity-based metric. This metric operates on the premise that redundant structures produce outputs highly similar to their inputs. Surprisingly, while attention layers are essential for transformers and distinguish them from other mainstream architectures, we found that a large proportion of attention layers exhibit excessively high similarity and can be safely pruned without degrading performance, leading to reduced memory and computation costs. Additionally, we further propose a method that jointly drops Attention and MLP layers, achieving improved performance and dropping ratios. Extensive experiments demonstrate the effectiveness of our methods, e.g., Llama-3-70B maintains comparable performance even after pruning half of the attention layers. Our findings provide valuable insights for future network architecture design. The code is released at: \url{https://github.com/Shwai-He/LLM-Drop}.
Authors: Isaac Ong, Amjad Almahairi, Vincent Wu, Wei-Lin Chiang, Tianhao Wu, Joseph E. Gonzalez, M Waleed Kadous, Ion Stoica
Abstract: Large language models (LLMs) exhibit impressive capabilities across a wide range of tasks, yet the choice of which model to use often involves a trade-off between performance and cost. More powerful models, though effective, come with higher expenses, while less capable models are more cost-effective. To address this dilemma, we propose several efficient router models that dynamically select between a stronger and a weaker LLM during inference, aiming to optimize the balance between cost and response quality. We develop a training framework for these routers leveraging human preference data and data augmentation techniques to enhance performance. Our evaluation on widely-recognized benchmarks shows that our approach significantly reduces costs-by over 2 times in certain cases-without compromising the quality of responses. Interestingly, our router models also demonstrate significant transfer learning capabilities, maintaining their performance even when the strong and weak models are changed at test time. This highlights the potential of these routers to provide a cost-effective yet high-performance solution for deploying LLMs.
Authors: Thong Nguyen, Yi Bin, Xiaobao Wu, Xinshuai Dong, Zhiyuan Hu, Khoi Le, Cong-Duy Nguyen, See-Kiong Ng, Luu Anh Tuan
Abstract: Data quality stands at the forefront of deciding the effectiveness of video-language representation learning. However, video-text pairs in previous data typically do not align perfectly with each other, which might lead to video-language representations that do not accurately reflect cross-modal semantics. Moreover, previous data also possess an uneven distribution of concepts, thereby hampering the downstream performance across unpopular subjects. To address these problems, we propose a contrastive objective with a subtractive angular margin to regularize cross-modal representations in their effort to reach perfect similarity. Furthermore, to adapt to the non-uniform concept distribution, we propose a multi-layer perceptron (MLP)-parameterized weighting function that maps loss values to sample weights which enable dynamic adjustment of the model's focus throughout the training. With the training guided by a small amount of unbiased meta-data and augmented by video-text data generated by large vision-language model, we improve video-language representations and achieve superior performances on commonly used video question answering and text-video retrieval datasets.
Authors: Nuo Chen, Yang Deng, Jia Li
Abstract: This survey explores the burgeoning field of role-playing with language models, focusing on their development from early persona-based models to advanced character-driven simulations facilitated by Large Language Models (LLMs). Initially confined to simple persona consistency due to limited model capabilities, role-playing tasks have now expanded to embrace complex character portrayals involving character consistency, behavioral alignment, and overall attractiveness. We provide a comprehensive taxonomy of the critical components in designing these systems, including data, models and alignment, agent architecture and evaluation. This survey not only outlines the current methodologies and challenges, such as managing dynamic personal profiles and achieving high-level persona consistency but also suggests avenues for future research in improving the depth and realism of role-playing applications. The goal is to guide future research by offering a structured overview of current methodologies and identifying potential areas for improvement. Related resources and papers are available at https://github.com/nuochenpku/Awesome-Role-Play-Papers.
URLs: https://github.com/nuochenpku/Awesome-Role-Play-Papers.
Authors: Christoph Minixhofer, Ond\v{r}ej Klejch, Peter Bell
Abstract: Many recently published Text-to-Speech (TTS) systems produce audio close to real speech. However, TTS evaluation needs to be revisited to make sense of the results obtained with the new architectures, approaches and datasets. We propose evaluating the quality of synthetic speech as a combination of multiple factors such as prosody, speaker identity, and intelligibility. Our approach assesses how well synthetic speech mirrors real speech by obtaining correlates of each factor and measuring their distance from both real speech datasets and noise datasets. We benchmark 35 TTS systems developed between 2008 and 2024 and show that our score computed as an unweighted average of factors strongly correlates with the human evaluations from each time period.
Authors: Yiluo Wei, Yiming Zhu, Pan Hui, Gareth Tyson
Abstract: The rise of generative AI is transforming the landscape of digital imagery, and exerting a significant influence on online creative communities. This has led to the emergence of AI-Generated Content (AIGC) social platforms, such as Civitai. These distinctive social platforms allow users to build and share their own generative AI models, thereby enhancing the potential for more diverse artistic expression. Designed in the vein of social networks, they also provide artists with the means to showcase their creations (generated from the models), engage in discussions, and obtain feedback, thus nurturing a sense of community. Yet, this openness also raises concerns about the abuse of such platforms, e.g., using models to disseminate deceptive deepfakes or infringe upon copyrights. To explore this, we conduct the first comprehensive empirical study of an AIGC social platform, focusing on its use for generating abusive content. As an exemplar, we construct a comprehensive dataset covering Civitai, the largest available AIGC social platform. Based on this dataset of 87K models and 2M images, we explore the characteristics of content and discuss strategies for moderation to better govern these platforms.
Authors: Audrey Huang, Wenhao Zhan, Tengyang Xie, Jason D. Lee, Wen Sun, Akshay Krishnamurthy, Dylan J. Foster
Abstract: Language model alignment methods, such as reinforcement learning from human feedback (RLHF), have led to impressive advances in language model capabilities, but existing techniques are limited by a widely observed phenomenon known as overoptimization, where the quality of the language model plateaus or degrades over the course of the alignment process. Overoptimization is often attributed to overfitting to an inaccurate reward model, and while it can be mitigated through online data collection, this is infeasible in many settings. This raises a fundamental question: Do existing offline alignment algorithms make the most of the data they have, or can their sample-efficiency be improved further? We address this question with a new algorithm for offline alignment, $\chi^2$-Preference Optimization ($\chi$PO). $\chi$PO is a one-line change to Direct Preference Optimization (DPO; Rafailov et al., 2023), which only involves modifying the logarithmic link function in the DPO objective. Despite this minimal change, $\chi$PO implicitly implements the principle of pessimism in the face of uncertainty via regularization with the $\chi^2$-divergence -- which quantifies uncertainty more effectively than KL-regularization -- and provably alleviates overoptimization, achieving sample-complexity guarantees based on single-policy concentrability -- the gold standard in offline reinforcement learning. $\chi$PO's simplicity and strong guarantees make it the first practical and general-purpose offline alignment algorithm that is provably robust to overoptimization.
Authors: Hossein Naderi, Alireza Shojaei, Lifu Huang
Abstract: Automating activities through robots in unstructured environments, such as construction sites, has been a long-standing desire. However, the high degree of unpredictable events in these settings has resulted in far less adoption compared to more structured settings, such as manufacturing, where robots can be hard-coded or trained on narrowly defined datasets. Recently, pretrained foundation models, such as Large Language Models (LLMs), have demonstrated superior generalization capabilities by providing zero-shot solutions for problems do not present in the training data, proposing them as a potential solution for introducing robots to unstructured environments. To this end, this study investigates potential opportunities and challenges of pretrained foundation models from a multi-dimensional perspective. The study systematically reviews application of foundation models in two field of robotic and unstructured environment and then synthesized them with deliberative acting theory. Findings showed that linguistic capabilities of LLMs have been utilized more than other features for improving perception in human-robot interactions. On the other hand, findings showed that the use of LLMs demonstrated more applications in project management and safety in construction, and natural hazard detection in disaster management. Synthesizing these findings, we located the current state-of-the-art in this field on a five-level scale of automation, placing them at conditional automation. This assessment was then used to envision future scenarios, challenges, and solutions toward autonomous safe unstructured environments. Our study can be seen as a benchmark to track our progress toward that future.