Authors: Nikolay Plotnikov, Alexander Antonov
Abstract: In this paper, we introduce four comprehensive datasets for the Chuvash language, aiming to support and enhance linguistic research and technological development for this underrepresented language. These datasets include a monolingual dataset, a parallel dataset with Russian, a parallel dataset with English, and an audio dataset. Each dataset is meticulously curated to serve various applications such as machine translation, linguistic analysis, and speech recognition, providing valuable resources for scholars and developers working with the Chuvash language. Together, these datasets represent a significant step towards preserving and promoting the Chuvash language in the digital age.
Authors: Shafiuddin Rehan Ahmed, Zhiyong Eric Wang, George Arthur Baker, Kevin Stowe, James H. Martin
Abstract: The most popular Cross-Document Event Coreference Resolution (CDEC) datasets fail to convey the true difficulty of the task, due to the lack of lexical diversity between coreferring event triggers (words or phrases that refer to an event). Furthermore, there is a dearth of event datasets for figurative language, limiting a crucial avenue of research in event comprehension. We address these two issues by introducing ECB+META, a lexically rich variant of Event Coref Bank Plus (ECB+) for CDEC on symbolic and metaphoric language. We use ChatGPT as a tool for the metaphoric transformation of sentences in the documents of ECB+, then tag the original event triggers in the transformed sentences in a semi-automated manner. In this way, we avoid the re-annotation of expensive coreference links. We present results that show existing methods that work well on ECB+ struggle with ECB+META, thereby paving the way for CDEC research on a much more challenging dataset. Code/data: https://github.com/ahmeshaf/llms_coref
Authors: Santiago Arena, Antonio Quintero-Rinc\'on
Abstract: Specific Language Impairment (SLI) is a disorder that affects communication and can affect both comprehension and expression. This study focuses on effectively detecting SLI in children using transcripts of spontaneous narratives from 1063 interviews. A three-stage cascading pipeline was proposed f. In the first stage, feature extraction and dimensionality reduction of the data are performed using the Random Forest (RF) and Spearman correlation methods. In the second stage, the most predictive variables from the first stage are estimated using logistic regression, which is used in the last stage to detect SLI in children from transcripts of spontaneous narratives using a nearest neighbor classifier. The results revealed an accuracy of 97.13% in identifying SLI, highlighting aspects such as the length of the responses, the quality of their utterances, and the complexity of the language. This new approach, framed in natural language processing, offers significant benefits to the field of SLI detection by avoiding complex subjective variables and focusing on quantitative metrics directly related to the child's performance.
Authors: Hilda Hadan, Derrick Wang, Reza Hadi Mogavi, Joseph Tu, Leah Zhang-Kennedy, Lennart E. Nacke
Abstract: Generative AI (GenAI) use in research writing is growing fast. However, it is unclear how peer reviewers recognize or misjudge AI-augmented manuscripts. To investigate the impact of AI-augmented writing on peer reviews, we conducted a snippet-based online survey with 17 peer reviewers from top-tier HCI conferences. Our findings indicate that while AI-augmented writing improves readability, language diversity, and informativeness, it often lacks research details and reflective insights from authors. Reviewers consistently struggled to distinguish between human and AI-augmented writing but their judgements remained consistent. They noted the loss of a "human touch" and subjective expressions in AI-augmented writing. Based on our findings, we advocate for reviewer guidelines that promote impartial evaluations of submissions, regardless of any personal biases towards GenAI. The quality of the research itself should remain a priority in reviews, regardless of any preconceived notions about the tools used to create it. We emphasize that researchers must maintain their authorship and control over the writing process, even when using GenAI's assistance.
Authors: Jisoo Mok, Mohammad Kachuee, Shuyang Dai, Shayan Ray, Tara Taghavi, Sungroh Yoon
Abstract: Task-orientated conversational agents interact with users and assist them via leveraging external APIs. A typical task-oriented conversational system can be broken down into three phases: external API selection, argument filling, and response generation. The focus of our work is the task of argument filling, which is in charge of accurately providing arguments required by the selected API. Upon comprehending the dialogue history and the pre-defined API schema, the argument filling task is expected to provide the external API with the necessary information to generate a desirable agent action. In this paper, we study the application of Large Language Models (LLMs) for the problem of API argument filling task. Our initial investigation reveals that LLMs require an additional grounding process to successfully perform argument filling, inspiring us to design training and prompting frameworks to ground their responses. Our experimental results demonstrate that when paired with proposed techniques, the argument filling performance of LLMs noticeably improves, paving a new way toward building an automated argument filling framework.
Authors: Bernadette J Tix
Abstract: This study investigates the impact of Large Language Models generating follow up questions in response to user requests for short text documents. Users provided prompts requesting documents they would like the AI to produce. The AI then generated questions to clarify the user needs before generating the requested documents. Users answered the questions and then indicated their preference between a document generated using both the initial prompt and the questions and answers, and a document generated using only the initial prompt, and gave feedback about their experience with the question-answering process. The findings of this study show clear benefits to question-asking both in document preference and in the qualitative user experience.
Authors: Sadar Jaf, Basel Barakat
Abstract: Despite the extensive communication benefits offered by social media platforms, numerous challenges must be addressed to ensure user safety. One of the most significant risks faced by users on these platforms is targeted hate speech. Social media platforms are widely utilised for generating datasets employed in training and evaluating machine learning algorithms for hate speech detection. However, existing public datasets exhibit numerous limitations, hindering the effective training of these algorithms and leading to inaccurate hate speech classification. This study provides a comprehensive empirical evaluation of several public datasets commonly used in automated hate speech classification. Through rigorous analysis, we present compelling evidence highlighting the limitations of current hate speech datasets. Additionally, we conduct a range of statistical analyses to elucidate the strengths and weaknesses inherent in these datasets. This work aims to advance the development of more accurate and reliable machine learning models for hate speech detection by addressing the dataset limitations identified.
Authors: Shezheng Song, Shasha Li, Jie Yu, Shan Zhao, Xiaopeng Li, Jun Ma, Xiaodong Liu, Zhuo Li, Xiaoguang Mao
Abstract: Our study delves into Multimodal Entity Linking, aligning the mention in multimodal information with entities in knowledge base. Existing methods are still facing challenges like ambiguous entity representations and limited image information utilization. Thus, we propose dynamic entity extraction using ChatGPT, which dynamically extracts entities and enhances datasets. We also propose a method: Dynamically Integrate Multimodal information with knowledge base (DIM), employing the capability of the Large Language Model (LLM) for visual understanding. The LLM, such as BLIP-2, extracts information relevant to entities in the image, which can facilitate improved extraction of entity features and linking them with the dynamic entity representations provided by ChatGPT. The experiments demonstrate that our proposed DIM method outperforms the majority of existing methods on the three original datasets, and achieves state-of-the-art (SOTA) on the dynamically enhanced datasets (Wiki+, Rich+, Diverse+). For reproducibility, our code and collected datasets are released on \url{https://github.com/season1blue/DIM}.
Authors: Aditya Makkar, Divya Makkar, Aarav Patel, Liam Hebert
Abstract: The lack of fluency in sign language remains a barrier to seamless communication for hearing and speech-impaired communities. In this work, we propose a low-cost, real-time ASL-to-speech translation glove and an exhaustive training dataset of sign language patterns. We then benchmarked this dataset with supervised learning models, such as LSTMs, GRUs and Transformers, where our best model achieved 92% accuracy. The SignSpeak dataset has 7200 samples encompassing 36 classes (A-Z, 1-10) and aims to capture realistic signing patterns by using five low-cost flex sensors to measure finger positions at each time step at 36 Hz. Our open-source dataset, models and glove designs, provide an accurate and efficient ASL translator while maintaining cost-effectiveness, establishing a framework for future work to build on.
Authors: Xukun Liu, Bowen Lei, Ruqi Zhang, Dongkuan Xu
Abstract: Large language model (LLM) decoding involves generating a sequence of tokens based on a given context, where each token is predicted one at a time using the model's learned probabilities. The typical autoregressive decoding method requires a separate forward pass through the model for each token generated, which is computationally inefficient and poses challenges for deploying LLMs in latency-sensitive scenarios. The main limitations of current decoding methods stem from their inefficiencies and resource demands. Existing approaches either necessitate fine-tuning smaller models, which is resource-intensive, or rely on fixed retrieval schemes to construct drafts for the next tokens, which lack adaptability and fail to generalize across different models and contexts. To address these issues, we introduce a novel methodology called ADED, which accelerates LLM decoding without requiring fine-tuning. Our approach involves an adaptive draft-verification process that evolves over time to improve efficiency. We utilize a tri-gram matrix-based LLM representation to dynamically approximate the output distribution of the LLM, allowing the model to adjust to changing token probabilities during the decoding process. Additionally, we implement a draft construction mechanism that effectively balances exploration and exploitation, ensuring that the drafts generated are both diverse and close to the true output distribution of the LLM. The importance of this design lies in its ability to optimize the draft distribution adaptively, leading to faster and more accurate decoding. Through extensive experiments on various benchmark datasets and LLM architectures, we demonstrate that ADED significantly accelerates the decoding process while maintaining high accuracy, making it suitable for deployment in a wide range of practical applications.
Authors: Peiyang Wu, Nan Guo, Xiao Xiao, Wenming Li, Xiaochun Ye, Dongrui Fan
Abstract: Recently, large language models (LLMs) have demonstrated excellent performance in understanding human instructions and generating code, which has inspired researchers to explore the feasibility of generating RTL code with LLMs. However, the existing approaches to fine-tune LLMs on RTL codes typically are conducted on fixed datasets, which do not fully stimulate the capability of LLMs and require large amounts of reference data. To mitigate these issues , we introduce a simple yet effective iterative training paradigm named ITERTL. During each iteration, samples are drawn from the model trained in the previous cycle. Then these new samples are employed for training in this loop. Through this iterative approach, the distribution mismatch between the model and the training samples is reduced. Additionally, the model is thus enabled to explore a broader generative space and receive more comprehensive feedback. Theoretical analyses are conducted to investigate the mechanism of the effectiveness. Experimental results show the model trained through our proposed approach can compete with and even outperform the state-of-the-art (SOTA) open-source model with nearly 37\% reference samples, achieving remarkable 42.9\% and 62.2\% pass@1 rate on two VerilogEval evaluation datasets respectively. While using the same amount of reference samples, our method can achieved a relative improvement of 16.9\% and 12.5\% in pass@1 compared to the non-iterative method. This study facilitates the application of LLMs for generating RTL code in practical scenarios with limited data.
Authors: Zhong-Zhi Li, Ming-Liang Zhang, Fei Yin, Zhi-Long Ji, Jin-Feng Bai, Zhen-Ru Pan, Fan-Hu Zeng, Jian Xu, Jia-Xin Zhang, Cheng-Lin Liu
Abstract: Due to the rapid advancements in multimodal large language models, evaluating their multimodal mathematical capabilities continues to receive wide attention. Despite the datasets like MathVista proposed benchmarks for assessing mathematical capabilities in multimodal scenarios, there is still a lack of corresponding evaluation tools and datasets for fine-grained assessment in the context of K12 education in Chinese language. To systematically evaluate the capability of multimodal large models in solving Chinese multimodal mathematical problems, we propose a Chinese Multi-modal Math Skill Evaluation Benchmark, named CMMaTH, contraining 23k multimodal K12 math related questions, forming the largest Chinese multimodal mathematical problem benchmark to date. CMMaTH questions from elementary to high school levels, provide increased diversity in problem types, solution objectives, visual elements, detailed knowledge points, and standard solution annotations. We have constructed an open-source tool GradeGPT integrated with the CMMaTH dataset, facilitating stable, rapid, and cost-free model evaluation. Our data and code are available.
Authors: Rakesh M. Verma, Nachum Dershowitz
Abstract: We show the fraught side of the academic publishing realm and illustrate it through a recent case study with an NLP journal.
Authors: Dimitrios C. Gklezakos, Timothy Misiak, Diamond Bishop
Abstract: From organizing recorded videos and meetings into chapters, to breaking down large inputs in order to fit them into the context window of commoditized Large Language Models (LLMs), topic segmentation of large transcripts emerges as a task of increasing significance. Still, accurate segmentation presents many challenges, including (a) the noisy nature of the Automatic Speech Recognition (ASR) software typically used to obtain the transcripts, (b) the lack of diverse labeled data and (c) the difficulty in pin-pointing the ground-truth number of segments. In this work we present TreeSeg, an approach that combines off-the-shelf embedding models with divisive clustering, to generate hierarchical, structured segmentations of transcripts in the form of binary trees. Our approach is robust to noise and can handle large transcripts efficiently. We evaluate TreeSeg on the ICSI and AMI corpora, demonstrating that it outperforms all baselines. Finally, we introduce TinyRec, a small-scale corpus of manually annotated transcripts, obtained from self-recorded video sessions.
Authors: Timothy Nguyen
Abstract: Transformer based large-language models (LLMs) display extreme proficiency with language yet a precise understanding of how they work remains elusive. One way of demystifying transformer predictions would be to describe how they depend on their context in terms of simple template functions. This paper takes a first step in this direction by considering families of functions (i.e. rules) formed out of simple N-gram based statistics of the training data. By studying how well these rulesets approximate transformer predictions, we obtain a variety of novel discoveries: a simple method to detect overfitting during training without using a holdout set, a quantitative measure of how transformers progress from learning simple to more complex statistical rules over the course of training, a model-variance criterion governing when transformer predictions tend to be described by N-gram rules, and insights into how well transformers can be approximated by N-gram rulesets in the limit where these rulesets become increasingly complex. In this latter direction, we find that for 78% of LLM next-token distributions on TinyStories, their top-1 predictions agree with those provided by our N-gram rulesets.
Authors: Giorgio Roffo
Abstract: This tutorial explores the advancements and challenges in the development of Large Language Models (LLMs) such as ChatGPT and Gemini. It addresses inherent limitations like temporal knowledge cutoffs, mathematical inaccuracies, and the generation of incorrect information, proposing solutions like Retrieval Augmented Generation (RAG), Program-Aided Language Models (PAL), and frameworks such as ReAct and LangChain. The integration of these techniques enhances LLM performance and reliability, especially in multi-step reasoning and complex task execution. The paper also covers fine-tuning strategies, including instruction fine-tuning, parameter-efficient methods like LoRA, and Reinforcement Learning from Human Feedback (RLHF) as well as Reinforced Self-Training (ReST). Additionally, it provides a comprehensive survey of transformer architectures and training techniques for LLMs. The toolbox for implementing these techniques is publicly available at https://github.com/giorgioroffo/large_language_models_open_suite
URLs: https://github.com/giorgioroffo/large_language_models_open_suite
Authors: Faeze Brahman, Sachin Kumar, Vidhisha Balachandran, Pradeep Dasigi, Valentina Pyatkin, Abhilasha Ravichander, Sarah Wiegreffe, Nouha Dziri, Khyathi Chandu, Jack Hessel, Yulia Tsvetkov, Noah A. Smith, Yejin Choi, Hannaneh Hajishirzi
Abstract: Chat-based language models are designed to be helpful, yet they should not comply with every user request. While most existing work primarily focuses on refusal of "unsafe" queries, we posit that the scope of noncompliance should be broadened. We introduce a comprehensive taxonomy of contextual noncompliance describing when and how models should not comply with user requests. Our taxonomy spans a wide range of categories including incomplete, unsupported, indeterminate, and humanizing requests (in addition to unsafe requests). To test noncompliance capabilities of language models, we use this taxonomy to develop a new evaluation suite of 1000 noncompliance prompts. We find that most existing models show significantly high compliance rates in certain previously understudied categories with models like GPT-4 incorrectly complying with as many as 30% of requests. To address these gaps, we explore different training strategies using a synthetically-generated training set of requests and expected noncompliant responses. Our experiments demonstrate that while direct finetuning of instruction-tuned models can lead to both over-refusal and a decline in general capabilities, using parameter efficient methods like low rank adapters helps to strike a good balance between appropriate noncompliance and other capabilities.
Authors: Tengfei Xue, Xuefeng Li, Roman Smirnov, Tahir Azim, Arash Sadrieh, Babak Pahlavan
Abstract: Retrieval-augmented generation (RAG) techniques are widely used today to retrieve and present information in a conversational format. This paper presents a set of enhancements to traditional RAG techniques, focusing on large language models (LLMs) fine-tuned and hosted on AWS Trainium and Inferentia2 AI chips via SageMaker. These chips are characterized by their elasticity, affordability, and efficient performance for AI compute tasks. Besides enabling deployment on these chips, this work aims to improve tool usage, add citation capabilities, and mitigate the risks of hallucinations and unsafe responses due to context bias. We benchmark our RAG system's performance on the Natural Questions and HotPotQA datasets, achieving an accuracy of 62% and 59% respectively, exceeding other models such as DBRX and Mixtral Instruct.
Authors: Daniel Goldstein, Fares Obeid, Eric Alcaide, Guangyu Song, Eugene Cheah
Abstract: We introduce GoldFinch, a hybrid Linear Attention/Transformer sequence model that uses a new technique to efficiently generate a highly compressed and reusable KV-Cache in linear time and space with respect to sequence length. GoldFinch stacks our new GOLD transformer on top of an enhanced version of the Finch (RWKV-6) architecture. We train up to 1.5B parameter class models of the Finch, Llama, and GoldFinch architectures, and find dramatically improved modeling performance relative to both Finch and Llama. Our cache size savings increase linearly with model layer count, ranging from 756-2550 times smaller than the traditional transformer cache for common sizes, enabling inference of extremely large context lengths even on limited hardware. Although autoregressive generation has O(n) time complexity per token because of attention, pre-fill computation of the entire initial cache state for a submitted context costs only O(1) time per token due to the use of a recurrent neural network (RNN) to generate this cache. We release our trained weights and training code under the Apache 2.0 license for community use.
Authors: Minh Nguyen, Franck Dernoncourt, Seunghyun Yoon, Hanieh Deilamsalehy, Hao Tan, Ryan Rossi, Quan Hung Tran, Trung Bui, Thien Huu Nguyen
Abstract: We introduce an approach to identifying speaker names in dialogue transcripts, a crucial task for enhancing content accessibility and searchability in digital media archives. Despite the advancements in speech recognition, the task of text-based speaker identification (SpeakerID) has received limited attention, lacking large-scale, diverse datasets for effective model training. Addressing these gaps, we present a novel, large-scale dataset derived from the MediaSum corpus, encompassing transcripts from a wide range of media sources. We propose novel transformer-based models tailored for SpeakerID, leveraging contextual cues within dialogues to accurately attribute speaker names. Through extensive experiments, our best model achieves a great precision of 80.3\%, setting a new benchmark for SpeakerID. The data and code are publicly available here: \url{https://github.com/adobe-research/speaker-identification}
URLs: https://github.com/adobe-research/speaker-identification
Authors: Marc Pickett, Jeremy Hartman, Ayan Kumar Bhowmick, Raquib-ul Alam, Aditya Vempaty
Abstract: A common way to extend the memory of large language models (LLMs) is by retrieval augmented generation (RAG), which inserts text retrieved from a larger memory into an LLM's context window. However, the context window is typically limited to several thousand tokens, which limits the number of retrieved passages that can inform a model's response. For this reason, it's important to avoid occupying context window space with redundant information by ensuring a degree of diversity among retrieved passages. At the same time, the information should also be relevant to the current task. Most prior methods that encourage diversity among retrieved results, such as Maximal Marginal Relevance (MMR), do so by incorporating an objective that explicitly trades off diversity and relevance. We propose a novel simple optimization metric based on relevant information gain, a probabilistic measure of the total information relevant to a query for a set of retrieved results. By optimizing this metric, diversity organically emerges from our system. When used as a drop-in replacement for the retrieval component of a RAG system, this method yields state-of-the-art performance on question answering tasks from the Retrieval Augmented Generation Benchmark (RGB), outperforming existing metrics that directly optimize for relevance and diversity.
Authors: Bunyamin Keles, Murat Gunay, Serdar I. Caglar
Abstract: Machine translation is indispensable in healthcare for enabling the global dissemination of medical knowledge across languages. However, complex medical terminology poses unique challenges to achieving adequate translation quality and accuracy. This study introduces a novel "LLMs-in-the-loop" approach to develop supervised neural machine translation models optimized specifically for medical texts. While large language models (LLMs) have demonstrated powerful capabilities, this research shows that small, specialized models trained on high-quality in-domain (mostly synthetic) data can outperform even vastly larger LLMs. Custom parallel corpora in six languages were compiled from scientific articles, synthetically generated clinical documents, and medical texts. Our LLMs-in-the-loop methodology employs synthetic data generation, rigorous evaluation, and agent orchestration to enhance performance. We developed small medical translation models using the MarianMT base model. We introduce a new medical translation test dataset to standardize evaluation in this domain. Assessed using BLEU, METEOR, ROUGE, and BERT scores on this test set, our MarianMT-based models outperform Google Translate, DeepL, and GPT-4-Turbo. Results demonstrate that our LLMs-in-the-loop approach, combined with fine-tuning high-quality, domain-specific data, enables specialized models to outperform general-purpose and some larger systems. This research, part of a broader series on expert small models, paves the way for future healthcare-related AI developments, including deidentification and bio-medical entity extraction models. Our study underscores the potential of tailored neural translation models and the LLMs-in-the-loop methodology to advance the field through improved data generation, evaluation, agent, and modeling techniques.
Authors: Hubert Plisiecki, Piotr Koc, Maria Flakus, Artur Pokropek
Abstract: This study explores the use of large language models (LLMs) to predict emotion intensity in Polish political texts, a resource-poor language context. The research compares the performance of several LLMs against a supervised model trained on an annotated corpus of 10,000 social media texts, evaluated for the intensity of emotions by expert judges. The findings indicate that while the supervised model generally outperforms LLMs, offering higher accuracy and lower variance, LLMs present a viable alternative, especially given the high costs associated with data annotation. The study highlights the potential of LLMs in low-resource language settings and underscores the need for further research on emotion intensity prediction and its application across different languages and continuous features. The implications suggest a nuanced decision-making process to choose the right approach to emotion prediction for researchers and practitioners based on resource availability and the specific requirements of their tasks.
Authors: Nicholas Deas, Elsbeth Turcan, Iv\'an P\'erez Mej\'ia, Kathleen McKeown
Abstract: In the field of emotion analysis, much NLP research focuses on identifying a limited number of discrete emotion categories, often applied across languages. These basic sets, however, are rarely designed with textual data in mind, and culture, language, and dialect can influence how particular emotions are interpreted. In this work, we broaden our scope to a practically unbounded set of \textit{affective states}, which includes any terms that humans use to describe their experiences of feeling. We collect and publish MASIVE, a dataset of Reddit posts in English and Spanish containing over 1,000 unique affective states each. We then define the new problem of \textit{affective state identification} for language generation models framed as a masked span prediction task. On this task, we find that smaller finetuned multilingual models outperform much larger LLMs, even on region-specific Spanish affective states. Additionally, we show that pretraining on MASIVE improves model performance on existing emotion benchmarks. Finally, through machine translation experiments, we find that native speaker-written data is vital to good performance on this task.
Authors: Amit Roth, Arnon Turetzky, Yossi Adi
Abstract: We tackle the task of text-to-speech (TTS) in Hebrew. Traditional Hebrew contains Diacritics, which dictate the way individuals should pronounce given words, however, modern Hebrew rarely uses them. The lack of diacritics in modern Hebrew results in readers expected to conclude the correct pronunciation and understand which phonemes to use based on the context. This imposes a fundamental challenge on TTS systems to accurately map between text-to-speech. In this work, we propose to adopt a language modeling Diacritics-Free approach, for the task of Hebrew TTS. The model operates on discrete speech representations and is conditioned on a word-piece tokenizer. We optimize the proposed method using in-the-wild weakly supervised data and compare it to several diacritic-based TTS systems. Results suggest the proposed method is superior to the evaluated baselines considering both content preservation and naturalness of the generated speech. Samples can be found under the following link: pages.cs.huji.ac.il/adiyoss-lab/HebTTS/
Authors: Lauren Levine, Cindy Tung Li, Lydia Bremer-McCollum, Nicholas Wagner, Amir Zeldes
Abstract: Ancient manuscripts are frequently damaged, containing gaps in the text known as lacunae. In this paper, we present a bidirectional RNN model for character prediction of Coptic characters in manuscript lacunae. Our best model performs with 72% accuracy on single character reconstruction, but falls to 37% when reconstructing lacunae of various lengths. While not suitable for definitive manuscript reconstruction, we argue that our RNN model can help scholars rank the likelihood of textual reconstructions. As evidence, we use our RNN model to rank reconstructions in two early Coptic manuscripts. Our investigation shows that neural models can augment traditional methods of textual restoration, providing scholars with an additional tool to assess lacunae in Coptic manuscripts.
Authors: Cathy Jiao, Gary Gao, Chenyan Xiong
Abstract: Data valuation quantifies the value of training data, and is used for data attribution (i.e., determining the contribution of training data towards model predictions), and data selection; both of which are important for curating high-quality datasets to train large language models. In our paper, we show that data valuation through in-context probing (i.e., prompting a LLM) approximates influence functions for selecting training data. We provide a theoretical sketch on this connection based on transformer models performing "implicit" gradient descent on its in-context inputs. Our empirical findings show that in-context probing and gradient-based influence frameworks are similar in how they rank training data. Furthermore, fine-tuning experiments on data selected by either method reveal similar model performance.
Authors: Haoyang Wen, Honglei Zhuang, Hamed Zamani, Alexander Hauptmann, Michael Bendersky
Abstract: Knowledge-intensive visual question answering requires models to effectively use external knowledge to help answer visual questions. A typical pipeline includes a knowledge retriever and an answer generator. However, a retriever that utilizes local information, such as an image patch, may not provide reliable question-candidate relevance scores. Besides, the two-tower architecture also limits the relevance score modeling of a retriever to select top candidates for answer generator reasoning. In this paper, we introduce an additional module, a multi-modal reranker, to improve the ranking quality of knowledge candidates for answer generation. Our reranking module takes multi-modal information from both candidates and questions and performs cross-item interaction for better relevance score modeling. Experiments on OK-VQA and A-OKVQA show that multi-modal reranker from distant supervision provides consistent improvements. We also find a training-testing discrepancy with reranking in answer generation, where performance improves if training knowledge candidates are similar to or noisier than those used in testing.
Authors: Thao Minh Nguyen Phan, Cong-Tinh Dao, Chenwei Wu, Jian-Zhe Wang, Shun Liu, Jun-En Ding, David Restrepo, Feng Liu, Fang-Ming Hung, Wen-Chih Peng
Abstract: Electronic health records (EHRs) are multimodal by nature, consisting of structured tabular features like lab tests and unstructured clinical notes. In real-life clinical practice, doctors use complementary multimodal EHR data sources to get a clearer picture of patients' health and support clinical decision-making. However, most EHR predictive models do not reflect these procedures, as they either focus on a single modality or overlook the inter-modality interactions/redundancy. In this work, we propose MEDFuse, a Multimodal EHR Data Fusion framework that incorporates masked lab-test modeling and large language models (LLMs) to effectively integrate structured and unstructured medical data. MEDFuse leverages multimodal embeddings extracted from two sources: LLMs fine-tuned on free clinical text and masked tabular transformers trained on structured lab test results. We design a disentangled transformer module, optimized by a mutual information loss to 1) decouple modality-specific and modality-shared information and 2) extract useful joint representation from the noise and redundancy present in clinical notes. Through comprehensive validation on the public MIMIC-III dataset and the in-house FEMH dataset, MEDFuse demonstrates great potential in advancing clinical predictions, achieving over 90% F1 score in the 10-disease multi-label classification task.
Authors: Kushan Hewapathirana (Department of Computer Science & Engineering, University of Moratuwa, Sri Lanka, ConscientAI, Sri Lanka), Nisansa de Silva (Department of Computer Science & Engineering, University of Moratuwa, Sri Lanka), C. D. Athuraliya (ConscientAI, Sri Lanka)
Abstract: In the rapidly evolving digital era, there is an increasing demand for concise information as individuals seek to distil key insights from various sources. Recent attention from researchers on Multi-document Summarisation (MDS) has resulted in diverse datasets covering customer reviews, academic papers, medical and legal documents, and news articles. However, the English-centric nature of these datasets has created a conspicuous void for multilingual datasets in today's globalised digital landscape, where linguistic diversity is celebrated. Media platforms such as British Broadcasting Corporation (BBC) have disseminated news in 20+ languages for decades. With only 380 million people speaking English natively as their first language, accounting for less than 5% of the global population, the vast majority primarily relies on other languages. These facts underscore the need for inclusivity in MDS research, utilising resources from diverse languages. Recognising this gap, we present the Multilingual Dataset for Multi-document Summarisation (M2DS), which, to the best of our knowledge, is the first dataset of its kind. It includes document-summary pairs in five languages from BBC articles published during the 2010-2023 period. This paper introduces M2DS, emphasising its unique multilingual aspect, and includes baseline scores from state-of-the-art MDS models evaluated on our dataset.
Authors: Jintang Xue, Yun-Cheng Wang, Chengwei Wei, C. -C. Jay Kuo
Abstract: As a fundamental task in natural language processing, word embedding converts each word into a representation in a vector space. A challenge with word embedding is that as the vocabulary grows, the vector space's dimension increases and it can lead to a vast model size. Storing and processing word vectors are resource-demanding, especially for mobile edge-devices applications. This paper explores word embedding dimension reduction. To balance computational costs and performance, we propose an efficient and effective weakly-supervised feature selection method, named WordFS. It has two variants, each utilizing novel criteria for feature selection. Experiments conducted on various tasks (e.g., word and sentence similarity and binary and multi-class classification) indicate that the proposed WordFS model outperforms other dimension reduction methods at lower computational costs.
Authors: Jie Zhang, Dongrui Liu, Chen Qian, Ziyue Gan, Yong Liu, Yu Qiao, Jing Shao
Abstract: Personality psychologists have analyzed the relationship between personality and safety behaviors in human society. Although Large Language Models (LLMs) demonstrate personality traits, the relationship between personality traits and safety abilities in LLMs still remains a mystery. In this paper, we discover that LLMs' personality traits are closely related to their safety abilities, i.e., toxicity, privacy, and fairness, based on the reliable MBTI-M scale. Meanwhile, the safety alignment generally increases various LLMs' Extraversion, Sensing, and Judging traits. According to such findings, we can edit LLMs' personality traits and improve their safety performance, e.g., inducing personality from ISTJ to ISTP resulted in a relative improvement of approximately 43% and 10% in privacy and fairness performance, respectively. Additionally, we find that LLMs with different personality traits are differentially susceptible to jailbreak. This study pioneers the investigation of LLM safety from a personality perspective, providing new insights into LLM safety enhancement.
Authors: Jeonghyun Park, Hwanhee Lee
Abstract: Conversational search seeks to retrieve relevant passages for the given questions in Conversational QA (ConvQA). Questions in ConvQA face challenges such as omissions and coreferences, making it difficult to obtain desired search results. Conversational Query Reformulation (CQR) transforms these current queries into de-contextualized forms to resolve these issues. However, existing CQR methods focus on rewriting human-friendly queries, which may not always yield optimal search results for the retriever. To overcome this challenge, we introduce GuideCQR, a framework that utilizes guided documents to refine queries, ensuring that they are optimal for retrievers. Specifically, we augment keywords, generate expected answers from the re-ranked documents, and unify them with the filtering process. Experimental results show that queries enhanced by guided documents outperform previous CQR methods. Especially, GuideCQR surpasses the performance of Large Language Model (LLM) prompt-powered approaches and demonstrates the importance of the guided documents in formulating retriever-friendly queries across diverse setups.
Authors: Indranil Bandyopadhyay, Rahul Karmakar
Abstract: Sentiment analysis (SA), is an approach of natural language processing (NLP) for determining a text's emotional tone by analyzing subjective information such as views, feelings, and attitudes toward specific topics, products, services, events, or experiences. This study attempts to develop an advanced deep learning (DL) model for SA to understand global audience emotions through tweets in the context of the Olympic Games. The findings represent global attitudes around the Olympics and contribute to advancing the SA models. We have used NLP for tweet pre-processing and sophisticated DL models for arguing with SA, this research enhances the reliability and accuracy of sentiment classification. The study focuses on data selection, preprocessing, visualization, feature extraction, and model building, featuring a baseline Na\"ive Bayes (NB) model and three advanced DL models: Convolutional Neural Network (CNN), Bidirectional Long Short-Term Memory (BiLSTM), and Bidirectional Encoder Representations from Transformers (BERT). The results of the experiments show that the BERT model can efficiently classify sentiments related to the Olympics, achieving the highest accuracy of 99.23%.
Authors: Houjun Liu, Brian MacWhinney
Abstract: Language development researchers are interested in comparing the process of language learning across languages. Unfortunately, it has been difficult to construct a consistent quantitative framework for such comparisons. However, recent advances in AI (Artificial Intelligence) and ML (Machine Learning) are providing new methods for ASR (automatic speech recognition) and NLP (natural language processing) that can be brought to bear on this problem. Using the Batchalign2 program (Liu et al., 2023), we have been transcribing and linking data for the CHILDES database and have applied the UD (Universal Dependencies) framework to provide a consistent and comparable morphosyntactic analysis for 27 languages. These new resources open possibilities for deeper crosslinguistic study of language learning.
Authors: Zheni Zeng, Jiayi Chen, Huimin Chen, Yukun Yan, Yuxuan Chen, Zhiyuan Liu, Maosong Sun
Abstract: Large language models exhibit aspects of human-level intelligence that catalyze their application as human-like agents in domains such as social simulations, human-machine interactions, and collaborative multi-agent systems. However, the absence of distinct personalities, such as displaying ingratiating behaviors, inconsistent opinions, and uniform response patterns, diminish LLMs utility in practical applications. Addressing this, the development of personality traits in LLMs emerges as a crucial area of research to unlock their latent potential. Existing methods to personify LLMs generally involve strategies like employing stylized training data for instruction tuning or using prompt engineering to simulate different personalities. These methods only capture superficial linguistic styles instead of the core of personalities and are therefore not stable. In this study, we propose PersLLM, integrating psychology-grounded principles of personality: social practice, consistency, and dynamic development, into a comprehensive training methodology. We incorporate personality traits directly into the model parameters, enhancing the model's resistance to induction, promoting consistency, and supporting the dynamic evolution of personality. Single-agent evaluation validates our method's superiority, as it produces responses more aligned with reference personalities compared to other approaches. Case studies for multi-agent communication highlight its benefits in enhancing opinion consistency within individual agents and fostering collaborative creativity among multiple agents in dialogue contexts, potentially benefiting human simulation and multi-agent cooperation. Additionally, human-agent interaction evaluations indicate that our personified models significantly enhance interactive experiences, underscoring the practical implications of our research.
Authors: Arda Y\"uksel, Abdullatif K\"oksal, L\"utfi Kerem \c{S}enel, Anna Korhonen, Hinrich Sch\"utze
Abstract: Multiple choice question answering tasks evaluate the reasoning, comprehension, and mathematical abilities of Large Language Models (LLMs). While existing benchmarks employ automatic translation for multilingual evaluation, this approach is error-prone and potentially introduces culturally biased questions, especially in social sciences. We introduce the first multitask, multiple-choice Turkish QA benchmark, TurkishMMLU, to evaluate LLMs' understanding of the Turkish language. TurkishMMLU includes over 10,000 questions, covering 9 different subjects from Turkish high-school education curricula. These questions are written by curriculum experts, suitable for the high-school curricula in Turkey, covering subjects ranging from natural sciences and math questions to more culturally representative topics such as Turkish Literature and the history of the Turkish Republic. We evaluate over 20 LLMs, including multilingual open-source (e.g., Gemma, Llama, MT5), closed-source (GPT 4o, Claude, Gemini), and Turkish-adapted (e.g., Trendyol) models. We provide an extensive evaluation, including zero-shot and few-shot evaluation of LLMs, chain-of-thought reasoning, and question difficulty analysis along with model performance. We provide an in-depth analysis of the Turkish capabilities and limitations of current LLMs to provide insights for future LLMs for the Turkish language. We publicly release our code for the dataset and evaluation: https://github.com/ArdaYueksel/TurkishMMLU.
Authors: Haisong Gong, Huanhuan Ma, Qiang Liu, Shu Wu, Liang Wang
Abstract: Claim verification is a task that involves assessing the truthfulness of a given claim based on multiple evidence pieces. Using large language models (LLMs) for claim verification is a promising way. However, simply feeding all the evidence pieces to an LLM and asking if the claim is factual does not yield good results. The challenge lies in the noisy nature of both the evidence and the claim: evidence passages typically contain irrelevant information, with the key facts hidden within the context, while claims often convey multiple aspects simultaneously. To navigate this "noisy crowd" of information, we propose EACon (Evidence Abstraction and Claim Deconstruction), a framework designed to find key information within evidence and verify each aspect of a claim separately. EACon first finds keywords from the claim and employs fuzzy matching to select relevant keywords for each raw evidence piece. These keywords serve as a guide to extract and summarize critical information into abstracted evidence. Subsequently, EACon deconstructs the original claim into subclaims, which are then verified against both abstracted and raw evidence individually. We evaluate EACon using two open-source LLMs on two challenging datasets. Results demonstrate that EACon consistently and substantially improve LLMs' performance in claim verification.
Authors: Seyedeh Fatemeh Ebrahimi, Karim Akhavan Azari, Amirmasoud Iravani, Hadi Alizadeh, Zeinab Sadat Taghavi, Hossein Sameti
Abstract: Semantic Textual Relatedness holds significant relevance in Natural Language Processing, finding applications across various domains. Traditionally, approaches to STR have relied on knowledge-based and statistical methods. However, with the emergence of Large Language Models, there has been a paradigm shift, ushering in new methodologies. In this paper, we delve into the investigation of sentence-level STR within Track A (Supervised) by leveraging fine-tuning techniques on the RoBERTa transformer. Our study focuses on assessing the efficacy of this approach across different languages. Notably, our findings indicate promising advancements in STR performance, particularly in Latin languages. Specifically, our results demonstrate notable improvements in English, achieving a correlation of 0.82 and securing a commendable 19th rank. Similarly, in Spanish, we achieved a correlation of 0.67, securing the 15th position. However, our approach encounters challenges in languages like Arabic, where we observed a correlation of only 0.38, resulting in a 20th rank.
Authors: Wanqi Yang, Yunqiu Xu, Yanda Li, Kunze Wang, Binbin Huang, Ling Chen
Abstract: In this study, we explore an emerging research area of Continual Learning for Temporal Sensitive Question Answering (CLTSQA). Previous research has primarily focused on Temporal Sensitive Question Answering (TSQA), often overlooking the unpredictable nature of future events. In real-world applications, it's crucial for models to continually acquire knowledge over time, rather than relying on a static, complete dataset. Our paper investigates strategies that enable models to adapt to the ever-evolving information landscape, thereby addressing the challenges inherent in CLTSQA. To support our research, we first create a novel dataset, divided into five subsets, designed specifically for various stages of continual learning. We then propose a training framework for CLTSQA that integrates temporal memory replay and temporal contrastive learning. Our experimental results highlight two significant insights: First, the CLTSQA task introduces unique challenges for existing models. Second, our proposed framework effectively navigates these challenges, resulting in improved performance.
Authors: Kun Sun, Rong Wang
Abstract: The development of different theories of discourse structure has led to the establishment of discourse corpora based on these theories. However, the existence of discourse corpora established on different theoretical bases creates challenges when it comes to exploring them in a consistent and cohesive way. This study has as its primary focus the conversion of PDTB annotations into dependency structures. It employs refined BERT-based discourse parsers to test the validity of the dependency data derived from the PDTB-style corpora in English, Chinese, and several other languages. By converting both PDTB and RST annotations for the same texts into dependencies, this study also applies ``dependency distance'' metrics to examine the correlation between RST dependencies and PDTB dependencies in English. The results show that the PDTB dependency data is valid and that there is a strong correlation between the two types of dependency distance. This study presents a comprehensive approach for analyzing and evaluating discourse corpora by employing discourse dependencies to achieve unified analysis. By applying dependency representations, we can extract data from PDTB, RST, and SDRT corpora in a coherent and unified manner. Moreover, the cross-linguistic validation establishes the framework's generalizability beyond English. The establishment of this comprehensive dependency framework overcomes limitations of existing discourse corpora, supporting a diverse range of algorithms and facilitating further studies in computational discourse analysis and language sciences.
Authors: Rahul Kumar, Shubham Kakde, Divyansh Rajput, Daud Ibrahim, Rishabh Nahata, Pidathala Sowjanya, Deepak Kumar
Abstract: We present a novel approach to data preparation for developing multilingual Indic large language model. Our meticulous data acquisition spans open-source and proprietary sources, including Common Crawl, Indic books, news articles, and Wikipedia, ensuring a diverse and rich linguistic representation. For each Indic language, we design a custom preprocessing pipeline to effectively eliminate redundant and low-quality text content. Additionally, we perform deduplication on Common Crawl data to address the redundancy present in 70% of the crawled web pages. This study focuses on developing high-quality data, optimizing tokenization for our multilingual dataset for Indic large language models with 3B and 7B parameters, engineered for superior performance in Indic languages. We introduce a novel multilingual tokenizer training strategy, demonstrating our custom-trained Indic tokenizer outperforms the state-of-the-art OpenAI Tiktoken tokenizer, achieving a superior token-to-word ratio for Indic languages.
Authors: Mustafa Dogan, Ilker Kesen, Iacer Calixto, Aykut Erdem, Erkut Erdem
Abstract: The linguistic capabilities of Multimodal Large Language Models (MLLMs) are critical for their effective application across diverse tasks. This study aims to evaluate the performance of MLLMs on the VALSE benchmark, focusing on the efficacy of few-shot In-Context Learning (ICL), and Chain-of-Thought (CoT) prompting. We conducted a comprehensive assessment of state-of-the-art MLLMs, varying in model size and pretraining datasets. The experimental results reveal that ICL and CoT prompting significantly boost model performance, particularly in tasks requiring complex reasoning and contextual understanding. Models pretrained on captioning datasets show superior zero-shot performance, while those trained on interleaved image-text data benefit from few-shot learning. Our findings provide valuable insights into optimizing MLLMs for better grounding of language in visual contexts, highlighting the importance of the composition of pretraining data and the potential of few-shot learning strategies to improve the reasoning abilities of MLLMs.
Authors: Andrea W Wen-Yi, Kathryn Adamson, Nathalie Greenfield, Rachel Goldberg, Sandra Babcock, David Mimno, Allison Koenecke
Abstract: The language used by US courtroom actors in criminal trials has long been studied for biases. However, systematic studies for bias in high-stakes court trials have been difficult, due to the nuanced nature of bias and the legal expertise required. New large language models offer the possibility to automate annotation, saving time and cost. But validating these approaches requires both high quantitative performance as well as an understanding of how automated methods fit in existing workflows, and what they really offer. In this paper we present a case study of adding an automated system to a complex and high-stakes problem: identifying gender-biased language in US capital trials for women defendants. Our team of experienced death-penalty lawyers and NLP technologists pursued a three-phase study: first annotating manually, then training and evaluating computational models, and finally comparing human annotations to model predictions. Unlike many typical NLP tasks, annotating for gender bias in months-long capital trials was a complicated task that involves with many individual judgment calls. In contrast to standard arguments for automation that are based on efficiency and scalability, legal experts found the computational models most useful in challenging their personal bias in annotation and providing opportunities to refine and build consensus on rules for annotation. This suggests that seeking to replace experts with computational models is both unrealistic and undesirable. Rather, computational models offer valuable opportunities to assist the legal experts in annotation-based studies.
Authors: Yunfan Shao, Linyang Li, Yichuan Ma, Peiji Li, Demin Song, Qinyuan Cheng, Shimin Li, Xiaonan Li, Pengyu Wang, Qipeng Guo, Hang Yan, Xipeng Qiu, Xuanjing Huang, Dahua Lin
Abstract: Complex reasoning is an impressive ability shown by large language models (LLMs). Most LLMs are skilled in deductive reasoning, such as chain-of-thought prompting or iterative tool-using to solve challenging tasks step-by-step. In this paper, we hope to focus on evaluating and teaching LLMs to conduct inductive reasoning, that is, LLMs are supposed to infer underlying rules by observing examples or sequential transformations. However, collecting large-scale and diverse human-generated inductive data is challenging. We focus on data synthesis in the code domain and propose a \textbf{Case2Code} task by exploiting the expressiveness and correctness of programs. Specifically, we collect a diverse set of executable programs, synthesize input-output transformations for each program, and force LLMs to infer the underlying code implementations based on the synthetic I/O cases. We first evaluate representative LLMs on the synthesized Case2Code task and demonstrate that the Case-to-code induction is challenging for LLMs. Then, we synthesize large-scale Case2Code training samples to train LLMs to perform inductive reasoning. Experimental results show that such induction training benefits not only in distribution Case2Code performance but also enhances various coding abilities of trained LLMs, demonstrating the great potential of learning inductive reasoning via synthetic data.
Authors: Donghoon Han, Eunhwan Park, Gisang Lee, Adam Lee, Nojun Kwak
Abstract: The rapid expansion of multimedia content has made accurately retrieving relevant videos from large collections increasingly challenging. Recent advancements in text-video retrieval have focused on cross-modal interactions, large-scale foundation model training, and probabilistic modeling, yet often neglect the crucial user perspective, leading to discrepancies between user queries and the content retrieved. To address this, we introduce MERLIN (Multimodal Embedding Refinement via LLM-based Iterative Navigation), a novel, training-free pipeline that leverages Large Language Models (LLMs) for iterative feedback learning. MERLIN refines query embeddings from a user perspective, enhancing alignment between queries and video content through a dynamic question answering process. Experimental results on datasets like MSR-VTT, MSVD, and ActivityNet demonstrate that MERLIN substantially improves Recall@1, outperforming existing systems and confirming the benefits of integrating LLMs into multimodal retrieval systems for more responsive and context-aware multimedia retrieval.
Authors: Fengyu Cai, Xinran Zhao, Hongming Zhang, Iryna Gurevych, Heinz Koeppl
Abstract: Recent advances in measuring hardness-wise properties of data guide language models in sample selection within low-resource scenarios. However, class-specific properties are overlooked for task setup and learning. How will these properties influence model learning and is it generalizable across datasets? To answer this question, this work formally initiates the concept of $\textit{class-wise hardness}$. Experiments across eight natural language understanding (NLU) datasets demonstrate a consistent hardness distribution across learning paradigms, models, and human judgment. Subsequent experiments unveil a notable challenge in measuring such class-wise hardness with instance-level metrics in previous works. To address this, we propose $\textit{GeoHard}$ for class-wise hardness measurement by modeling class geometry in the semantic embedding space. $\textit{GeoHard}$ surpasses instance-level metrics by over 59 percent on $\textit{Pearson}$'s correlation on measuring class-wise hardness. Our analysis theoretically and empirically underscores the generality of $\textit{GeoHard}$ as a fresh perspective on data diagnosis. Additionally, we showcase how understanding class-wise hardness can practically aid in improving task learning.
Authors: Ha Young Kim, Niranjan Balasubramanian, Byungkon Kang
Abstract: It has become common practice now to use random initialization schemes, rather than the pre-trained embeddings, when training transformer based models from scratch. Indeed, we find that pre-trained word embeddings from GloVe, and some sub-word embeddings extracted from language models such as T5 and mT5 fare much worse compared to random initialization. This is counter-intuitive given the well-known representational and transfer-learning advantages of pre-training. Interestingly, we also find that BERT and mBERT embeddings fare better than random initialization, showing the advantages of pre-trained representations. In this work, we posit two potential factors that contribute to these mixed results: the model sensitivity to parameter distribution and the embedding interactions with position encodings. We observe that pre-trained GloVe, T5, and mT5 embeddings have a wider distribution of values. As argued in the initialization studies, such large value initializations can lead to poor training because of saturated outputs. Further, the larger embedding values can, in effect, absorb the smaller position encoding values when added together, thus losing position information. Standardizing the pre-trained embeddings to a narrow range (e.g. as prescribed by Xavier) leads to substantial gains for Glove, T5, and mT5 embeddings. On the other hand, BERT pre-trained embeddings, while larger, are still relatively closer to Xavier initialization range which may allow it to effectively transfer the pre-trained knowledge.
Authors: Xiaoyu Tan, Haoyu Wang, Xihe Qiu, Yuan Cheng, Yinghui Xu, Wei Chu, Yuan Qi
Abstract: Structured data, rich in logical and relational information, has the potential to enhance the reasoning abilities of large language models (LLMs). Still, its integration poses a challenge due to the risk of overwhelming LLMs with excessive tokens and irrelevant context information. To address this, we propose Struct-X, a novel framework that operates through five key phases: ``read-model-fill-reflect-reason'' efficiently enabling LLMs to utilize structured data. It begins by encoding structured data into a topological space using graph embeddings, followed by filling in missing entity information with knowledge retrieval modules, and filtering out irrelevant tokens via a self-supervised module. The final phase involves constructing a topological network with selected tokens to further reduce the total token length for more effective LLM inference. Additionally, Struct-X includes an Auxiliary Module trained to generate prompts, aiding LLMs in analyzing structured data. Extensive experiments on benchmarks, including the knowledge graph question-answer task and the long document reading comprehension task, show that Struct-X notably improves LLM reasoning, demonstrating the effectiveness of structured data augmentation in improving LLM inference with complex input context.
Authors: Ingeol Baek, Jimin Lee, Joonho Yang, Hwanhee Lee
Abstract: Query rewriting aims to generate a new query that can complement the original query to improve the information retrieval system. Recent studies on query rewriting, such as query2doc (Q2D), query2expand (Q2E) and querey2cot (Q2C), rely on the internal knowledge of Large Language Models (LLMs) to generate a relevant passage to add information to the query. Nevertheless, the efficacy of these methodologies may markedly decline in instances where the requisite knowledge is not encapsulated within the model's intrinsic parameters. In this paper, we propose a novel structured query rewriting method called Crafting the Path tailored for retrieval systems. Crafting the Path involves a three-step process that crafts query-related information necessary for finding the passages to be searched in each step. Specifically, the Crafting the Path begins with Query Concept Comprehension, proceeds to Query Type Identification, and finally conducts Expected Answer Extraction. Experimental results show that our method outperforms previous rewriting methods, especially in less familiar domains for LLMs. We demonstrate that our method is less dependent on the internal parameter knowledge of the model and generates queries with fewer factual inaccuracies. Furthermore, we observe that Crafting the Path has less latency compared to the baselines.
Authors: Xihe Qiu, Haoyu Wang, Xiaoyu Tan, Chao Qu, Yujie Xiong, Yuan Cheng, Yinghui Xu, Wei Chu, Yuan Qi
Abstract: Effective collaboration in multi-agent systems requires communicating goals and intentions between agents. Current agent frameworks often suffer from dependencies on single-agent execution and lack robust inter-module communication, frequently leading to suboptimal multi-agent reinforcement learning (MARL) policies and inadequate task coordination. To address these challenges, we present a framework for training large language models (LLMs) as collaborative agents to enable coordinated behaviors in cooperative MARL. Each agent maintains a private intention consisting of its current goal and associated sub-tasks. Agents broadcast their intentions periodically, allowing other agents to infer coordination tasks. A propagation network transforms broadcast intentions into teammate-specific communication messages, sharing relevant goals with designated teammates. The architecture of our framework is structured into planning, grounding, and execution modules. During execution, multiple agents interact in a downstream environment and communicate intentions to enable coordinated behaviors. The grounding module dynamically adapts comprehension strategies based on emerging coordination patterns, while feedback from execution agents influnces the planning module, enabling the dynamic re-planning of sub-tasks. Results in collaborative environment simulation demonstrate intention propagation reduces miscoordination errors by aligning sub-task dependencies between agents. Agents learn when to communicate intentions and which teammates require task details, resulting in emergent coordinated behaviors. This demonstrates the efficacy of intention sharing for cooperative multi-agent RL based on LLMs.
Authors: Ting Jiang, Minghui Song, Zihan Zhang, Haizhen Huang, Weiwei Deng, Feng Sun, Qi Zhang, Deqing Wang, Fuzhen Zhuang
Abstract: Multimodal large language models (MLLMs) have shown promising advancements in general visual and language understanding. However, the representation of multimodal information using MLLMs remains largely unexplored. In this work, we introduce a new framework, E5-V, designed to adapt MLLMs for achieving universal multimodal embeddings. Our findings highlight the significant potential of MLLMs in representing multimodal inputs compared to previous approaches. By leveraging MLLMs with prompts, E5-V effectively bridges the modality gap between different types of inputs, demonstrating strong performance in multimodal embeddings even without fine-tuning. We propose a single modality training approach for E5-V, where the model is trained exclusively on text pairs. This method demonstrates significant improvements over traditional multimodal training on image-text pairs, while reducing training costs by approximately 95%. Additionally, this approach eliminates the need for costly multimodal training data collection. Extensive experiments across four types of tasks demonstrate the effectiveness of E5-V. As a universal multimodal model, E5-V not only achieves but often surpasses state-of-the-art performance in each task, despite being trained on a single modality.
Authors: Claudio Pinhanez, Paulo Cavalin, Luciana Storto, Thomas Fimbow, Alexander Cobbinah, Julio Nogima, Marisa Vasconcelos, Pedro Domingues, Priscila de Souza Mizukami, Nicole Grell, Majo\'i Gongora, Isabel Gon\c{c}alves
Abstract: Since 2022 we have been exploring application areas and technologies in which Artificial Intelligence (AI) and modern Natural Language Processing (NLP), such as Large Language Models (LLMs), can be employed to foster the usage and facilitate the documentation of Indigenous languages which are in danger of disappearing. We start by discussing the decreasing diversity of languages in the world and how working with Indigenous languages poses unique ethical challenges for AI and NLP. To address those challenges, we propose an alternative development AI cycle based on community engagement and usage. Then, we report encouraging results in the development of high-quality machine learning translators for Indigenous languages by fine-tuning state-of-the-art (SOTA) translators with tiny amounts of data and discuss how to avoid some common pitfalls in the process. We also present prototypes we have built in projects done in 2023 and 2024 with Indigenous communities in Brazil, aimed at facilitating writing, and discuss the development of Indigenous Language Models (ILMs) as a replicable and scalable way to create spell-checkers, next-word predictors, and similar tools. Finally, we discuss how we envision a future for language documentation where dying languages are preserved as interactive language models.
Authors: Aman Sinha, Timothee Mickus, Marianne Clausel, Mathieu Constant, Xavier Coubez
Abstract: The success of pretrained language models (PLMs) across a spate of use-cases has led to significant investment from the NLP community towards building domain-specific foundational models. On the other hand, in mission critical settings such as biomedical applications, other aspects also factor in-chief of which is a model's ability to produce reasonable estimates of its own uncertainty. In the present study, we discuss these two desiderata through the lens of how they shape the entropy of a model's output probability distribution. We find that domain specificity and uncertainty awareness can often be successfully combined, but the exact task at hand weighs in much more strongly.
Authors: Chenze Shao, Fandong Meng, Jie Zhou
Abstract: As Large Language Models (LLMs) achieve remarkable progress in language understanding and generation, their training efficiency has become a critical concern. Traditionally, LLMs are trained to predict the next token in a sequence. Despite the success of token-level training, it suffers from considerable computational costs due to the need to process an extensive number of tokens. To mitigate this issue, this paper introduces patch-level training for LLMs, which reduces the sequence length by compressing multiple tokens into a single patch. During patch-level training, we feed the language model shorter sequences of patches and train it to predict the next patch, thereby processing the majority of the training data at a significantly reduced computational cost. Following this, the model continues token-level training on the remaining training data to align with the inference mode. Experiments on a diverse range of models (370M-2.7B parameters) demonstrate that patch-level training can reduce overall computational costs to 0.5$\times$, without compromising the model performance compared to token-level training. Source code: \url{https://github.com/shaochenze/PatchTrain}.
Authors: Youmin Ko, Hyemin Yang, Taeuk Kim, Hyunjoon Kim
Abstract: Fine-tuning pre-trained language models (PLMs) has recently shown a potential to improve knowledge graph completion (KGC). However, most PLM-based methods encode only textual information, neglecting various topological structures of knowledge graphs (KGs). In this paper, we empirically validate the significant relations between the structural properties of KGs and the performance of the PLM-based methods. To leverage the structural knowledge, we propose a Subgraph-Aware Training framework for KGC (SATKGC) that combines (i) subgraph-aware mini-batching to encourage hard negative sampling, and (ii) a new contrastive learning method to focus more on harder entities and harder negative triples in terms of the structural properties. To the best of our knowledge, this is the first study to comprehensively incorporate the structural inductive bias of the subgraphs into fine-tuning PLMs. Extensive experiments on four KGC benchmarks demonstrate the superiority of SATKGC. Our code is available.
Authors: Ben Yao, Yazhou Zhang, Qiuchi Li, Jing Qin
Abstract: Elaborating a series of intermediate reasoning steps significantly improves the ability of large language models (LLMs) to solve complex problems, as such steps would evoke LLMs to think sequentially. However, human sarcasm understanding is often considered an intuitive and holistic cognitive process, in which various linguistic, contextual, and emotional cues are integrated to form a comprehensive understanding of the speaker's true intention, which is argued not be limited to a step-by-step reasoning process. To verify this argument, we introduce a new prompting framework called SarcasmCue, which contains four prompting strategies, $viz.$ chain of contradiction (CoC), graph of cues (GoC), bagging of cues (BoC) and tensor of cues (ToC), which elicits LLMs to detect human sarcasm by considering sequential and non-sequential prompting methods. Through a comprehensive empirical comparison on four benchmarking datasets, we show that the proposed four prompting methods outperforms standard IO prompting, CoT and ToT with a considerable margin, and non-sequential prompting generally outperforms sequential prompting.
Authors: Chris Madge, Massimo Poesio
Abstract: In this work we proposing adapting the Minecraft builder task into an LLM benchmark suitable for evaluating LLM ability in spatially orientated tasks, and informing builder agent design. Previous works have proposed corpora with varying complex structures, and human written instructions. We instead attempt to provide a comprehensive synthetic benchmark for testing builder agents over a series of distinct tasks that comprise of common building operations. We believe this approach allows us to probe specific strengths and weaknesses of different agents, and test the ability of LLMs in the challenging area of spatial reasoning and vector based math.
Authors: Manar Abdelatty, Sherief Reda
Abstract: Hardware design engineers routinely work with multiple Process Design Kits (PDKs) from various fabrication labs, each containing several standard cell libraries, optimized for specific metric such as speed, power, or density. These libraries include multiple views such as liberty files for timing information, LEF files for abstract layout details, and technology LEF for process design rules. Navigating this complex landscape to retrieve specific information about gates or design rules is often time-consuming and error-prone. To address this, we present HDLCopilot, an LLM-powered PDK query system that allows engineers to streamline interactions with PDKs in natural language format, making information retrieval accurate and more efficient. HDLCopilot achieves an accuracy of 94.23\% on an evaluation set comprised of diverse and complex natural language queries. HDLCopilot positions itself as a powerful assistant in the hardware design process, enhancing productivity and reducing potential human errors.
Authors: Kaichen Zhang, Bo Li, Peiyuan Zhang, Fanyi Pu, Joshua Adrian Cahyono, Kairui Hu, Shuai Liu, Yuanhan Zhang, Jingkang Yang, Chunyuan Li, Ziwei Liu
Abstract: The advances of large foundation models necessitate wide-coverage, low-cost, and zero-contamination benchmarks. Despite continuous exploration of language model evaluations, comprehensive studies on the evaluation of Large Multi-modal Models (LMMs) remain limited. In this work, we introduce LMMS-EVAL, a unified and standardized multimodal benchmark framework with over 50 tasks and more than 10 models to promote transparent and reproducible evaluations. Although LMMS-EVAL offers comprehensive coverage, we find it still falls short in achieving low cost and zero contamination. To approach this evaluation trilemma, we further introduce LMMS-EVAL LITE, a pruned evaluation toolkit that emphasizes both coverage and efficiency. Additionally, we present Multimodal LIVEBENCH that utilizes continuously updating news and online forums to assess models' generalization abilities in the wild, featuring a low-cost and zero-contamination evaluation approach. In summary, our work highlights the importance of considering the evaluation trilemma and provides practical solutions to navigate the trade-offs in evaluating large multi-modal models, paving the way for more effective and reliable benchmarking of LMMs. We opensource our codebase and maintain leaderboard of LIVEBENCH at https://github.com/EvolvingLMMs-Lab/lmms-eval and https://huggingface.co/spaces/lmms-lab/LiveBench.
URLs: https://github.com/EvolvingLMMs-Lab/lmms-eval, https://huggingface.co/spaces/lmms-lab/LiveBench.
Authors: Jon McCormack, Elliott Wilson, Nina Rajcic, Maria Teresa Llano
Abstract: This paper presents the design and initial assessment of a novel device that uses generative AI to facilitate creative ideation, inspiration, and reflective thought. Inspired by magnetic poetry, which was originally designed to help overcome writer's block, the device allows participants to compose short poetic texts from a limited vocabulary by physically placing words on the device's surface. Upon composing the text, the system employs a large language model (LLM) to generate a response, displayed on an e-ink screen. We explored various strategies for internally sequencing prompts to foster creative thinking, including analogy, allegorical interpretations, and ideation. We installed the device in our research laboratory for two weeks and held a focus group at the conclusion to evaluate the design. The design choice to limit interactions with the LLM to poetic text, coupled with the tactile experience of assembling the poem, fostered a deeper and more enjoyable engagement with the LLM compared to traditional chatbot or screen-based interactions. This approach gives users the opportunity to reflect on the AI-generated responses in a manner conducive to creative thought.
Authors: Michael McGettrick, Paul McGettrick
Abstract: We estimate the Kolmogorov complexity of melodies in Irish traditional dance music using Lempel-Ziv compression. The "tunes" of the music are presented in so-called "ABC notation" as simply a sequence of letters from an alphabet: We have no rhythmic variation, with all notes being of equal length. Our estimation of algorithmic complexity can be used to distinguish "simple" or "easy" tunes (with more repetition) from "difficult" ones (with less repetition) which should prove useful for students learning tunes. We further present a comparison of two tune categories (reels and jigs) in terms of their complexity.
Authors: Junhyuk Choi, Yeseon Hong, Bugeun Kim
Abstract: Large Language Models (LLMs) have recently been widely adopted on interactive systems requiring communications. As the false belief in a model can harm the usability of such systems, LLMs should not have cognitive biases that humans have. Especially psychologists focused on the False Consensus Effect (FCE), which can distract smooth communication by posing false beliefs. However, previous studies have less examined FCE in LLMs thoroughly, which needs more consideration of confounding biases, general situations, and prompt changes. Therefore, in this paper, we conduct two studies to deeply examine the FCE phenomenon in LLMs. In Study 1, we investigate whether LLMs have FCE. In Study 2, we explore how various prompting styles affect the demonstration of FCE. As a result of these studies, we identified that popular LLMs have FCE. Also, the result specifies the conditions when the strength of FCE becomes larger or smaller compared to normal usage.
Authors: Mladen Popovi\'c, Maruf A. Dhali, Lambert Schomaker, Johannes van der Plicht, Kaare Lund Rasmussen, Jacopo La Nasa, Ilaria Degano, Maria Perla Colombini, Eibert Tigchelaar
Abstract: Determining the chronology of ancient handwritten manuscripts is essential for reconstructing the evolution of ideas. For the Dead Sea Scrolls, this is particularly important. However, there is an almost complete lack of date-bearing manuscripts evenly distributed across the timeline and written in similar scripts available for palaeographic comparison. Here, we present Enoch, a state-of-the-art AI-based date-prediction model, trained on the basis of new radiocarbon-dated samples of the scrolls. Enoch uses established handwriting-style descriptors and applies Bayesian ridge regression. The challenge of this study is that the number of radiocarbon-dated manuscripts is small, while current machine learning requires an abundance of training data. We show that by using combined angular and allographic writing style feature vectors and applying Bayesian ridge regression, Enoch could predict the radiocarbon-based dates from style, supported by leave-one-out validation, with varied MAEs of 27.9 to 30.7 years relative to the radiocarbon dating. Enoch was then used to estimate the dates of 135 unseen manuscripts, revealing that 79 per cent of the samples were considered 'realistic' upon palaeographic post-hoc evaluation. We present a new chronology of the scrolls. The radiocarbon ranges and Enoch's style-based predictions are often older than the traditionally assumed palaeographic estimates. In the range of 300-50 BCE, Enoch's date prediction provides an improved granularity. The study is in line with current developments in multimodal machine-learning techniques, and the methods can be used for date prediction in other partially-dated manuscript collections. This research shows how Enoch's quantitative, probability-based approach can be a tool for palaeographers and historians, re-dating ancient Jewish key texts and contributing to current debates on Jewish and Christian origins.
Authors: Khai Le-Duc, Ryan Zhang, Ngoc Son Nguyen, Tan-Hanh Pham, Anh Dao, Ba Hung Ngo, Anh Totti Nguyen, Truong-Son Hy
Abstract: Vision-language models have been extensively explored across a wide range of tasks, achieving satisfactory performance; however, their application in medical imaging remains underexplored. In this work, we propose a unified framework - LiteGPT - for the medical imaging. We leverage multiple pre-trained visual encoders to enrich information and enhance the performance of vision-language models. To the best of our knowledge, this is the first study to utilize vision-language models for the novel task of joint localization and classification in medical images. Besides, we are pioneers in providing baselines for disease localization in chest X-rays. Finally, we set new state-of-the-art performance in the image classification task on the well-benchmarked VinDr-CXR dataset. All code and models are publicly available online: https://github.com/leduckhai/LiteGPT
Authors: Kareem Amin, Alex Bie, Weiwei Kong, Alexey Kurakin, Natalia Ponomareva, Umar Syed, Andreas Terzis, Sergei Vassilvitskii
Abstract: We present an approach for generating differentially private synthetic text using large language models (LLMs), via private prediction. In the private prediction framework, we only require the output synthetic data to satisfy differential privacy guarantees. This is in contrast to approaches that train a generative model on potentially sensitive user-supplied source data and seek to ensure the model itself is safe to release. We prompt a pretrained LLM with source data, but ensure that next-token predictions are made with differential privacy guarantees. Previous work in this paradigm reported generating a small number of examples (<10) at reasonable privacy levels, an amount of data that is useful only for downstream in-context learning or prompting. In contrast, we make changes that allow us to generate thousands of high-quality synthetic data points, greatly expanding the set of potential applications. Our improvements come from an improved privacy analysis and a better private selection mechanism, which makes use of the equivalence between the softmax layer for sampling tokens in LLMs and the exponential mechanism. Furthermore, we introduce a novel use of public predictions via the sparse vector technique, in which we do not pay privacy costs for tokens that are predictable without sensitive data; we find this to be particularly effective for structured data.
Authors: Yuyang Jiang, Chacha Chen, Dang Nguyen, Benjamin M. Mervak, Chenhao Tan
Abstract: GPT-4V's purported strong multimodal abilities raise interests in using it to automate radiology report writing, but there lacks thorough evaluations. In this work, we perform a systematic evaluation of GPT-4V in generating radiology reports on two chest X-ray report datasets: MIMIC-CXR and IU X-Ray. We attempt to directly generate reports using GPT-4V through different prompting strategies and find that it fails terribly in both lexical metrics and clinical efficacy metrics. To understand the low performance, we decompose the task into two steps: 1) the medical image reasoning step of predicting medical condition labels from images; and 2) the report synthesis step of generating reports from (groundtruth) conditions. We show that GPT-4V's performance in image reasoning is consistently low across different prompts. In fact, the distributions of model-predicted labels remain constant regardless of which groundtruth conditions are present on the image, suggesting that the model is not interpreting chest X-rays meaningfully. Even when given groundtruth conditions in report synthesis, its generated reports are less correct and less natural-sounding than a finetuned LLaMA-2. Altogether, our findings cast doubt on the viability of using GPT-4V in a radiology workflow.
Authors: Gavin Leech, Juan J. Vazquez, Misha Yagudin, Niclas Kupper, Laurence Aitchison
Abstract: Evaluating modern ML models is hard. The strong incentive for researchers and companies to report a state-of-the-art result on some metric often leads to questionable research practices (QRPs): bad practices which fall short of outright research fraud. We describe 43 such practices which can undermine reported results, giving examples where possible. Our list emphasises the evaluation of large language models (LLMs) on public benchmarks. We also discuss "irreproducible research practices", i.e. decisions that make it difficult or impossible for other researchers to reproduce, build on or audit previous research.
Authors: Yufan Shen, Chuwei Luo, Zhaoqing Zhu, Yang Chen, Qi Zheng, Zhi Yu, Jiajun Bu, Cong Yao
Abstract: Recently, large language models (LLMs) and multimodal large language models (MLLMs) have demonstrated promising results on document visual question answering (VQA) task, particularly after training on document instruction datasets. An effective evaluation method for document instruction data is crucial in constructing instruction data with high efficacy, which, in turn, facilitates the training of LLMs and MLLMs for document VQA. However, most existing evaluation methods for instruction data are limited to the textual content of the instructions themselves, thereby hindering the effective assessment of document instruction datasets and constraining their construction. In this paper, we propose ProcTag, a data-oriented method that assesses the efficacy of document instruction data. ProcTag innovatively performs tagging on the execution process of instructions rather than the instruction text itself. By leveraging the diversity and complexity of these tags to assess the efficacy of the given dataset, ProcTag enables selective sampling or filtering of document instructions. Furthermore, DocLayPrompt, a novel semi-structured layout-aware document prompting strategy, is proposed for effectively representing documents. Experiments demonstrate that sampling existing open-sourced and generated document VQA/instruction datasets with ProcTag significantly outperforms current methods for evaluating instruction data. Impressively, with ProcTag-based sampling in the generated document datasets, only 30.5\% of the document instructions are required to achieve 100\% efficacy compared to the complete dataset. The code is publicly available at https://github.com/AlibabaResearch/AdvancedLiterateMachinery/tree/main/DocumentUnderstanding/ProcTag.
Authors: Gengze Zhou, Yicong Hong, Zun Wang, Xin Eric Wang, Qi Wu
Abstract: Capitalizing on the remarkable advancements in Large Language Models (LLMs), there is a burgeoning initiative to harness LLMs for instruction following robotic navigation. Such a trend underscores the potential of LLMs to generalize navigational reasoning and diverse language understanding. However, a significant discrepancy in agent performance is observed when integrating LLMs in the Vision-and-Language navigation (VLN) tasks compared to previous downstream specialist models. Furthermore, the inherent capacity of language to interpret and facilitate communication in agent interactions is often underutilized in these integrations. In this work, we strive to bridge the divide between VLN-specialized models and LLM-based navigation paradigms, while maintaining the interpretative prowess of LLMs in generating linguistic navigational reasoning. By aligning visual content in a frozen LLM, we encompass visual observation comprehension for LLMs and exploit a way to incorporate LLMs and navigation policy networks for effective action predictions and navigational reasoning. We demonstrate the data efficiency of the proposed methods and eliminate the gap between LM-based agents and state-of-the-art VLN specialists.
Authors: Haoyang Gui, Thales Bertaglia, Catalina Goanta, Sybe de Vries, Gerasimos Spanakis
Abstract: Content monetization on social media fuels a growing influencer economy. Influencer marketing remains largely undisclosed or inappropriately disclosed on social media. Non-disclosure issues have become a priority for national and supranational authorities worldwide, who are starting to impose increasingly harsher sanctions on them. This paper proposes a transparent methodology for measuring whether and how influencers comply with disclosures based on legal standards. We introduce a novel distinction between disclosures that are legally sufficient (green) and legally insufficient (yellow). We apply this methodology to an original dataset reflecting the content of 150 Dutch influencers publicly registered with the Dutch Media Authority based on recently introduced registration obligations. The dataset consists of 292,315 posts and is multi-language (English and Dutch) and cross-platform (Instagram, YouTube and TikTok). We find that influencer marketing remains generally underdisclosed on social media, and that bigger influencers are not necessarily more compliant with disclosure standards.
Authors: Wentao Xu
Abstract: Understanding the dynamics of language toxicity on social media is important for us to investigate the propagation of misinformation and the development of echo chambers for political scenarios such as U.S. presidential elections. Recent research has used large-scale data to investigate the dynamics across social media platforms. However, research on the toxicity dynamics is not enough. This study aims to provide a first exploration of the potential language toxicity flow among Left, Right and Center users. Specifically, we aim to examine whether Left users were easier to be attacked by language toxicity. In this study, more than 500M Twitter posts were examined. It was discovered that Left users received much more toxic replies than Right and Center users.
Authors: Angie Boggust, Hyemin Bang, Hendrik Strobelt, Arvind Satyanarayan
Abstract: Abstraction -- the process of generalizing specific examples into broad reusable patterns -- is central to how people efficiently process and store information and apply their knowledge to new data. Promisingly, research has shown that ML models learn representations that span levels of abstraction, from specific concepts like "bolo tie" and "car tire" to more general concepts like "CEO" and "model". However, existing techniques analyze these representations in isolation, treating learned concepts as independent artifacts rather than an interconnected web of abstraction. As a result, although we can identify the concepts a model uses to produce its output, it is difficult to assess if it has learned a human-aligned abstraction of the concepts that will generalize to new data. To address this gap, we introduce abstraction alignment, a methodology to measure the agreement between a model's learned abstraction and the expected human abstraction. We quantify abstraction alignment by comparing model outputs against a human abstraction graph, such as linguistic relationships or medical disease hierarchies. In evaluation tasks interpreting image models, benchmarking language models, and analyzing medical datasets, abstraction alignment provides a deeper understanding of model behavior and dataset content, differentiating errors based on their agreement with human knowledge, expanding the verbosity of current model quality metrics, and revealing ways to improve existing human abstractions.
Authors: William Brannon, Doug Beeferman, Hang Jiang, Andrew Heyward, Deb Roy
Abstract: Understanding and making use of audience feedback is important but difficult for journalists, who now face an impractically large volume of audience comments online. We introduce AudienceView, an online tool to help journalists categorize and interpret this feedback by leveraging large language models (LLMs). AudienceView identifies themes and topics, connects them back to specific comments, provides ways to visualize the sentiment and distribution of the comments, and helps users develop ideas for subsequent reporting projects. We consider how such tools can be useful in a journalist's workflow, and emphasize the importance of contextual awareness and human judgment.
Authors: Chiman Salavati, Shannon Song, Willmar Sosa Diaz, Scott A. Hale, Roberto E. Montenegro, Fabricio Murai, Shiri Dori-Hacohen
Abstract: Biased information (recently termed bisinformation) continues to be taught in medical curricula, often long after having been debunked. In this paper, we introduce BRICC, a firstin-class initiative that seeks to mitigate medical bisinformation using machine learning to systematically identify and flag text with potential biases, for subsequent review in an expert-in-the-loop fashion, thus greatly accelerating an otherwise labor-intensive process. A gold-standard BRICC dataset was developed throughout several years, and contains over 12K pages of instructional materials. Medical experts meticulously annotated these documents for bias according to comprehensive coding guidelines, emphasizing gender, sex, age, geography, ethnicity, and race. Using this labeled dataset, we trained, validated, and tested medical bias classifiers. We test three classifier approaches: a binary type-specific classifier, a general bias classifier; an ensemble combining bias type-specific classifiers independently-trained; and a multitask learning (MTL) model tasked with predicting both general and type-specific biases. While MTL led to some improvement on race bias detection in terms of F1-score, it did not outperform binary classifiers trained specifically on each task. On general bias detection, the binary classifier achieves up to 0.923 of AUC, a 27.8% improvement over the baseline. This work lays the foundations for debiasing medical curricula by exploring a novel dataset and evaluating different training model strategies. Hence, it offers new pathways for more nuanced and effective mitigation of bisinformation.
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: Aparna Ananthasubramaniam, Yufei Zhu, David Jurgens, Daniel Romero
Abstract: Although the spread of behaviors is influenced by many social factors, existing literature tends to study the effects of single factors -- most often, properties of the social network -- on the final cascade. In order to move towards a more integrated view of cascades, this paper offers the first comprehensive investigation into the role of two social factors in the diffusion of 1,337 popular hashtags representing the production of novel culture on Twitter: 1) the topology of the Twitter social network and 2) performance of each user's probable demographic identity. Here, we show that cascades are best modeled using a combination of network and identity, rather than either factor alone. This combined model best reproduces a composite index of ten cascade properties across all 1,337 hashtags. However, there is important heterogeneity in what social factors are required to reproduce different properties of hashtag cascades. For instance, while a combined network+identity model best predicts the popularity of cascades, a network-only model has better performance in predicting cascade growth and an identity-only model in adopter composition. We are able to predict what type of hashtag is best modeled by each combination of features and use this to further improve performance. Additionally, consistent with prior literature on the combined network+identity model most outperforms the single-factor counterfactuals among hashtags used for expressing racial or regional identity, stance-taking, talking about sports, or variants of existing cultural trends with very slow- or fast-growing communicative need. In sum, our results imply the utility of multi-factor models in predicting cascades, in order to account for the varied ways in which network, identity, and other social factors play a role in the diffusion of hashtags on Twitter.
Authors: Songming Zhang, Yunlong Liang, Shuaibo Wang, Wenjuan Han, Jian Liu, Jinan Xu, Yufeng Chen
Abstract: Knowledge distillation (KD) is a promising technique for model compression in neural machine translation. However, where the knowledge hides in KD is still not clear, which may hinder the development of KD. In this work, we first unravel this mystery from an empirical perspective and show that the knowledge comes from the top-1 predictions of teachers, which also helps us build a potential connection between word- and sequence-level KD. Further, we point out two inherent issues in vanilla word-level KD based on this finding. Firstly, the current objective of KD spreads its focus to whole distributions to learn the knowledge, yet lacks special treatment on the most crucial top-1 information. Secondly, the knowledge is largely covered by the golden information due to the fact that most top-1 predictions of teachers overlap with ground-truth tokens, which further restricts the potential of KD. To address these issues, we propose a novel method named \textbf{T}op-1 \textbf{I}nformation \textbf{E}nhanced \textbf{K}nowledge \textbf{D}istillation (TIE-KD). Specifically, we design a hierarchical ranking loss to enforce the learning of the top-1 information from the teacher. Additionally, we develop an iterative KD procedure to infuse more additional knowledge by distilling on the data without ground-truth targets. Experiments on WMT'14 English-German, WMT'14 English-French and WMT'16 English-Romanian demonstrate that our method can respectively boost Transformer$_{base}$ students by +1.04, +0.60 and +1.11 BLEU scores and significantly outperform the vanilla word-level KD baseline. Besides, our method shows higher generalizability on different teacher-student capacity gaps than existing KD techniques.
Authors: Yuan Sui, Mengyu Zhou, Mingjie Zhou, Shi Han, Dongmei Zhang
Abstract: Large language models (LLMs) are becoming attractive as few-shot reasoners to solve Natural Language (NL)-related tasks. However, the understanding of their capability to process structured data like tables remains an under-explored area. While tables can be serialized as input for LLMs, there is a lack of comprehensive studies on whether LLMs genuinely comprehend this data. In this paper, we try to understand this by designing a benchmark to evaluate the structural understanding capabilities of LLMs through seven distinct tasks, e.g., cell lookup, row retrieval and size detection. Specially, we perform a series of evaluations on the recent most advanced LLM models, GPT-3.5 and GPT-4 and observe that performance varied with different input choices, including table input format, content order, role prompting, and partition marks. Drawing from the insights gained through the benchmark evaluations, we propose $\textit{self-augmentation}$ for effective structural prompting, such as critical value / range identification using internal knowledge of LLMs. When combined with carefully chosen input choices, these structural prompting methods lead to promising improvements in LLM performance on a variety of tabular tasks, e.g., TabFact($\uparrow2.31\%$), HybridQA($\uparrow2.13\%$), SQA($\uparrow2.72\%$), Feverous($\uparrow0.84\%$), and ToTTo($\uparrow5.68\%$). We believe that our open source benchmark and proposed prompting methods can serve as a simple yet generic selection for future research. The code and data of this paper will be temporality released at https://anonymous.4open.science/r/StructuredLLM-76F3/README.md and will be replaced with an official one at https://github.com/microsoft/TableProvider later.
URLs: https://anonymous.4open.science/r/StructuredLLM-76F3/README.md, https://github.com/microsoft/TableProvider
Authors: Tian Liang, Zhiwei He, Wenxiang Jiao, Xing Wang, Rui Wang, Yujiu Yang, Zhaopeng Tu, Shuming Shi
Abstract: Modern large language models (LLMs) like ChatGPT have shown remarkable performance on general language tasks but still struggle on complex reasoning tasks, which drives the research on cognitive behaviors of LLMs to explore human-like problem-solving strategies. Along this direction, one representative strategy is self-reflection, which asks an LLM to refine the solution with the feedback generated by itself iteratively. However, our study shows that such reflection-style methods suffer from the Degeneration-of-Thought (DoT) problem: once the LLM has established confidence in its solutions, it is unable to generate novel thoughts later through reflection even if its initial stance is incorrect. To address the DoT problem, we propose a Multi-Agent Debate (MAD) framework, in which multiple agents express their arguments in the state of "tit for tat" and a judge manages the debate process to obtain a final solution. Clearly, our MAD framework encourages divergent thinking in LLMs which would be helpful for tasks that require deep levels of contemplation. Experiment results on two challenging datasets, commonsense machine translation and counter-intuitive arithmetic reasoning, demonstrate the effectiveness of our MAD framework. Extensive analyses suggest that the adaptive break of debate and the modest level of "tit for tat" state are required for MAD to obtain good performance. Moreover, we find that LLMs might not be a fair judge if different LLMs are used for agents. Code is available at https://github.com/Skytliang/Multi-Agents-Debate.
Authors: Jiaao Chen, Xiaoman Pan, Dian Yu, Kaiqiang Song, Xiaoyang Wang, Dong Yu, Jianshu Chen
Abstract: We investigate how to elicit compositional generalization capabilities in large language models (LLMs). Compositional generalization empowers LLMs to solve complex problems by combining foundational skills, a critical reasoning ability akin to human intelligence. However, even the most advanced LLMs currently struggle with this form of reasoning. We examine this problem within the framework of in-context learning and find that demonstrating both foundational skills and compositional examples grounded in these skills within the same prompt context is crucial. We refer to this prompt structure as skills-in-context (SKiC). With as few as two exemplars, this in-context learning structure enables LLMs to tackle more challenging problems requiring innovative skill combinations, achieving near-perfect systematic generalization across a broad range of tasks. Intriguingly, SKiC also unlocks the latent potential of LLMs, allowing them to more actively utilize pre-existing internal skills acquired during earlier pretraining stages to solve complex reasoning problems. The SKiC structure is robust across different skill constructions and exemplar choices and demonstrates strong transferability to new tasks. Finally, inspired by our in-context learning study, we show that fine-tuning LLMs with SKiC-style data can elicit zero-shot weak-to-strong generalization, enabling the models to solve much harder problems directly with standard prompting.
Authors: Christian Huber, Tu Anh Dinh, Carlos Mullov, Ngoc Quan Pham, Thai Binh Nguyen, Fabian Retkowski, Stefan Constantin, Enes Yavuz Ugan, Danni Liu, Zhaolin Li, Sai Koneru, Jan Niehues, Alexander Waibel
Abstract: The challenge of low-latency speech translation has recently draw significant interest in the research community as shown by several publications and shared tasks. Therefore, it is essential to evaluate these different approaches in realistic scenarios. However, currently only specific aspects of the systems are evaluated and often it is not possible to compare different approaches. In this work, we propose the first framework to perform and evaluate the various aspects of low-latency speech translation under realistic conditions. The evaluation is carried out in an end-to-end fashion. This includes the segmentation of the audio as well as the run-time of the different components. Secondly, we compare different approaches to low-latency speech translation using this framework. We evaluate models with the option to revise the output as well as methods with fixed output. Furthermore, we directly compare state-of-the-art cascaded as well as end-to-end systems. Finally, the framework allows to automatically evaluate the translation quality as well as latency and also provides a web interface to show the low-latency model outputs to the user.
Authors: Chengpeng Li, Zheng Yuan, Hongyi Yuan, Guanting Dong, Keming Lu, Jiancan Wu, Chuanqi Tan, Xiang Wang, Chang Zhou
Abstract: In math reasoning with large language models (LLMs), fine-tuning data augmentation by query evolution and diverse reasoning paths is empirically verified effective, profoundly narrowing the gap between open-sourced LLMs and cutting-edge proprietary LLMs. In this paper, we conduct an investigation for such data augmentation in math reasoning and are intended to answer: (1) What strategies of data augmentation are more effective; (2) What is the scaling relationship between the amount of augmented data and model performance; and (3) Can data augmentation incentivize generalization to out-of-domain mathematical reasoning tasks? To this end, we create two new dataset AugGSM8K and AugMATH, by complicating and diversifying the queries and sampling multiple reasoning paths from GSM8K and MATH. We obtained a series of LLMs called MuggleMath by fine-tuning LLaMA models on AugGSM8K and AugMATH. MuggleMath substantially achieves new state-of-the-art on GSM8K and MATH. A log-linear relationship and a segmented log-linear are presented between MuggleMath's performance and the amount of augmented data on GSM8K and MATH, respectively. We also find that it is weak in out-of-domain math reasoning generalization from AugGSM8K to MATH and from AugMATH to GSM8K, which suggests that augmenting queries that cover a broader range of subjects is more beneficial for generalization. We release our codes and augmented data in https://github.com/OFA-Sys/gsm8k-ScRel.
Authors: Brian Gordon, Yonatan Bitton, Yonatan Shafir, Roopal Garg, Xi Chen, Dani Lischinski, Daniel Cohen-Or, Idan Szpektor
Abstract: While existing image-text alignment models reach high quality binary assessments, they fall short of pinpointing the exact source of misalignment. In this paper, we present a method to provide detailed textual and visual explanation of detected misalignments between text-image pairs. We leverage large language models and visual grounding models to automatically construct a training set that holds plausible misaligned captions for a given image and corresponding textual explanations and visual indicators. We also publish a new human curated test set comprising ground-truth textual and visual misalignment annotations. Empirical results show that fine-tuning vision language models on our training set enables them to articulate misalignments and visually indicate them within images, outperforming strong baselines both on the binary alignment classification and the explanation generation tasks. Our method code and human curated test set are available at: https://mismatch-quest.github.io/
Authors: Christian Huber, Alexander Waibel
Abstract: Despite recent advances, Automatic Speech Recognition (ASR) systems are still far from perfect. Typical errors include acronyms, named entities and domain-specific special words for which little or no data is available. To address the problem of recognizing these words, we propose an self-supervised continual learning approach. Given the audio of a lecture talk with corresponding slides, we bias the model towards decoding new words from the slides by using a memory-enhanced ASR model from previous work. Then, we perform inference on the talk, collecting utterances that contain detected new words into an adaptation dataset. Continual learning is then performed on this set by adapting low-rank matrix weights added to each weight matrix of the model. The whole procedure is iterated for many talks. We show that with this approach, we obtain increasing performance on the new words when they occur more frequently (more than 80% recall) while preserving the general performance of the model.
Authors: Yanis Labrak, Adrien Bazoge, Emmanuel Morin, Pierre-Antoine Gourraud, Mickael Rouvier, Richard Dufour
Abstract: Large Language Models (LLMs) have demonstrated remarkable versatility in recent years, offering potential applications across specialized domains such as healthcare and medicine. Despite the availability of various open-source LLMs tailored for health contexts, adapting general-purpose LLMs to the medical domain presents significant challenges. In this paper, we introduce BioMistral, an open-source LLM tailored for the biomedical domain, utilizing Mistral as its foundation model and further pre-trained on PubMed Central. We conduct a comprehensive evaluation of BioMistral on a benchmark comprising 10 established medical question-answering (QA) tasks in English. We also explore lightweight models obtained through quantization and model merging approaches. Our results demonstrate BioMistral's superior performance compared to existing open-source medical models and its competitive edge against proprietary counterparts. Finally, to address the limited availability of data beyond English and to assess the multilingual generalization of medical LLMs, we automatically translated and evaluated this benchmark into 7 other languages. This marks the first large-scale multilingual evaluation of LLMs in the medical domain. Datasets, multilingual evaluation benchmarks, scripts, and all the models obtained during our experiments are freely released.
Authors: Sahand Sabour, Siyang Liu, Zheyuan Zhang, June M. Liu, Jinfeng Zhou, Alvionna S. Sunaryo, Juanzi Li, Tatia M. C. Lee, Rada Mihalcea, Minlie Huang
Abstract: Recent advances in Large Language Models (LLMs) have highlighted the need for robust, comprehensive, and challenging benchmarks. Yet, research on evaluating their Emotional Intelligence (EI) is considerably limited. Existing benchmarks have two major shortcomings: first, they mainly focus on emotion recognition, neglecting essential EI capabilities such as emotion regulation and thought facilitation through emotion understanding; second, they are primarily constructed from existing datasets, which include frequent patterns, explicit information, and annotation errors, leading to unreliable evaluation. We propose EmoBench, a benchmark that draws upon established psychological theories and proposes a comprehensive definition for machine EI, including Emotional Understanding and Emotional Application. EmoBench includes a set of 400 hand-crafted questions in English and Chinese, which are meticulously designed to require thorough reasoning and understanding. Our findings reveal a considerable gap between the EI of existing LLMs and the average human, highlighting a promising direction for future research. Our code and data are publicly available at https://github.com/Sahandfer/EmoBench.
Authors: Ivan Rep, David Duki\'c, Jan \v{S}najder
Abstract: While BERT produces high-quality sentence embeddings, its pre-training computational cost is a significant drawback. In contrast, ELECTRA delivers a cost-effective pre-training objective and downstream task performance improvements, but not as performant sentence embeddings. The community tacitly stopped utilizing ELECTRA's sentence embeddings for semantic textual similarity (STS). We notice a significant drop in performance when using the ELECTRA discriminator's last layer in comparison to earlier layers. We explore this drop and devise a way to repair ELECTRA's embeddings, proposing a novel truncated model fine-tuning (TMFT) method. TMFT improves the Spearman correlation coefficient by over 8 points while increasing parameter efficiency on the STS benchmark dataset. We extend our analysis to various model sizes and languages. Further, we discover the surprising efficacy of ELECTRA's generator model, which performs on par with BERT, using significantly fewer parameters and a substantially smaller embedding size. Finally, we observe further boosts by combining TMFT with a word similarity task or domain adaptive pre-training.
Authors: William Merrill, Zhaofeng Wu, Norihito Naka, Yoon Kim, Tal Linzen
Abstract: Do LMs infer the semantics of text from co-occurrence patterns in their training data? Merrill et al. (2022) argue that, in theory, sentence co-occurrence probabilities predicted by an optimal LM should reflect the entailment relationship of the constituent sentences, but it is unclear whether probabilities predicted by neural LMs encode entailment in this way because of strong assumptions made by Merrill et al. (namely, that humans always avoid redundancy). In this work, we investigate whether their theory can be used to decode entailment relations from neural LMs. We find that a test similar to theirs can decode entailment relations between natural sentences, well above random chance, though not perfectly, across many datasets and LMs. This suggests LMs implicitly model aspects of semantics to predict semantic effects on sentence co-occurrence patterns. However, we find the test that predicts entailment in practice works in the opposite direction to the theoretical test. We thus revisit the assumptions underlying the original test, finding its derivation did not adequately account for redundancy in human-written text. We argue that better accounting for redundancy related to explanations might derive the observed flipped test and, more generally, improve computational models of speakers in linguistics.
Authors: Tom Hosking, Hao Tang, Mirella Lapata
Abstract: We propose a method for unsupervised abstractive opinion summarization, that combines the attributability and scalability of extractive approaches with the coherence and fluency of Large Language Models (LLMs). Our method, HIRO, learns an index structure that maps sentences to a path through a semantically organized discrete hierarchy. At inference time, we populate the index and use it to identify and retrieve clusters of sentences containing popular opinions from input reviews. Then, we use a pretrained LLM to generate a readable summary that is grounded in these extracted evidential clusters. The modularity of our approach allows us to evaluate its efficacy at each stage. We show that HIRO learns an encoding space that is more semantically structured than prior work, and generates summaries that are more representative of the opinions in the input reviews. Human evaluation confirms that HIRO generates significantly more coherent, detailed and accurate summaries.
Authors: Jingzhe Shi, Jialuo Li, Qinwei Ma, Zaiwen Yang, Huan Ma, Lei Li
Abstract: Businesses and software platforms are increasingly turning to Large Language Models (LLMs) such as GPT-3.5, GPT-4, GLM-3, and LLaMa-2 for chat assistance with file access or as reasoning agents for customer service. However, current LLM-based customer service models have limited integration with customer profiles and lack the operational capabilities necessary for effective service. Moreover, existing API integrations emphasize diversity over the precision and error avoidance essential in real-world customer service scenarios. To address these issues, we propose an LLM agent named CHOPS (CHat with custOmer Profile in existing System), designed to: (1) efficiently utilize existing databases or systems for accessing user information or interacting with these systems following existing guidelines; (2) provide accurate and reasonable responses or carry out required operations in the system while avoiding harmful operations; and (3) leverage a combination of small and large LLMs to achieve satisfying performance at a reasonable inference cost. We introduce a practical dataset, the CPHOS-dataset, which includes a database, guiding files, and QA pairs collected from CPHOS, an online platform that facilitates the organization of simulated Physics Olympiads for high school teachers and students. We have conducted extensive experiments to validate the performance of our proposed CHOPS architecture using the CPHOS-dataset, with the aim of demonstrating how LLMs can enhance or serve as alternatives to human customer service. Code for our proposed architecture and dataset can be found at {https://github.com/JingzheShi/CHOPS}.
Authors: Anton Sch\"afer, Thomas Hofmann, Imanol Schlag, Tiago Pimentel
Abstract: Tokenisation is a core part of language models (LMs). It involves splitting a character sequence into subwords which are assigned arbitrary indices before being served to the LM. While typically lossless, however, this process may lead to less sample efficient LM training: as it removes character-level information, it could make it harder for LMs to generalise across similar subwords, such as now and Now. We refer to such subwords as near duplicates. In this paper, we study the impact of near duplicate subwords on LM training efficiency. First, we design an experiment that gives us an upper bound to how much we should expect a model to improve if we could perfectly generalise across near duplicates. We do this by duplicating each subword in our LM's vocabulary, creating perfectly equivalent classes of subwords. Experimentally, we find that LMs need roughly 17% more data when trained in a fully duplicated setting. Second, we investigate the impact of naturally occurring near duplicates on LMs. Here, we see that merging them considerably hurts LM performance. Therefore, although subword duplication negatively impacts LM training efficiency, naturally occurring near duplicates may not be as similar as anticipated, limiting the potential for performance improvements.
Authors: Anton Sch\"afer, Shauli Ravfogel, Thomas Hofmann, Tiago Pimentel, Imanol Schlag
Abstract: Multilinguality is crucial for extending recent advancements in language modelling to diverse linguistic communities. To maintain high performance while representing multiple languages, multilingual models ideally align representations, allowing what is learned in one language to generalise to others. Prior research has emphasised the importance of parallel data and shared vocabulary elements as key factors for such alignment. In this study, we investigate an unintuitive novel driver of cross-lingual generalisation: language imbalance. In controlled experiments on perfectly equivalent cloned languages, we observe that the existence of a predominant language during training boosts the performance of less frequent languages and leads to stronger alignment of model representations across languages. Furthermore, we find that this trend is amplified with scale: with large enough models or long enough training, we observe that bilingual training data with a 90/10 language split yields better performance on both languages than a balanced 50/50 split. Building on these insights, we design training schemes that can improve performance in all cloned languages, even without altering the training data. As we extend our analysis to real languages, we find that infrequent languages still benefit from frequent ones, yet whether language imbalance causes cross-lingual generalisation there is not conclusive.
Authors: Matteo Merler, Katsiaryna Haitsiukevich, Nicola Dainese, Pekka Marttinen
Abstract: State of the art Symbolic Regression (SR) methods currently build specialized models, while the application of Large Language Models (LLMs) remains largely unexplored. In this work, we introduce the first comprehensive framework that utilizes LLMs for the task of SR. We propose In-Context Symbolic Regression (ICSR), an SR method which iteratively refines a functional form with an LLM and determines its coefficients with an external optimizer. ICSR leverages LLMs' strong mathematical prior both to propose an initial set of possible functions given the observations and to refine them based on their errors. Our findings reveal that LLMs are able to successfully find symbolic equations that fit the given data, matching or outperforming the overall performance of the best SR baselines on four popular benchmarks, while yielding simpler equations with better out of distribution generalization.
Authors: Vyas Raina, Rao Ma, Charles McGhee, Kate Knill, Mark Gales
Abstract: Recent developments in large speech foundation models like Whisper have led to their widespread use in many automatic speech recognition (ASR) applications. These systems incorporate `special tokens' in their vocabulary, such as $\texttt{<|endoftext|>}$, to guide their language generation process. However, we demonstrate that these tokens can be exploited by adversarial attacks to manipulate the model's behavior. We propose a simple yet effective method to learn a universal acoustic realization of Whisper's $\texttt{<|endoftext|>}$ token, which, when prepended to any speech signal, encourages the model to ignore the speech and only transcribe the special token, effectively `muting' the model. Our experiments demonstrate that the same, universal 0.64-second adversarial audio segment can successfully mute a target Whisper ASR model for over 97\% of speech samples. Moreover, we find that this universal adversarial audio segment often transfers to new datasets and tasks. Overall this work demonstrates the vulnerability of Whisper models to `muting' adversarial attacks, where such attacks can pose both risks and potential benefits in real-world settings: for example the attack can be used to bypass speech moderation systems, or conversely the attack can also be used to protect private speech data.
Authors: Boxi Cao, Keming Lu, Xinyu Lu, Jiawei Chen, Mengjie Ren, Hao Xiang, Peilin Liu, Yaojie Lu, Ben He, Xianpei Han, Le Sun, Hongyu Lin, Bowen Yu
Abstract: Alignment is the most critical step in building large language models (LLMs) that meet human needs. With the rapid development of LLMs gradually surpassing human capabilities, traditional alignment methods based on human-annotation are increasingly unable to meet the scalability demands. Therefore, there is an urgent need to explore new sources of automated alignment signals and technical approaches. In this paper, we systematically review the recently emerging methods of automated alignment, attempting to explore how to achieve effective, scalable, automated alignment once the capabilities of LLMs exceed those of humans. Specifically, we categorize existing automated alignment methods into 4 major categories based on the sources of alignment signals and discuss the current status and potential development of each category. Additionally, we explore the underlying mechanisms that enable automated alignment and discuss the essential factors that make automated alignment technologies feasible and effective from the fundamental role of alignment.
Authors: Vera Neplenbroek, Arianna Bisazza, Raquel Fern\'andez
Abstract: Generative large language models (LLMs) have been shown to exhibit harmful biases and stereotypes. While safety fine-tuning typically takes place in English, if at all, these models are being used by speakers of many different languages. There is existing evidence that the performance of these models is inconsistent across languages and that they discriminate based on demographic factors of the user. Motivated by this, we investigate whether the social stereotypes exhibited by LLMs differ as a function of the language used to prompt them, while controlling for cultural differences and task accuracy. To this end, we present MBBQ (Multilingual Bias Benchmark for Question-answering), a carefully curated version of the English BBQ dataset extended to Dutch, Spanish, and Turkish, which measures stereotypes commonly held across these languages. We further complement MBBQ with a parallel control dataset to measure task performance on the question-answering task independently of bias. Our results based on several open-source and proprietary LLMs confirm that some non-English languages suffer from bias more than English, even when controlling for cultural shifts. Moreover, we observe significant cross-lingual differences in bias behaviour for all except the most accurate models. With the release of MBBQ, we hope to encourage further research on bias in multilingual settings. The dataset and code are available at https://github.com/Veranep/MBBQ.
Authors: Matteo Esposito, Francesco Palagiano, Valentina Lenarduzzi, Davide Taibi
Abstract: Context. Risk analysis assesses potential risks in specific scenarios. Risk analysis principles are context-less; the same methodology can be applied to a risk connected to health and information technology security. Risk analysis requires a vast knowledge of national and international regulations and standards and is time and effort-intensive. A large language model can quickly summarize information in less time than a human and can be fine-tuned to specific tasks. Aim. Our empirical study aims to investigate the effectiveness of Retrieval-Augmented Generation and fine-tuned LLM in Risk analysis. To our knowledge, no prior study has explored its capabilities in risk analysis. Method. We manually curated \totalscenarios unique scenarios leading to \totalsamples representative samples from over 50 mission-critical analyses archived by the industrial context team in the last five years. We compared the base GPT-3.5 and GPT-4 models versus their Retrieval-Augmented Generation and fine-tuned counterparts. We employ two human experts as competitors of the models and three other three human experts to review the models and the former human expert's analysis. The reviewers analyzed 5,000 scenario analyses. Results and Conclusions. HEs demonstrated higher accuracy, but LLMs are quicker and more actionable. Moreover, our findings show that RAG-assisted LLMs have the lowest hallucination rates, effectively uncovering hidden risks and complementing human expertise. Thus, the choice of model depends on specific needs, with FTMs for accuracy, RAG for hidden risks discovery, and base models for comprehensiveness and actionability. Therefore, experts can leverage LLMs for an effective complementing companion in risk analysis within a condensed timeframe. They can also save costs by averting unnecessary expenses associated with implementing unwarranted countermeasures.
Authors: Guanting Dong, Keming Lu, Chengpeng Li, Tingyu Xia, Bowen Yu, Chang Zhou, Jingren Zhou
Abstract: One core capability of large language models (LLMs) is to follow natural language instructions. However, the issue of automatically constructing high-quality training data to enhance the complex instruction-following abilities of LLMs without manual annotation remains unresolved. In this paper, we introduce AutoIF, the first scalable and reliable method for automatically generating instruction-following training data. AutoIF transforms the validation of instruction-following data quality into code verification, requiring LLMs to generate instructions, the corresponding code to check the correctness of the instruction responses, and unit test samples to verify the code's correctness. Then, execution feedback-based rejection sampling can generate data for Supervised Fine-Tuning (SFT) and Reinforcement Learning from Human Feedback (RLHF) training. AutoIF achieves significant improvements across three training algorithms, SFT, Offline DPO, and Online DPO, when applied to the top open-source LLMs, Qwen2 and LLaMA3, in self-alignment and strong-to-weak distillation settings. Our code is publicly available at https://github.com/QwenLM/AutoIF.
Authors: Jikai Wang, Yi Su, Juntao Li, Qingrong Xia, Zi Ye, Xinyu Duan, Zhefeng Wang, Min Zhang
Abstract: Autoregressive language models demonstrate excellent performance in various scenarios. However, the inference efficiency is limited by its one-step-one-word generation mode, which has become a pressing problem recently as the models become increasingly larger. Speculative decoding employs a "draft and then verify" mechanism to allow multiple tokens to be generated in one step, realizing lossless acceleration. Existing methods mainly adopt fixed heuristic draft structures, which fail to adapt to different situations to maximize the acceptance length during verification. To alleviate this dilemma, we proposed OPT-Tree, an algorithm to construct adaptive and scalable draft trees. It searches the optimal tree structure that maximizes the mathematical expectation of the acceptance length in each decoding step. Experimental results reveal that OPT-Tree outperforms the existing draft structures and achieves a speed-up ratio of up to 3.2 compared with autoregressive decoding. If the draft model is powerful enough and the node budget is sufficient, it can generate more than ten tokens in a single step. Our code is available at https://github.com/Jikai0Wang/OPT-Tree.
Authors: Chengpeng Li, Guanting Dong, Mingfeng Xue, Ru Peng, Xiang Wang, Dayiheng Liu
Abstract: Large language models (LLMs) have made impressive progress in handling simple math problems, yet they still struggle with more challenging and complex mathematical tasks. In this paper, we introduce a series of LLMs that employs the Decomposition of thought with code assistance and self-correction for mathematical reasoning, dubbed as DotaMath. DotaMath models tackle complex mathematical tasks by decomposing them into simpler logical subtasks, leveraging code to solve these subtasks, obtaining fine-grained feedback from the code interpreter, and engaging in self-reflection and correction. By annotating diverse interactive tool-use trajectories and employing query evolution on GSM8K and MATH datasets, we generate an instruction fine-tuning dataset called DotaMathQA with 574K query-response pairs. We train a series of base LLMs using imitation learning on DotaMathQA, resulting in DotaMath models that achieve remarkable performance compared to open-source LLMs across various in-domain and out-of-domain benchmarks. Notably, DotaMath-deepseek-7B showcases an outstanding performance of 64.8% on the competitive MATH dataset and 86.7% on GSM8K. Besides, DotaMath-deepseek-7B maintains strong competitiveness on a series of in-domain and out-of-domain benchmarks (Avg. 80.1%). Looking forward, we anticipate that the DotaMath paradigm will open new pathways for addressing intricate mathematical problems. Our code is publicly available at https://github.com/ChengpengLi1003/DotaMath.
Authors: Saar Kuzi, Shervin Malmasi
Abstract: Consumers on a shopping mission often leverage both product search and information seeking systems, such as web search engines and Question Answering (QA) systems, in an iterative process to improve their understanding of available products and reach a purchase decision. While product search is useful for shoppers to find the actual products meeting their requirements in the catalog, information seeking systems can be utilized to answer any questions they may have to refine those requirements. The recent success of Large Language Models (LLMs) has opened up an opportunity to bridge the gap between the two tasks to help customers achieve their goals quickly and effectively by integrating conversational QA within product search. In this paper, we propose to recommend users Question-Answer (Q&A) pairs that are relevant to their product search and can help them make a purchase decision. We discuss the different aspects of the problem including the requirements and characteristics of the Q&A pairs, their generation, and the optimization of the Q&A recommendation task. We highlight the challenges, open problems, and suggested solutions to encourage future research in this emerging area.
Authors: Qingcheng Zeng, Mingyu Jin, Qinkai Yu, Zhenting Wang, Wenyue Hua, Zihao Zhou, Guangyan Sun, Yanda Meng, Shiqing Ma, Qifan Wang, Felix Juefei-Xu, Kaize Ding, Fan Yang, Ruixiang Tang, Yongfeng Zhang
Abstract: Large Language Models (LLMs) are employed across various high-stakes domains, where the reliability of their outputs is crucial. One commonly used method to assess the reliability of LLMs' responses is uncertainty estimation, which gauges the likelihood of their answers being correct. While many studies focus on improving the accuracy of uncertainty estimations for LLMs, our research investigates the fragility of uncertainty estimation and explores potential attacks. We demonstrate that an attacker can embed a backdoor in LLMs, which, when activated by a specific trigger in the input, manipulates the model's uncertainty without affecting the final output. Specifically, the proposed backdoor attack method can alter an LLM's output probability distribution, causing the probability distribution to converge towards an attacker-predefined distribution while ensuring that the top-1 prediction remains unchanged. Our experimental results demonstrate that this attack effectively undermines the model's self-evaluation reliability in multiple-choice questions. For instance, we achieved a 100 attack success rate (ASR) across three different triggering strategies in four models. Further, we investigate whether this manipulation generalizes across different prompts and domains. This work highlights a significant threat to the reliability of LLMs and underscores the need for future defenses against such attacks. The code is available at https://github.com/qcznlp/uncertainty_attack.
Authors: Shaomang Huang, Jianfeng Pan, Hanzhong Zheng
Abstract: In the domain of Large Language Model (LLM), LLMs demonstrate significant capabilities in natural language understanding and generation. With the growing needs of applying LLMs on various domains, it is a research question that how to efficiently train and build a model that has expertise in different domains but with a low training cost. We propose CCoE architecture, a framework of easily coupling multiple strong domain experts together to fuse into a big LLM, provides a collective way of utilizing the different domain expert LLMs. Besides, training a large collaborative of multiple expert LLMs requires a high requirements on training sources. CCoE bypasses this problem through isolating other experts and train each expert separately. The design of CCoE assembles multiple expert LLMs through the CoE (Collaboration of Experts) layer. Each CoE layer could have one or more expert LLMs. Expert LLMs have different number of layers and have been well-trained for different domain tasks. Each expert is fine-tuned to be able to achieve the comparable results with SOTA domain LLMs. We start from 5 experts in the domain of Code, Math, Law, text-to-SQL and Medical. The results indicate that our CCoE framework can easily and efficiently boost nearly 10%-20% performance on original base model in different domains but using less resources on training, as well as inference.
Authors: Xingmeng Zhao, Dan Schumacher, Sashank Nalluri, Xavier Walton, Suhana Shrestha, Anthony Rios
Abstract: Increasing cycling for transportation or recreation can boost health and reduce the environmental impacts of vehicles. However, news agencies' ideologies and reporting styles often influence public perception of cycling. For example, if news agencies overly report cycling accidents, it may make people perceive cyclists as "dangerous," reducing the number of cyclists who opt to cycle. Additionally, a decline in cycling can result in less government funding for safe infrastructure. In this paper, we develop a method for detecting the perceived perception of cyclists within news headlines. We introduce a new dataset called ``Bike Frames'' to accomplish this. The dataset consists of 31,480 news headlines and 1,500 annotations. Our focus is on analyzing 11,385 headlines from the United States. We also introduce the BikeFrame Chain-of-Code framework to predict cyclist perception, identify accident-related headlines, and determine fault. This framework uses pseudocode for precise logic and integrates news agency bias analysis for improved predictions over traditional chain-of-thought reasoning in large language models. Our method substantially outperforms other methods, and most importantly, we find that incorporating news bias information substantially impacts performance, improving the average F1 from .739 to .815. Finally, we perform a comprehensive case study on US-based news headlines, finding reporting differences between news agencies and cycling-specific websites as well as differences in reporting depending on the gender of cyclists. WARNING: This paper contains descriptions of accidents and death.
Authors: Shreyas Vaidya, Arvind Kumar Sharma, Prajwal Gatti, Anand Mishra
Abstract: In this work, we study the task of visually translating scene text from a source language (e.g., Hindi) to a target language (e.g., English). Visual translation involves not just the recognition and translation of scene text but also the generation of the translated image that preserves visual features of the source scene text, such as font, size, and background. There are several challenges associated with this task, such as translation with limited context, deciding between translation and transliteration, accommodating varying text lengths within fixed spatial boundaries, and preserving the font and background styles of the source scene text in the target language. To address this problem, we make the following contributions: (i) We study visual translation as a standalone problem for the first time in the literature. (ii) We present a cascaded framework for visual translation that combines state-of-the-art modules for scene text recognition, machine translation, and scene text synthesis as a baseline for the task. (iii) We propose a set of task-specific design enhancements to design a variant of the baseline to obtain performance improvements. (iv) Currently, the existing related literature lacks any comprehensive performance evaluation for this novel task. To fill this gap, we introduce several automatic and user-assisted evaluation metrics designed explicitly for evaluating visual translation. Further, we evaluate presented baselines for translating scene text between Hindi and English. Our experiments demonstrate that although we can effectively perform visual translation over a large collection of scene text images, the presented baseline only partially addresses challenges posed by visual translation tasks. We firmly believe that this new task and the limitations of existing models, as reported in this paper, should encourage further research in visual translation.
Authors: Eran Malach
Abstract: Large language models display remarkable capabilities in logical and mathematical reasoning, allowing them to solve complex tasks. Interestingly, these abilities emerge in networks trained on the simple task of next-token prediction. In this work, we present a theoretical framework for studying auto-regressive next-token predictors. We demonstrate that even simple models such as linear next-token predictors, trained on Chain-of-Thought (CoT) data, can approximate any function efficiently computed by a Turing machine. We introduce a new complexity measure -- length complexity -- which measures the number of intermediate tokens in a CoT sequence required to approximate some target function, and analyze the interplay between length complexity and other notions of complexity. Finally, we show experimentally that simple next-token predictors, such as linear networks and shallow Multi-Layer Perceptrons (MLPs), display non-trivial performance on text generation and arithmetic tasks. Our results demonstrate that the power of today's LLMs can be attributed, to a great extent, to the auto-regressive next-token training scheme, and not necessarily to a particular choice of architecture.
Authors: Brian Yang, Huangyuan Su, Nikolaos Gkanatsios, Tsung-Wei Ke, Ayush Jain, Jeff Schneider, Katerina Fragkiadaki
Abstract: Diffusion models excel at modeling complex and multimodal trajectory distributions for decision-making and control. Reward-gradient guided denoising has been recently proposed to generate trajectories that maximize both a differentiable reward function and the likelihood under the data distribution captured by a diffusion model. Reward-gradient guided denoising requires a differentiable reward function fitted to both clean and noised samples, limiting its applicability as a general trajectory optimizer. In this paper, we propose DiffusionES, a method that combines gradient-free optimization with trajectory denoising to optimize black-box non-differentiable objectives while staying in the data manifold. Diffusion-ES samples trajectories during evolutionary search from a diffusion model and scores them using a black-box reward function. It mutates high-scoring trajectories using a truncated diffusion process that applies a small number of noising and denoising steps, allowing for much more efficient exploration of the solution space. We show that DiffusionES achieves state-of-the-art performance on nuPlan, an established closed-loop planning benchmark for autonomous driving. Diffusion-ES outperforms existing sampling-based planners, reactive deterministic or diffusion-based policies, and reward-gradient guidance. Additionally, we show that unlike prior guidance methods, our method can optimize non-differentiable language-shaped reward functions generated by few-shot LLM prompting. When guided by a human teacher that issues instructions to follow, our method can generate novel, highly complex behaviors, such as aggressive lane weaving, which are not present in the training data. This allows us to solve the hardest nuPlan scenarios which are beyond the capabilities of existing trajectory optimization methods and driving policies.
Authors: Benjamin Bergner, Andrii Skliar, Amelie Royer, Tijmen Blankevoort, Yuki Asano, Babak Ehteshami Bejnordi
Abstract: Large language models (LLMs) have become ubiquitous in practice and are widely used for generation tasks such as translation, summarization and instruction following. However, their enormous size and reliance on autoregressive decoding increase deployment costs and complicate their use in latency-critical applications. In this work, we propose a hybrid approach that combines language models of different sizes to increase the efficiency of autoregressive decoding while maintaining high performance. Our method utilizes a pretrained frozen LLM that encodes all prompt tokens once in parallel, and uses the resulting representations to condition and guide a small language model (SLM), which then generates the response more efficiently. We investigate the combination of encoder-decoder LLMs with both encoder-decoder and decoder-only SLMs from different model families and only require fine-tuning of the SLM. Experiments with various benchmarks show substantial speedups of up to $4\times$, with minor performance penalties of $1-2\%$ for translation and summarization tasks compared to the LLM.
Authors: Qinyu Zhao, Ming Xu, Kartik Gupta, Akshay Asthana, Liang Zheng, Stephen Gould
Abstract: Large vision-language models (LVLMs), designed to interpret and respond to human instructions, occasionally generate hallucinated or harmful content due to inappropriate instructions. This study uses linear probing to shed light on the hidden knowledge at the output layers of LVLMs. We demonstrate that the logit distributions of the first tokens contain sufficient information to determine whether to respond to the instructions, including recognizing unanswerable visual questions, defending against jailbreaking attacks, and identifying deceptive questions. Such hidden knowledge is gradually lost in logits of subsequent tokens during response generation. Then, we illustrate a simple decoding strategy at the generation of the first token, effectively improving the generated content. In experiments, we find a few interesting insights: First, the CLIP model already contains a strong signal for solving these tasks, which indicates potential bias in the existing datasets. Second, we observe performance improvement by utilizing the first logit distributions on three additional tasks, including indicating uncertainty in math solving, mitigating hallucination, and image classification. Last, with the same training data, simply finetuning LVLMs improves models' performance but is still inferior to linear probing on these tasks.
Authors: Abhishek Ghose, Emma Thuong Nguyen
Abstract: Active learning (AL) techniques optimally utilize a labeling budget by iteratively selecting instances that are most valuable for learning. However, they lack ``prerequisite checks'', i.e., there are no prescribed criteria to pick an AL algorithm best suited for a dataset. A practitioner must pick a technique they \emph{trust} would beat random sampling, based on prior reported results, and hope that it is resilient to the many variables in their environment: dataset, labeling budget and prediction pipelines. The important questions then are: how often on average, do we expect any AL technique to reliably beat the computationally cheap and easy-to-implement strategy of random sampling? Does it at least make sense to use AL in an ``Always ON'' mode in a prediction pipeline, so that while it might not always help, it never under-performs random sampling? How much of a role does the prediction pipeline play in AL's success? We examine these questions in detail for the task of text classification using pre-trained representations, which are ubiquitous today. Our primary contribution here is a rigorous evaluation of AL techniques, old and new, across setups that vary wrt datasets, text representations and classifiers. This unlocks multiple insights around warm-up times, i.e., number of labels before gains from AL are seen, viability of an ``Always ON'' mode and the relative significance of different factors. Additionally, we release a framework for rigorous benchmarking of AL techniques for text classification.
Authors: Akhil Padmanabha, Jessie Yuan, Janavi Gupta, Zulekha Karachiwalla, Carmel Majidi, Henny Admoni, Zackory Erickson
Abstract: Physically assistive robots present an opportunity to significantly increase the well-being and independence of individuals with motor impairments or other forms of disability who are unable to complete activities of daily living. Speech interfaces, especially ones that utilize Large Language Models (LLMs), can enable individuals to effectively and naturally communicate high-level commands and nuanced preferences to robots. Frameworks for integrating LLMs as interfaces to robots for high level task planning and code generation have been proposed, but fail to incorporate human-centric considerations which are essential while developing assistive interfaces. In this work, we present a framework for incorporating LLMs as speech interfaces for physically assistive robots, constructed iteratively with 3 stages of testing involving a feeding robot, culminating in an evaluation with 11 older adults at an independent living facility. We use both quantitative and qualitative data from the final study to validate our framework and additionally provide design guidelines for using LLMs as speech interfaces for assistive robots. Videos and supporting files are located on our project website: https://sites.google.com/andrew.cmu.edu/voicepilot/
Authors: Navonil Majumder, Chia-Yu Hung, Deepanway Ghosal, Wei-Ning Hsu, Rada Mihalcea, Soujanya Poria
Abstract: Generative multimodal content is increasingly prevalent in much of the content creation arena, as it has the potential to allow artists and media personnel to create pre-production mockups by quickly bringing their ideas to life. The generation of audio from text prompts is an important aspect of such processes in the music and film industry. Many of the recent diffusion-based text-to-audio models focus on training increasingly sophisticated diffusion models on a large set of datasets of prompt-audio pairs. These models do not explicitly focus on the presence of concepts or events and their temporal ordering in the output audio with respect to the input prompt. Our hypothesis is focusing on how these aspects of audio generation could improve audio generation performance in the presence of limited data. As such, in this work, using an existing text-to-audio model Tango, we synthetically create a preference dataset where each prompt has a winner audio output and some loser audio outputs for the diffusion model to learn from. The loser outputs, in theory, have some concepts from the prompt missing or in an incorrect order. We fine-tune the publicly available Tango text-to-audio model using diffusion-DPO (direct preference optimization) loss on our preference dataset and show that it leads to improved audio output over Tango and AudioLDM2, in terms of both automatic- and manual-evaluation metrics.
Authors: Jian Hu, Xibin Wu, Weixun Wang, Xianyu, Dehao Zhang, Yu Cao
Abstract: As large language models (LLMs) continue to grow by scaling laws, reinforcement learning from human feedback (RLHF) has gained significant attention due to its outstanding performance. However, unlike pretraining or fine-tuning a single model, scaling reinforcement learning from human feedback (RLHF) for training large language models poses coordination challenges across four models. We present OpenRLHF, an open-source framework enabling efficient RLHF scaling. Unlike existing RLHF frameworks that co-locate four models on the same GPUs, OpenRLHF re-designs scheduling for the models beyond 70B parameters using Ray, vLLM, and DeepSpeed, leveraging improved resource utilization and diverse training approaches. Integrating seamlessly with Hugging Face, OpenRLHF provides an out-of-the-box solution with optimized algorithms and launch scripts, which ensures user-friendliness. OpenRLHF implements RLHF, DPO, rejection sampling, and other alignment techniques. Empowering state-of-the-art LLM development, OpenRLHF's code is available at \url{https://github.com/OpenRLHF/OpenRLHF}.
Authors: Yasin Abbasi Yadkori, Ilja Kuzborskij, Andr\'as Gy\"orgy, Csaba Szepesv\'ari
Abstract: We explore uncertainty quantification in large language models (LLMs), with the goal to identify when uncertainty in responses given a query is large. We simultaneously consider both epistemic and aleatoric uncertainties, where the former comes from the lack of knowledge about the ground truth (such as about facts or the language), and the latter comes from irreducible randomness (such as multiple possible answers). In particular, we derive an information-theoretic metric that allows to reliably detect when only epistemic uncertainty is large, in which case the output of the model is unreliable. This condition can be computed based solely on the output of the model obtained simply by some special iterative prompting based on the previous responses. Such quantification, for instance, allows to detect hallucinations (cases when epistemic uncertainty is high) in both single- and multi-answer responses. This is in contrast to many standard uncertainty quantification strategies (such as thresholding the log-likelihood of a response) where hallucinations in the multi-answer case cannot be detected. We conduct a series of experiments which demonstrate the advantage of our formulation. Further, our investigations shed some light on how the probabilities assigned to a given output by an LLM can be amplified by iterative prompting, which might be of independent interest.
Authors: Chenming Zhu, Tai Wang, Wenwei Zhang, Kai Chen, Xihui Liu
Abstract: Although great progress has been made in 3D visual grounding, current models still rely on explicit textual descriptions for grounding and lack the ability to reason human intentions from implicit instructions. We propose a new task called 3D reasoning grounding and introduce a new benchmark ScanReason which provides over 10K question-answer-location pairs from five reasoning types that require the synerization of reasoning and grounding. We further design our approach, ReGround3D, composed of the visual-centric reasoning module empowered by Multi-modal Large Language Model (MLLM) and the 3D grounding module to obtain accurate object locations by looking back to the enhanced geometry and fine-grained details from the 3D scenes. A chain-of-grounding mechanism is proposed to further boost the performance with interleaved reasoning and grounding steps during inference. Extensive experiments on the proposed benchmark validate the effectiveness of our proposed approach.
Authors: Liang Zeng, Liangjun Zhong, Liang Zhao, Tianwen Wei, Liu Yang, Jujie He, Cheng Cheng, Rui Hu, Yang Liu, Shuicheng Yan, Han Fang, Yahui Zhou
Abstract: In this paper, we investigate the underlying factors that potentially enhance the mathematical reasoning capabilities of large language models (LLMs). We argue that the data scaling law for math reasoning capabilities in modern LLMs is far from being saturated, highlighting how the model's quality improves with increases in data quantity. To support this claim, we introduce the Skywork-Math model series, supervised fine-tuned (SFT) on common 7B LLMs using our proposed 2.5M-instance Skywork-MathQA dataset. Skywork-Math 7B has achieved impressive accuracies of 51.2% on the competition-level MATH benchmark and 83.9% on the GSM8K benchmark using only SFT data, outperforming an early version of GPT-4 on MATH. The superior performance of Skywork-Math models contributes to our novel two-stage data synthesis and model SFT pipelines, which include three different augmentation methods and a diverse seed problem set, ensuring both the quantity and quality of Skywork-MathQA dataset across varying difficulty levels. Most importantly, we provide several practical takeaways to enhance math reasoning abilities in LLMs for both research and industry applications.
Authors: Samyak Jain, Ekdeep Singh Lubana, Kemal Oksuz, Tom Joy, Philip H. S. Torr, Amartya Sanyal, Puneet K. Dokania
Abstract: Safety fine-tuning helps align Large Language Models (LLMs) with human preferences for their safe deployment. To better understand the underlying factors that make models safe via safety fine-tuning, we design a synthetic data generation framework that captures salient aspects of an unsafe input by modeling the interaction between the task the model is asked to perform (e.g., "design") versus the specific concepts the task is asked to be performed upon (e.g., a "cycle" vs. a "bomb"). Using this, we investigate three well-known safety fine-tuning methods -- supervised safety fine-tuning, direct preference optimization, and unlearning -- and provide significant evidence demonstrating that these methods minimally transform MLP weights to specifically align unsafe inputs into its weights' null space. This yields a clustering of inputs based on whether the model deems them safe or not. Correspondingly, when an adversarial input (e.g., a jailbreak) is provided, its activations are closer to safer samples, leading to the model processing such an input as if it were safe. We validate our findings, wherever possible, on real-world models -- specifically, Llama-2 7B and Llama-3 8B.
Authors: Gon\c{c}alo Hora de Carvalho, Robert Pollice, Oscar Knap
Abstract: We explore the hypothesis that LLMs, such as GPT-3.5 and GPT-4, possess broader cognitive functions, particularly in non-linguistic domains. Our approach extends beyond standard linguistic benchmarks by incorporating games like Tic-Tac-Toe, Connect Four, and Battleship, encoded via ASCII, to assess strategic thinking and decision-making. To evaluate the models' ability to generalize beyond their training data, we introduce two additional games. The first game, LEGO Connect Language (LCL), tests the models' capacity to understand spatial logic and follow assembly instructions. The second game, the game of shapes, challenges the models to identify shapes represented by 1s within a matrix of zeros, further testing their spatial reasoning skills. This "show, don't tell" strategy uses games instead of simply querying the models. Our results show that despite their proficiency on standard benchmarks, GPT-3.5 and GPT-4's abilities to play and reason about fully observable games without pre-training is mediocre. Both models fail to anticipate losing moves in Tic-Tac-Toe and Connect Four, and they are unable to play Battleship correctly. While GPT-4 shows some success in the game of shapes, both models fail at the assembly tasks presented in the LCL game. These results suggest that while GPT models can emulate conversational proficiency and basic rule comprehension, their performance in strategic gameplay and spatial reasoning tasks is very limited. Importantly, this reveals a blind spot in current LLM benchmarks that we highlight with our gameplay benchmark suite ChildPlay (https://github.com/child-play-neurips/child-play). Our findings provide a cautionary tale about claims of emergent intelligence and reasoning capabilities of LLMs that are roughly the size of GPT-3.5 and GPT-4.
Authors: Xuhong Wang, Haoyu Jiang, Yi Yu, Jingru Yu, Yilun Lin, Ping Yi, Yingchun Wang, Qiao Yu, Li Li, Fei-Yue Wang
Abstract: Large Language Models (LLMs) are increasingly integrated into diverse industries, posing substantial security risks due to unauthorized replication and misuse. To mitigate these concerns, robust identification mechanisms are widely acknowledged as an effective strategy. Identification systems for LLMs now rely heavily on watermarking technology to manage and protect intellectual property and ensure data security. However, previous studies have primarily concentrated on the basic principles of algorithms and lacked a comprehensive analysis of watermarking theory and practice from the perspective of intelligent identification. To bridge this gap, firstly, we explore how a robust identity recognition system can be effectively implemented and managed within LLMs by various participants using watermarking technology. Secondly, we propose a mathematical framework based on mutual information theory, which systematizes the identification process to achieve more precise and customized watermarking. Additionally, we present a comprehensive evaluation of performance metrics for LLM watermarking, reflecting participant preferences and advancing discussions on its identification applications. Lastly, we outline the existing challenges in current watermarking technologies and theoretical frameworks, and provide directional guidance to address these challenges. Our systematic classification and detailed exposition aim to enhance the comparison and evaluation of various methods, fostering further research and development toward a transparent, secure, and equitable LLM ecosystem.
Authors: Kalliopi Basioti, Mohamed A. Abdelsalam, Federico Fancellu, Vladimir Pavlovic, Afsaneh Fazly
Abstract: Controllable Image Captioning (CIC) aims at generating natural language descriptions for an image, conditioned on information provided by end users, e.g., regions, entities or events of interest. However, available image-language datasets mainly contain captions that describe the entirety of an image, making them ineffective for training CIC models that can potentially attend to any subset of regions or relationships. To tackle this challenge, we propose a novel, fully automatic method to sample additional focused and visually grounded captions using a unified structured semantic representation built on top of the existing set of captions associated with an image. We leverage Abstract Meaning Representation (AMR), a cross-lingual graph-based semantic formalism, to encode all possible spatio-semantic relations between entities, beyond the typical spatial-relations-only focus of current methods. We use this Structured Semantic Augmentation (SSA) framework to augment existing image-caption datasets with the grounded controlled captions, increasing their spatial and semantic diversity and focal coverage. We then develop a new model, CIC-BART-SSA, specifically tailored for the CIC task, that sources its control signals from SSA-diversified datasets. We empirically show that, compared to SOTA CIC models, CIC-BART-SSA generates captions that are superior in diversity and text quality, are competitive in controllability, and, importantly, minimize the gap between broad and highly focused controlled captioning performance by efficiently generalizing to the challenging highly focused scenarios. Code is available at https://github.com/SamsungLabs/CIC-BART-SSA.
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.