Authors: Linxin Song, Jiale Liu, Jieyu Zhang, Shaokun Zhang, Ao Luo, Shijian Wang, Qingyun Wu, Chi Wang
Abstract: Leveraging multiple large language model (LLM) agents has shown to be a promising approach for tackling complex tasks, while the effective design of multiple agents for a particular application remains an art. It is thus intriguing to answer a critical question: Given a task, how can we build a team of LLM agents to solve it effectively? Our new adaptive team-building paradigm offers a flexible solution, realized through a novel agent design named Captain Agent. It dynamically forms and manages teams for each step of a task-solving process, utilizing nested group conversations and reflection to ensure diverse expertise and prevent stereotypical outputs. It allows for a flexible yet structured approach to problem-solving and can help reduce redundancy and enhance output diversity. A comprehensive evaluation across six real-world scenarios demonstrates that Captain Agent significantly outperforms existing multi-agent methods with 21.94% improvement in average accuracy, providing outstanding performance without requiring task-specific prompt engineering.
Authors: Mithilesh Vaidya, Binaya Kumar Sahoo, Preeti Rao
Abstract: Reading fluency assessment is a critical component of literacy programmes, serving to guide and monitor early education interventions. Given the resource intensive nature of the exercise when conducted by teachers, the development of automatic tools that can operate on audio recordings of oral reading is attractive as an objective and highly scalable solution. Multiple complex aspects such as accuracy, rate and expressiveness underlie human judgements of reading fluency. In this work, we investigate end-to-end modeling on a training dataset of children's audio recordings of story texts labeled by human experts. The pre-trained wav2vec2.0 model is adopted due its potential to alleviate the challenges from the limited amount of labeled data. We report the performance of a number of system variations on the relevant measures, and also probe the learned embeddings for lexical and acoustic-prosodic features known to be important to the perception of reading fluency.
Authors: Yupei Wang, Renfen Hu, Zhe Zhao
Abstract: While current automated essay scoring (AES) methods show high agreement with human raters, their scoring mechanisms are not fully explored. Our proposed method, using counterfactual intervention assisted by Large Language Models (LLMs), reveals that when scoring essays, BERT-like models primarily focus on sentence-level features, while LLMs are attuned to conventions, language complexity, as well as organization, indicating a more comprehensive alignment with scoring rubrics. Moreover, LLMs can discern counterfactual interventions during feedback. Our approach improves understanding of neural AES methods and can also apply to other domains seeking transparency in model-driven decisions. The codes and data will be released at GitHub.
Authors: Everlyn Asiko Chimoto, Jay Gala, Orevaoghene Ahia, Julia Kreutzer, Bruce A. Bassett, Sara Hooker
Abstract: Neural Machine Translation models are extremely data and compute-hungry. However, not all data points contribute equally to model training and generalization. Data pruning to remove the low-value data points has the benefit of drastically reducing the compute budget without significant drop in model performance. In this paper, we propose a new data pruning technique: Checkpoints Across Time (CAT), that leverages early model training dynamics to identify the most relevant data points for model performance. We benchmark CAT against several data pruning techniques including COMET-QE, LASER and LaBSE. We find that CAT outperforms the benchmarks on Indo-European languages on multiple test sets. When applied to English-German, English-French and English-Swahili translation tasks, CAT achieves comparable performance to using the full dataset, while pruning up to 50% of training data. We inspect the data points that CAT selects and find that it tends to favour longer sentences and sentences with unique or rare words.
Authors: Peng Wang, Songshuo Lu, Yaohua Tang, Sijie Yan, Yuanjun Xiong, Wei Xia
Abstract: We present a generative dialogue system capable of operating in a full-duplex manner, allowing for seamless interaction. It is based on a large language model (LLM) carefully aligned to be aware of a perception module, a motor function module, and the concept of a simple finite state machine (called neural FSM) with two states. The perception and motor function modules operate simultaneously, allowing the system to simultaneously speak and listen to the user. The LLM generates textual tokens for inquiry responses and makes autonomous decisions to start responding to, wait for, or interrupt the user by emitting control tokens to the neural FSM. All these tasks of the LLM are carried out as next token prediction on a serialized view of the dialogue in real-time. In automatic quality evaluations simulating real-life interaction, the proposed system reduces the average conversation response latency by more than 3 folds compared with LLM-based half-duplex dialogue systems while responding within less than 500 milliseconds in more than 50% of evaluated interactions. Running a LLM with only 8 billion parameters, our system exhibits a 8% higher interruption precision rate than the best available commercial LLM for voice-based dialogue.
Authors: Sudeshna Das, Yao Ge, Yuting Guo, Swati Rajwal, JaMor Hairston, Jeanne Powell, Drew Walker, Snigdha Peddireddy, Sahithi Lakamana, Selen Bozkurt, Matthew Reyna, Reza Sameni, Yunyu Xiao, Sangmi Kim, Rasheeta Chandler, Natalie Hernandez, Danielle Mowery, Rachel Wightman, Jennifer Love, Anthony Spadaro, Jeanmarie Perrone, Abeed Sarker
Abstract: Retrieval augmented generation (RAG) provides the capability to constrain generative model outputs, and mitigate the possibility of hallucination, by providing relevant in-context text. The number of tokens a generative large language model (LLM) can incorporate as context is finite, thus limiting the volume of knowledge from which to generate an answer. We propose a two-layer RAG framework for query-focused answer generation and evaluate a proof-of-concept for this framework in the context of query-focused summary generation from social media forums, focusing on emerging drug-related information. The evaluations demonstrate the effectiveness of the two-layer framework in resource constrained settings to enable researchers in obtaining near real-time data from users.
Authors: Pierre Chambon, Jean-Benoit Delbrouck, Thomas Sounack, Shih-Cheng Huang, Zhihong Chen, Maya Varma, Steven QH Truong, Chu The Chuong, Curtis P. Langlotz
Abstract: Since the release of the original CheXpert paper five years ago, CheXpert has become one of the most widely used and cited clinical AI datasets. The emergence of vision language models has sparked an increase in demands for sharing reports linked to CheXpert images, along with a growing interest among AI fairness researchers in obtaining demographic data. To address this, CheXpert Plus serves as a new collection of radiology data sources, made publicly available to enhance the scaling, performance, robustness, and fairness of models for all subsequent machine learning tasks in the field of radiology. CheXpert Plus is the largest text dataset publicly released in radiology, with a total of 36 million text tokens, including 13 million impression tokens. To the best of our knowledge, it represents the largest text de-identification effort in radiology, with almost 1 million PHI spans anonymized. It is only the second time that a large-scale English paired dataset has been released in radiology, thereby enabling, for the first time, cross-institution training at scale. All reports are paired with high-quality images in DICOM format, along with numerous image and patient metadata covering various clinical and socio-economic groups, as well as many pathology labels and RadGraph annotations. We hope this dataset will boost research for AI models that can further assist radiologists and help improve medical care. Data is available at the following URL: https://stanfordaimi.azurewebsites.net/datasets/5158c524-d3ab-4e02-96e9-6ee9efc110a1 Models are available at the following URL: https://github.com/Stanford-AIMI/chexpert-plus
URLs: https://stanfordaimi.azurewebsites.net/datasets/5158c524-d3ab-4e02-96e9-6ee9efc110a1, https://github.com/Stanford-AIMI/chexpert-plus
Authors: Michael Fore, Simranjit Singh, Chaehong Lee, Amritanshu Pandey, Antonios Anastasopoulos, Dimitrios Stamoulis
Abstract: Misinformation regarding climate change is a key roadblock in addressing one of the most serious threats to humanity. This paper investigates factual accuracy in large language models (LLMs) regarding climate information. Using true/false labeled Q&A data for fine-tuning and evaluating LLMs on climate-related claims, we compare open-source models, assessing their ability to generate truthful responses to climate change questions. We investigate the detectability of models intentionally poisoned with false climate information, finding that such poisoning may not affect the accuracy of a model's responses in other domains. Furthermore, we compare the effectiveness of unlearning algorithms, fine-tuning, and Retrieval-Augmented Generation (RAG) for factually grounding LLMs on climate change topics. Our evaluation reveals that unlearning algorithms can be effective for nuanced conceptual claims, despite previous findings suggesting their inefficacy in privacy contexts. These insights aim to guide the development of more factually reliable LLMs and highlight the need for additional work to secure LLMs against misinformation attacks.
Authors: Umar Ibrahim, Abubakar Yakubu Zandam, Fatima Muhammad Adam, Aminu Musa
Abstract: Aspect-based Sentiment Analysis (ABSA) is crucial for understanding sentiment nuances in text, especially across diverse languages and cultures. This paper introduces a novel Deep Convolutional Neural Network (CNN)-based model tailored for aspect and polarity classification in Hausa movie reviews, an underrepresented language in sentiment analysis research. A comprehensive Hausa ABSA dataset is created, filling a significant gap in resource availability. The dataset, preprocessed using sci-kit-learn for TF-IDF transformation, includes manually annotated aspect-level feature ontology words and sentiment polarity assignments. The proposed model combines CNNs with attention mechanisms for aspect-word prediction, leveraging contextual information and sentiment polarities. With 91% accuracy on aspect term extraction and 92% on sentiment polarity classification, the model outperforms traditional machine models, offering insights into specific aspects and sentiments. This study advances ABSA research, particularly in underrepresented languages, with implications for cross-cultural linguistic research.
Authors: Yao Yao, Zuchao Li, Hai Zhao
Abstract: The burgeoning size of Large Language Models (LLMs) has led to enhanced capabilities in generating responses, albeit at the expense of increased inference times and elevated resource demands. Existing methods of acceleration, predominantly hinged on knowledge distillation, generally necessitate fine-tuning of considerably large models, such as Llama-7B, posing a challenge for average users. Furthermore, present techniques for expediting inference and reducing costs operate independently. To address these issues, we introduce a novel and intuitive Guidance-based Knowledge Transfer (GKT) framework. This approach leverages a larger LLM as a ''teacher'' to create guidance prompts, paired with a smaller ''student'' model to finalize responses. Remarkably, GKT requires no fine-tuning and doesn't necessitate the teacher and student models to have the same vocabulary, allowing for extensive batch generation to accelerate the process while ensuring user customization. GKT can be seamlessly integrated into cloud-edge collaboration architectures, and is versatile enough for plug-and-play application across various models. It excels in both efficiency and affordability, epitomizing a ''cheap and cheerful'' solution. GKT achieves a maximum accuracy improvement of 14.18%, along with a 10.72 times speed-up on GSM8K and an accuracy improvement of 14.00 % along with a 7.73 times speed-up in CSQA. When utilizing ChatGPT as teacher model and Llama2-70B as the student model, we can achieve 95.00% of ChatGPT's performance at 52% of the cost. The results highlight substantial enhancements in accuracy and processing speed on the GSM8K and CSQA datasets, surpassing the performance of using either the student or teacher models in isolation.
Authors: Ernesto Quevedo, Jorge Yero, Rachel Koerner, Pablo Rivas, Tomas Cerny
Abstract: Concerns regarding the propensity of Large Language Models (LLMs) to produce inaccurate outputs, also known as hallucinations, have escalated. Detecting them is vital for ensuring the reliability of applications relying on LLM-generated content. Current methods often demand substantial resources and rely on extensive LLMs or employ supervised learning with multidimensional features or intricate linguistic and semantic analyses difficult to reproduce and largely depend on using the same LLM that hallucinated. This paper introduces a supervised learning approach employing two simple classifiers utilizing only four numerical features derived from tokens and vocabulary probabilities obtained from other LLM evaluators, which are not necessarily the same. The method yields promising results, surpassing state-of-the-art outcomes in multiple tasks across three different benchmarks. Additionally, we provide a comprehensive examination of the strengths and weaknesses of our approach, highlighting the significance of the features utilized and the LLM employed as an evaluator. We have released our code publicly at https://github.com/Baylor-AI/HalluDetect.
Authors: Ruiyi Wang, Stephanie Milani, Jamie C. Chiu, Shaun M. Eack, Travis Labrum, Samuel M. Murphy, Nev Jones, Kate Hardy, Hong Shen, Fei Fang, Zhiyu Zoey Chen
Abstract: Mental illness remains one of the most critical public health issues, with a significant gap between the available mental health support and patient needs. Many mental health professionals highlight a disconnect between their training and real-world patient interactions, leaving some trainees feeling unprepared and potentially affecting their early career success. In this paper, we propose PATIENT-{\Psi}, a novel patient simulation framework for cognitive behavior therapy (CBT) training. To build PATIENT-{\Psi}, we constructed diverse patient profiles and their corresponding cognitive models based on CBT principles, and then used large language models (LLMs) programmed with the patient cognitive models to act as a simulated therapy patient. We propose an interactive training scheme, PATIENT-{\Psi}-TRAINER, for mental health trainees to practice a key skill in CBT -- formulating the cognitive model of the patient -- through role-playing a therapy session with PATIENT-{\Psi}. To evaluate PATIENT-{\Psi}, we conducted a user study of 4 mental health trainees and 10 experts. The results demonstrate that practice using PATIENT-{\Psi}-TRAINER greatly enhances the perceived skill acquisition and confidence of the trainees beyond existing forms of training such as textbooks, videos, and role-play with non-patients. Based on the experts' perceptions, PATIENT-{\Psi} is perceived to be closer to real patient interactions than GPT-4, and PATIENT-{\Psi}-TRAINER holds strong promise to improve trainee competencies. Our pioneering patient simulation training framework, using LLMs, holds great potential to enhance and advance mental health training, ultimately leading to improved patient care and outcomes. We will release all our data, code, and the training platform.
Authors: Yutao Zhu, Zhaoheng Huang, Zhicheng Dou, Ji-Rong Wen
Abstract: Retrieval-augmented generation (RAG) is a promising way to improve large language models (LLMs) for generating more factual, accurate, and up-to-date content. Existing methods either optimize prompts to guide LLMs in leveraging retrieved information or directly fine-tune the LLMs to adapt to RAG scenarios. Although fine-tuning can yield better performance, it often compromises the LLMs' general generation capabilities by modifying their parameters. This limitation poses challenges in practical applications, especially when LLMs are already deployed, as parameter adjustments may affect their original functionality. To address this, we propose a novel method that involves learning scalable and pluggable virtual tokens for RAG. By maintaining the LLMs' original parameters and fine-tuning only the embeddings of these pluggable tokens, our approach not only enhances LLMs' performance but also preserves their general generation capacities. Furthermore, we design several training strategies to improve the scalability, flexibility, and generalizability of our method. Comprehensive experiments across nine question-answering tasks demonstrate the superiority of our approach.
Authors: Lavanya Prahallad, Radhika Mamidi
Abstract: Gender bias in machine translation (MT) systems poses a significant challenge to achieving accurate and inclusive translations. This paper examines gender bias in machine translation systems for languages such as Telugu and Kannada from the Dravidian family, analyzing how gender inflections affect translation accuracy and neutrality using Google Translate and ChatGPT. It finds that while plural forms can reduce bias, individual-centric sentences often maintain the bias due to historical stereotypes. The study evaluates the Chain of Thought processing, noting significant bias mitigation from 80% to 4% in Telugu and from 40% to 0% in Kannada. It also compares Telugu and Kannada translations, emphasizing the need for language specific strategies to address these challenges and suggesting directions for future research to enhance fairness in both data preparation and prompts during inference.
Authors: Kaixuan Huang, Xudong Guo, Mengdi Wang
Abstract: Speculative decoding reduces the inference latency of a target large language model via utilizing a smaller and faster draft model. Its performance depends on a hyperparameter K -- the candidate length, i.e., the number of candidate tokens for the target model to verify in each round. However, previous methods often use simple heuristics to choose K, which may result in sub-optimal performance. We study the choice of the candidate length K and formulate it as a Markov Decision Process. We theoretically show that the optimal policy of this Markov decision process takes the form of a threshold policy, i.e., the current speculation should stop and be verified when the probability of getting a rejection exceeds a threshold value. Motivated by this theory, we propose SpecDec++, an enhanced version of speculative decoding that adaptively determines the candidate length on the fly. We augment the draft model with a trained acceptance prediction head to predict the conditional acceptance probability of the candidate tokens. SpecDec++ will stop the current speculation when the predicted probability that at least one token gets rejected exceeds a threshold. We implement SpecDec++ and apply it to the llama-2-chat 7B & 70B model pair. Our adaptive method achieves a 2.04x speedup on the Alpaca dataset (an additional 7.2% improvement over the baseline speculative decoding). On the GSM8K and HumanEval datasets, our method achieves a 2.26x speedup (9.4% improvement) and 2.23x speedup (11.1% improvement), respectively.
Authors: Chengwei Dai, Kun Li, Wei Zhou, Songlin Hu
Abstract: As Large Language Models (LLMs) scale up and gain powerful Chain-of-Thoughts (CoTs) reasoning abilities, practical resource constraints drive efforts to distill these capabilities into more compact Smaller Language Models (SLMs). We find that CoTs consist mainly of simple reasoning forms, with a small proportion ($\approx 4.7\%$) of key reasoning steps that truly impact conclusions. However, previous distillation methods typically involve supervised fine-tuning student SLMs only on correct CoTs data produced by teacher LLMs, resulting in students struggling to learn the key reasoning steps, instead imitating the teacher's reasoning forms and making errors or omissions on these steps. To address these issues, drawing an analogy to human learning, where analyzing mistakes according to correct solutions often reveals the crucial steps leading to successes or failures, we propose mistak\textbf{E}-\textbf{D}riven key reason\textbf{I}ng step distilla\textbf{T}ion (\textbf{EDIT}), a novel method that further aids SLMs learning key reasoning steps rather than mere simple fine-tuning. Firstly, to expose these crucial steps in CoTs, we design specific prompts to generate dual CoTs data with similar reasoning paths but divergent conclusions. Then, we apply the minimum edit distance algorithm on the dual CoTs data to locate these key steps and optimize the likelihood of these steps. Extensive experiments validate the effectiveness of EDIT across both in-domain and out-of-domain benchmark reasoning datasets. Further analysis shows that EDIT can generate high-quality CoTs with more correct key reasoning steps. Notably, we also explore how different mistake patterns affect performance and find that EDIT benefits more from logical errors than from knowledge or mathematical calculation errors in dual CoTs\footnote{Code can be found at \url{https://github.com/C-W-D/EDIT}}.
Authors: Jiatong Li, Renjun Hu, Kunzhe Huang, Yan Zhuang, Qi Liu, Mengxiao Zhu, Xing Shi, Wei Lin
Abstract: Expert-designed close-ended benchmarks serve as vital tools in assessing the knowledge capacity of large language models (LLMs). Despite their widespread use, concerns have mounted regarding their reliability due to limited test scenarios and an unavoidable risk of data contamination. To rectify this, we present PertEval, a toolkit devised for in-depth probing of LLMs' knowledge capacity through knowledge-invariant perturbations. These perturbations employ human-like restatement techniques to generate on-the-fly test samples from static benchmarks, meticulously retaining knowledge-critical content while altering irrelevant details. Our toolkit further includes a suite of transition analyses that compare performance on raw vs. perturbed test sets to precisely assess LLMs' genuine knowledge capacity. Six state-of-the-art LLMs are re-evaluated using PertEval. Results reveal significantly inflated performance of the LLMs on raw benchmarks, including an absolute 21% overestimation for GPT-4. Additionally, through a nuanced response pattern analysis, we discover that PertEval retains LLMs' uncertainty to specious knowledge, potentially being resolved through rote memorization and leading to inflated performance. We also find that the detailed transition analyses by PertEval could illuminate weaknesses in existing LLMs' knowledge mastery and guide the development of refinement. Given these insights, we posit that PertEval can act as an essential tool that, when applied alongside any close-ended benchmark, unveils the true knowledge capacity of LLMs, marking a significant step toward more trustworthy LLM evaluation.
Authors: Chong Li, Wen Yang, Jiajun Zhang, Jinliang Lu, Shaonan Wang, Chengqing Zong
Abstract: Large language models respond well in high-resource languages like English but struggle in low-resource languages. It may arise from the lack of high-quality instruction following data in these languages. Directly translating English samples into these languages can be a solution but unreliable, leading to responses with translation errors and lacking language-specific or cultural knowledge. To address this issue, we propose a novel method to construct cross-lingual instruction following samples with instruction in English and response in low-resource languages. Specifically, the language model first learns to generate appropriate English instructions according to the natural web texts in other languages as responses. The candidate cross-lingual instruction tuning samples are further refined and diversified. We have employed this method to build a large-scale cross-lingual instruction tuning dataset on 10 languages, namely X-Instruction. The instruction data built using our method incorporate more language-specific knowledge compared with the naive translation method. Experimental results have shown that the response quality of the model tuned on X-Instruction greatly exceeds the model distilled from a powerful teacher model, reaching or even surpassing the ones of ChatGPT. In addition, we find that models tuned on cross-lingual instruction following samples can follow the instruction in the output language without further tuning.
Authors: Kuo Liao, Shuang Li, Meng Zhao, Liqun Liu, Mengge Xue, Zhenyu Hu, Honglin Han, Chengguo Yin
Abstract: Recent strides in large language models (LLMs) have yielded remarkable performance, leveraging reinforcement learning from human feedback (RLHF) to significantly enhance generation and alignment capabilities. However, RLHF encounters numerous challenges, including the objective mismatch issue, leading to suboptimal performance in Natural Language Understanding (NLU) tasks. To address this limitation, we propose a novel Reinforcement Learning framework enhanced with Label-sensitive Reward (RLLR) to amplify the performance of LLMs in NLU tasks. By incorporating label-sensitive pairs into reinforcement learning, our method aims to adeptly capture nuanced label-sensitive semantic features during RL, thereby enhancing natural language understanding. Experiments conducted on five diverse foundation models across eight tasks showcase promising results. In comparison to Supervised Fine-tuning models (SFT), RLLR demonstrates an average performance improvement of 1.54%. Compared with RLHF models, the improvement averages at 0.69%. These results reveal the effectiveness of our method for LLMs in NLU tasks. Code and data available at: https://github.com/MagiaSN/ACL2024_RLLR.
Authors: Jeiyoon Park, Chanjun Park, Heuiseok Lim
Abstract: With the recent introduction of Assistants API, it is expected that document-based language models will be actively used in various domains, especially Role-playing. However, a key challenge lies in utilizing protagonist's persona: Assistants API often fails to achieve with its search because the information extraction part is different each time and it often omits important information such as protagonist's backstory or relationships. It is hard to maintain a consistent persona simply by using the persona document as input to the Assistants API. To address the challenge of achieving stable persona consistency, we propose CharacterGPT, a novel persona reconstruction framework to alleviate the shortcomings of the Assistants API. Our method involves Character Persona Training (CPT), an effective persona rebuilding process that updates the character persona by extracting the character's traits from given summary of the novel for each character as if the story in a novel progresses. In our experiments, we ask each character to take the Big Five Inventory personality test in various settings and analyze the results. To assess whether it can think outside the box, we let each character generate short novels. Extensive experiments and human evaluation demonstrate that CharacterGPT presents new possibilities for role-playing agent research.
Authors: Dylan Zhang, Justin Wang, Francois Charton
Abstract: Instruction tuning -- tuning large language models on instruction-output pairs -- is a promising technique for making models better adapted to the real world. Yet, the key factors driving the model's capability to understand and follow instructions not seen during training remain under-explored. Our investigation begins with a series of synthetic experiments within the theoretical framework of a Turing-complete algorithm called Markov algorithm, which allows fine-grained control over the instruction-tuning data. Generalization and robustness with respect to the training distribution emerge once a diverse enough set of tasks is provided, even though very few examples are provided for each task. We extend these initial results to a real-world application scenario of code generation and find that a more diverse instruction set, extending beyond code-related tasks, improves the performance of code generation. Our observations suggest that a more diverse semantic space for instruction-tuning sets greatly improves the model's ability to follow instructions and perform tasks.
Authors: Li Zhang, Peter Jansen, Tianyi Zhang, Peter Clark, Chris Callison-Burch, Niket Tandon
Abstract: Planning in textual environments have been shown to be a long-standing challenge even for current models. A recent, promising line of work uses LLMs to generate a formal representation of the environment that can be solved by a symbolic planner. However, existing methods rely on a fully-observed environment where all entity states are initially known, so a one-off representation can be constructed, leading to a complete plan. In contrast, we tackle partially-observed environments where there is initially no sufficient information to plan for the end-goal. We propose PDDLEGO that iteratively construct a planning representation that can lead to a partial plan for a given sub-goal. By accomplishing the sub-goal, more information is acquired to augment the representation, eventually achieving the end-goal. We show that plans produced by few-shot PDDLEGO are 43% more efficient than generating plans end-to-end on the Coin Collector simulation, with strong performance (98%) on the more complex Cooking World simulation where end-to-end LLMs fail to generate coherent plans (4%).
Authors: Ohjoon Kwon, Donghyeon Jeon, Nayoung Choi, Gyu-Hwung Cho, Changbong Kim, Hyunwoo Lee, Inho Kang, Sun Kim, Taiwoo Park
Abstract: Most prior safety research of large language models (LLMs) has focused on enhancing the alignment of LLMs to better suit the safety requirements of humans. However, internalizing such safeguard features into larger models brought challenges of higher training cost and unintended degradation of helpfulness. To overcome such challenges, a modular approach employing a smaller LLM to detect harmful user queries is regarded as a convenient solution in designing LLM-based system with safety requirements. In this paper, we leverage a smaller LLM for both harmful query detection and safeguard response generation. We introduce our safety requirements and the taxonomy of harmfulness categories, and then propose a multi-task learning mechanism fusing the two tasks into a single model. We demonstrate the effectiveness of our approach, providing on par or surpassing harmful query detection and safeguard response performance compared to the publicly available LLMs.
Authors: Jiahui Xu, Feng Jiang, Anningzhe Gao, Haizhou Li
Abstract: The advancement of large language models (LLMs) has propelled the development of dialogue systems. Unlike the popular ChatGPT-like assistant model, which only satisfies the user's preferences, task-oriented dialogue systems have also faced new requirements and challenges in the broader business field. They are expected to provide correct responses at each dialogue turn, at the same time, achieve the overall goal defined by the task. By understanding rhetorical structures and topic structures via topic segmentation and discourse parsing, a dialogue system may do a better planning to achieve both objectives. However, while both structures belong to discourse structure in linguistics, rhetorical structure and topic structure are mostly modeled separately or with one assisting the other in the prior work. The interaction between these two structures has not been considered for joint modeling and mutual learning. Furthermore, unsupervised learning techniques to achieve the above are not well explored. To fill this gap, we propose an unsupervised mutual learning framework of two structures leveraging the global and local connections between them. We extend the topic modeling between non-adjacent discourse units to ensure global structural relevance with rhetorical structures. We also incorporate rhetorical structures into the topic structure through a graph neural network model to ensure local coherence consistency. Finally, we utilize the similarity between the two fused structures for mutual learning. The experimental results demonstrate that our methods outperform all strong baselines on two dialogue rhetorical datasets (STAC and Molweni), as well as dialogue topic datasets (Doc2Dial and TIAGE).
Authors: Stephen Meisenbacher, Florian Matthes
Abstract: The study of Differential Privacy (DP) in Natural Language Processing often views the task of text privatization as a $\textit{rewriting}$ task, in which sensitive input texts are rewritten to hide explicit or implicit private information. In order to evaluate the privacy-preserving capabilities of a DP text rewriting mechanism, $\textit{empirical privacy}$ tests are frequently employed. In these tests, an adversary is modeled, who aims to infer sensitive information (e.g., gender) about the author behind a (privatized) text. Looking to improve the empirical protections provided by DP rewriting methods, we propose a simple post-processing method based on the goal of aligning rewritten texts with their original counterparts, where DP rewritten texts are rewritten $\textit{again}$. Our results shown that such an approach not only produces outputs that are more semantically reminiscent of the original inputs, but also texts which score on average better in empirical privacy evaluations. Therefore, our approach raises the bar for DP rewriting methods in their empirical privacy evaluations, providing an extra layer of protection against malicious adversaries.
Authors: Chengwei Dai, Kun Li, Wei Zhou, Songlin Hu
Abstract: Large language models (LLMs) exhibit enhanced reasoning at larger scales, driving efforts to distill these capabilities into smaller models via teacher-student learning. Previous works simply fine-tune student models on teachers' generated Chain-of-Thoughts (CoTs) data. Although these methods enhance in-domain (IND) reasoning performance, they struggle to generalize to out-of-domain (OOD) tasks. We believe that the widespread spurious correlations between questions and answers may lead the model to preset a specific answer which restricts the diversity and generalizability of its reasoning process. In this paper, we propose Cascading Decomposed CoTs Distillation (CasCoD) to address these issues by decomposing the traditional single-step learning process into two cascaded learning steps. Specifically, by restructuring the training objectives -- removing the answer from outputs and concatenating the question with the rationale as input -- CasCoD's two-step learning process ensures that students focus on learning rationales without interference from the preset answers, thus improving reasoning generalizability. Extensive experiments demonstrate the effectiveness of CasCoD on both IND and OOD benchmark reasoning datasets. Code can be found at https://github.com/C-W-D/CasCoD.
Authors: Chaochen Gao, Xing Wu, Qi Fu, Songlin Hu
Abstract: Large language models, initially pre-trained with a limited context length, can better handle longer texts by continuing training on a corpus with extended contexts. However, obtaining effective long-context data is challenging due to the scarcity and uneven distribution of long documents across different domains. To address this issue, we propose a Query-centric data synthesis method, abbreviated as Quest. Quest is an interpretable method based on the observation that documents retrieved by similar queries are relevant but low-redundant, thus well-suited for synthesizing long-context data. The method is also scalable and capable of constructing large amounts of long-context data. Using Quest, we synthesize a long-context dataset up to 128k context length, significantly outperforming other data synthesis methods on multiple long-context benchmark datasets. In addition, we further verify that the Quest method is predictable through scaling law experiments, making it a reliable solution for advancing long-context models.
Authors: Jia Li, Ge Li, Yunfei Zhao, Yongmin Li, Huanyu Liu, Hao Zhu, Lecheng Wang, Kaibo Liu, Zheng Fang, Lanshen Wang, Jiazheng Ding, Xuanming Zhang, Yuqi Zhu, Yihong Dong, Zhi Jin, Binhua Li, Fei Huang, Yongbin Li
Abstract: How to evaluate the coding abilities of Large Language Models (LLMs) remains an open question. We find that existing benchmarks are poorly aligned with real-world code repositories and are insufficient to evaluate the coding abilities of LLMs. To address the knowledge gap, we propose a new benchmark named DevEval, which has three advances. (1) DevEval aligns with real-world repositories in multiple dimensions, e.g., code distributions and dependency distributions. (2) DevEval is annotated by 13 developers and contains comprehensive annotations (e.g., requirements, original repositories, reference code, and reference dependencies). (3) DevEval comprises 1,874 testing samples from 117 repositories, covering 10 popular domains (e.g., Internet, Database). Based on DevEval, we propose repository-level code generation and evaluate 8 popular LLMs on DevEval (e.g., gpt-4, gpt-3.5, StarCoder 2, DeepSeek Coder, CodeLLaMa). Our experiments reveal these LLMs' coding abilities in real-world code repositories. For example, in our experiments, the highest Pass@1 of gpt-4-turbo is only 53.04%. We also analyze LLMs' failed cases and summarize their shortcomings. We hope DevEval can facilitate the development of LLMs in real code repositories. DevEval, prompts, and LLMs' predictions have been released.
Authors: Hao Zhao, Maksym Andriushchenko, Francesco Croce, Nicolas Flammarion
Abstract: In-context learning (ICL) allows LLMs to learn from examples without changing their weights, which is a particularly promising capability for long-context LLMs that can potentially learn from many examples. Recently, Lin et al. (2024) proposed URIAL, a method using only three in-context examples to align base LLMs, achieving non-trivial instruction following performance. In this work, we show that, while effective, ICL alignment with URIAL still underperforms compared to instruction fine-tuning on established benchmarks such as MT-Bench and AlpacaEval 2.0 (LC), especially with more capable base LMs. Unlike for tasks such as classification, translation, or summarization, adding more ICL demonstrations for long-context LLMs does not systematically improve instruction following performance. To address this limitation, we derive a greedy selection approach for ICL examples that noticeably improves performance, yet without bridging the gap to instruction fine-tuning. Finally, we provide a series of ablation studies to better understand the reasons behind the remaining gap, and we show how some aspects of ICL depart from the existing knowledge and are specific to the instruction tuning setting. Overall, our work advances the understanding of ICL as an alignment technique. We provide our code at https://github.com/tml-epfl/icl-alignment.
Authors: Yi Liu, Xiangyu Liu, Xiangrong Zhu, Wei Hu
Abstract: Multi-aspect controllable text generation aims to control the generated texts in attributes from multiple aspects (e.g., "positive" from sentiment and "sport" from topic). For ease of obtaining training samples, existing works neglect attribute correlations formed by the intertwining of different attributes. Particularly, the stereotype formed by imbalanced attribute correlations significantly affects multi-aspect control. In this paper, we propose MAGIC, a new multi-aspect controllable text generation method with disentangled counterfactual augmentation. We alleviate the issue of imbalanced attribute correlations during training using counterfactual feature vectors in the attribute latent space by disentanglement. During inference, we enhance attribute correlations by target-guided counterfactual augmentation to further improve multi-aspect control. Experiments show that MAGIC outperforms state-of-the-art baselines in both imbalanced and balanced attribute correlation scenarios. Our source code and data are available at https://github.com/nju-websoft/MAGIC.
Authors: Hossam M. Zawbaa, Wael Rashwan, Sourav Dutta, Haytham Assem
Abstract: Detecting out-of-scope user utterances is essential for task-oriented dialogues and intent classification. Current methodologies face difficulties with the unpredictable distribution of outliers and often rely on assumptions about data distributions. We present the Dual Encoder for Threshold-Based Re-Classification (DETER) to address these challenges. This end-to-end framework efficiently detects out-of-scope intents without requiring assumptions on data distributions or additional post-processing steps. The core of DETER utilizes dual text encoders, the Universal Sentence Encoder (USE) and the Transformer-based Denoising AutoEncoder (TSDAE), to generate user utterance embeddings, which are classified through a branched neural architecture. Further, DETER generates synthetic outliers using self-supervision and incorporates out-of-scope phrases from open-domain datasets. This approach ensures a comprehensive training set for out-of-scope detection. Additionally, a threshold-based re-classification mechanism refines the model's initial predictions. Evaluations on the CLINC-150, Stackoverflow, and Banking77 datasets demonstrate DETER's efficacy. Our model outperforms previous benchmarks, increasing up to 13% and 5% in F1 score for known and unknown intents on CLINC-150 and Stackoverflow, and 16% for known and 24% % for unknown intents on Banking77. The source code has been released at https://github.com/Hossam-Mohammed-tech/Intent\_Classification\_OOS.
URLs: https://github.com/Hossam-Mohammed-tech/Intent\_Classification\_OOS.
Authors: Avelina Asada Hadji-Kyriacou, Ognjen Arandjelovic
Abstract: Pre-trained Language Models (LMs) exhibit strong zero-shot and in-context learning capabilities; however, their behaviors are often difficult to control. By utilizing Reinforcement Learning from Human Feedback (RLHF), it is possible to fine-tune unsupervised LMs to follow instructions and produce outputs that reflect human preferences. Despite its benefits, RLHF has been shown to potentially harm a language model's reasoning capabilities and introduce artifacts such as hallucinations where the model may fabricate facts. To address this issue we introduce Direct Preference Heads (DPH), a fine-tuning framework that enables LMs to learn human preference signals through an auxiliary reward head without directly affecting the output distribution of the language modeling head. We perform a theoretical analysis of our objective function and find strong ties to Conservative Direct Preference Optimization (cDPO). Finally we evaluate our models on GLUE, RACE, and the GPT4All evaluation suite and demonstrate that our method produces models which achieve higher scores than those fine-tuned with Supervised Fine-Tuning (SFT) or Direct Preference Optimization (DPO) alone.
Authors: Elena Grazia Gado, Tommaso Martorella, Luca Zunino, Paola Mejia-Domenzain, Vinitra Swamy, Jibril Frej, Tanja K\"aser
Abstract: Intelligent Tutoring Systems (ITS) enhance personalized learning by predicting student answers to provide immediate and customized instruction. However, recent research has primarily focused on the correctness of the answer rather than the student's performance on specific answer choices, limiting insights into students' thought processes and potential misconceptions. To address this gap, we present MCQStudentBert, an answer forecasting model that leverages the capabilities of Large Language Models (LLMs) to integrate contextual understanding of students' answering history along with the text of the questions and answers. By predicting the specific answer choices students are likely to make, practitioners can easily extend the model to new answer choices or remove answer choices for the same multiple-choice question (MCQ) without retraining the model. In particular, we compare MLP, LSTM, BERT, and Mistral 7B architectures to generate embeddings from students' past interactions, which are then incorporated into a finetuned BERT's answer-forecasting mechanism. We apply our pipeline to a dataset of language learning MCQ, gathered from an ITS with over 10,000 students to explore the predictive accuracy of MCQStudentBert, which incorporates student interaction patterns, in comparison to correct answer prediction and traditional mastery-learning feature-based approaches. This work opens the door to more personalized content, modularization, and granular support.
Authors: David Stap, Eva Hasler, Bill Byrne, Christof Monz, Ke Tran
Abstract: Fine-tuning large language models (LLMs) for machine translation has shown improvements in overall translation quality. However, it is unclear what is the impact of fine-tuning on desirable LLM behaviors that are not present in neural machine translation models, such as steerability, inherent document-level translation abilities, and the ability to produce less literal translations. We perform an extensive translation evaluation on the LLaMA and Falcon family of models with model size ranging from 7 billion up to 65 billion parameters. Our results show that while fine-tuning improves the general translation quality of LLMs, several abilities degrade. In particular, we observe a decline in the ability to perform formality steering, to produce technical translations through few-shot examples, and to perform document-level translation. On the other hand, we observe that the model produces less literal translations after fine-tuning on parallel data. We show that by including monolingual data as part of the fine-tuning data we can maintain the abilities while simultaneously enhancing overall translation quality. Our findings emphasize the need for fine-tuning strategies that preserve the benefits of LLMs for machine translation.
Authors: Jingchang Chen, Hongxuan Tang, Zheng Chu, Qianglong Chen, Zekun Wang, Ming Liu, Bing Qin
Abstract: Despite recent progress made by large language models in code generation, they still struggle with programs that meet complex requirements. Recent work utilizes plan-and-solve decomposition to decrease the complexity and leverage self-tests to refine the generated program. Yet, planning deep-inside requirements in advance can be challenging, and the tests need to be accurate to accomplish self-improvement. To this end, we propose FunCoder, a code generation framework incorporating the divide-and-conquer strategy with functional consensus. Specifically, FunCoder recursively branches off sub-functions as smaller goals during code generation, represented by a tree hierarchy. These sub-functions are then composited to attain more complex objectives. Additionally, we designate functions via a consensus formed by identifying similarities in program behavior, mitigating error propagation. FunCoder outperforms state-of-the-art methods by +9.8% on average in HumanEval, MBPP, xCodeEval and MATH with GPT-3.5 and GPT-4. Moreover, our method demonstrates superiority on smaller models: With FunCoder, StableCode-3b surpasses GPT-3.5 by +18.6% and achieves 97.7% of GPT-4's performance on HumanEval. Further analysis reveals that our proposed dynamic function decomposition is capable of handling complex requirements, and the functional consensus prevails over self-testing in correctness evaluation.
Authors: Yingshan Chang, Yonatan Bisk
Abstract: Counting is a fundamental example of generalization, whether viewed through the mathematical lens of Peano's axioms defining the natural numbers or the cognitive science literature for children learning to count. The argument holds for both cases that learning to count means learning to count infinitely. While few papers have tried to distill transformer "reasoning" to the simplest case of counting, investigating length generalization does occur throughout the literature. In the "train short, test long" paradigm of NLP, length refers to the training sentence length. In formal language recognition, length refers to the input sequence length, or the maximum stack size induced by a pushdown automata. In general problem solving, length refers to the number of hops in a deductive reasoning chain or the recursion depth. For all cases, counting is central to task success. And crucially, generalizing counting inductively is central to success on OOD instances. This work provides extensive empirical results on training language models to count. We experiment with architectures ranging from RNNs, Transformers, State-Space Models and RWKV. We present carefully-designed task formats, auxiliary tasks and positional embeddings to avoid limitations in generalization with OOD-position and OOD-vocabulary. We find that while traditional RNNs trivially achieve inductive counting, Transformers have to rely on positional embeddings to count out-of-domain. As counting is the basis for many arguments concerning the expressivity of Transformers, our finding calls for the community to reexamine the application scope of primitive functions defined in formal characterizations. Finally, modern RNNs also largely underperform traditional RNNs in generalizing counting inductively. We discuss how design choices that enable parallelized training of modern RNNs cause them to lose merits of a recurrent nature.
Authors: Costas Mavromatis, George Karypis
Abstract: Knowledge Graphs (KGs) represent human-crafted factual knowledge in the form of triplets (head, relation, tail), which collectively form a graph. Question Answering over KGs (KGQA) is the task of answering natural questions grounding the reasoning to the information provided by the KG. Large Language Models (LLMs) are the state-of-the-art models for QA tasks due to their remarkable ability to understand natural language. On the other hand, Graph Neural Networks (GNNs) have been widely used for KGQA as they can handle the complex graph information stored in the KG. In this work, we introduce GNN-RAG, a novel method for combining language understanding abilities of LLMs with the reasoning abilities of GNNs in a retrieval-augmented generation (RAG) style. First, a GNN reasons over a dense KG subgraph to retrieve answer candidates for a given question. Second, the shortest paths in the KG that connect question entities and answer candidates are extracted to represent KG reasoning paths. The extracted paths are verbalized and given as input for LLM reasoning with RAG. In our GNN-RAG framework, the GNN acts as a dense subgraph reasoner to extract useful graph information, while the LLM leverages its natural language processing ability for ultimate KGQA. Furthermore, we develop a retrieval augmentation (RA) technique to further boost KGQA performance with GNN-RAG. Experimental results show that GNN-RAG achieves state-of-the-art performance in two widely used KGQA benchmarks (WebQSP and CWQ), outperforming or matching GPT-4 performance with a 7B tuned LLM. In addition, GNN-RAG excels on multi-hop and multi-entity questions outperforming competing approaches by 8.9--15.5% points at answer F1.
Authors: Frederick Riemenschneider, Kevin Krahn
Abstract: Historical languages present unique challenges to the NLP community, with one prominent hurdle being the limited resources available in their closed corpora. This work describes our submission to the constrained subtask of the SIGTYP 2024 shared task, focusing on PoS tagging, morphological tagging, and lemmatization for 13 historical languages. For PoS and morphological tagging we adapt a hierarchical tokenization method from Sun et al. (2023) and combine it with the advantages of the DeBERTa-V3 architecture, enabling our models to efficiently learn from every character in the training data. We also demonstrate the effectiveness of character-level T5 models on the lemmatization task. Pre-trained from scratch with limited data, our models achieved first place in the constrained subtask, nearly reaching the performance levels of the unconstrained task's winner. Our code is available at https://github.com/bowphs/SIGTYP-2024-hierarchical-transformers
URLs: https://github.com/bowphs/SIGTYP-2024-hierarchical-transformers
Authors: Rosario Uceda-Sosa, Karthikeyan Natesan Ramamurthy, Maria Chang, Moninder Singh
Abstract: The ability to summarize and organize knowledge into abstract concepts is key to learning and reasoning. Many industrial applications rely on the consistent and systematic use of concepts, especially when dealing with decision-critical knowledge. However, we demonstrate that, when methodically questioned, large language models (LLMs) often display and demonstrate significant inconsistencies in their knowledge. Computationally, the basic aspects of the conceptualization of a given domain can be represented as Is-A hierarchies in a knowledge graph (KG) or ontology, together with a few properties or axioms that enable straightforward reasoning. We show that even simple ontologies can be used to reveal conceptual inconsistencies across several LLMs. We also propose strategies that domain experts can use to evaluate and improve the coverage of key domain concepts in LLMs of various sizes. In particular, we have been able to significantly enhance the performance of LLMs of various sizes with openly available weights using simple knowledge-graph (KG) based prompting strategies.
Authors: Kuang-Ming Chen, Hung-yi Lee
Abstract: The rapid development of large language models (LLMs) in recent years has largely focused on English, resulting in models that respond exclusively in English. To adapt these models to other languages, continual pre-training (CP) is often employed, followed by supervised fine-tuning (SFT) to maintain conversational abilities. However, CP and SFT can reduce a model's ability to filter harmful content. We propose Instruction Continual Pre-training (InsCP), which integrates instruction tags into the CP process to prevent loss of conversational proficiency while acquiring new languages. Our experiments demonstrate that InsCP retains conversational and Reinforcement Learning from Human Feedback (RLHF) abilities. Empirical evaluations on language alignment, reliability, and knowledge benchmarks confirm the efficacy of InsCP. Notably, this approach requires only 0.1 billion tokens of high-quality instruction-following data, thereby reducing resource consumption.
Authors: Zichao Hu, Junyi Jessy Li, Arjun Guha, Joydeep Biswas
Abstract: Large language models (LLMs) have shown great promise at generating robot programs from natural language given domain-specific robot application programming interfaces (APIs). However, the performance gap between proprietary LLMs and smaller open-weight LLMs remains wide. This raises a question: Can we fine-tune smaller open-weight LLMs for generating domain-specific robot programs to close the performance gap with proprietary LLMs? While Self-Instruct is a promising solution by generating a diverse set of training data, it cannot verify the correctness of these programs. In contrast, a robot simulator with a well-defined world can identify execution errors but limits the diversity of programs that it can verify. In this work, we introduce Robo-Instruct, which brings the best of both worlds -- it promotes the diversity of Self-Instruct while providing the correctness of simulator-based checking. Robo-Instruct introduces RoboSim to synthesize a consistent world state on the fly by inferring properties relevant to the program being checked, and simulating actions accordingly. Furthermore, the instructions and programs generated by Self-Instruct may be subtly inconsistent -- such as the program missing a step implied by the instruction. Robo-Instruct further addresses this with InstAlign, an instruction-program alignment procedure that revises the task instruction to reflect the actual results of the generated program. Given a few seed task descriptions and the robot APIs, Robo-Instruct is capable of generating a training dataset using only a small open-weight model. This dataset can then be used to fine-tune small open-weight language models, enabling them to match or even exceed the performance of several proprietary LLMs, such as GPT-3.5-Turbo and Gemini-Pro.
Authors: Shuyang Jiang, Yusheng Liao, Ya Zhang, Yu Wang, Yanfeng Wang
Abstract: Fine-tuning on task-specific question-answer pairs is a predominant method for enhancing the performance of instruction-tuned large language models (LLMs) on downstream tasks. However, in certain specialized domains, such as healthcare or harmless content generation, it is nearly impossible to obtain a large volume of high-quality data that matches the downstream distribution. To improve the performance of LLMs in data-scarce domains with domain-mismatched data, we re-evaluated the Transformer architecture and discovered that not all parameter updates during fine-tuning contribute positively to downstream performance. Our analysis reveals that within the self-attention and feed-forward networks, only the fine-tuned attention parameters are particularly beneficial when the training set's distribution does not fully align with the test set. Based on this insight, we propose an effective inference-time intervention method: \uline{T}raining \uline{A}ll parameters but \uline{I}nferring with only \uline{A}ttention (\trainallInfAttn). We empirically validate \trainallInfAttn using two general instruction-tuning datasets and evaluate it on seven downstream tasks involving math, reasoning, and knowledge understanding across LLMs of different parameter sizes and fine-tuning techniques. Our comprehensive experiments demonstrate that \trainallInfAttn achieves superior improvements compared to both the fully fine-tuned model and the base model in most scenarios, with significant performance gains. The high tolerance of \trainallInfAttn to data mismatches makes it resistant to jailbreaking tuning and enhances specialized tasks using general data.
Authors: Andreas Koukounas, Georgios Mastrapas, Michael G\"unther, Bo Wang, Scott Martens, Isabelle Mohr, Saba Sturua, Mohammad Kalim Akram, Joan Fontanals Mart\'inez, Saahil Ognawala, Susana Guzman, Maximilian Werk, Nan Wang, Han Xiao
Abstract: Contrastive Language-Image Pretraining (CLIP) is widely used to train models to align images and texts in a common embedding space by mapping them to fixed-sized vectors. These models are key to multimodal information retrieval and related tasks. However, CLIP models generally underperform in text-only tasks compared to specialized text models. This creates inefficiencies for information retrieval systems that keep separate embeddings and models for text-only and multimodal tasks. We propose a novel, multi-task contrastive training method to address this issue, which we use to train the jina-clip-v1 model to achieve the state-of-the-art performance on both text-image and text-text retrieval tasks.
Authors: Chen Zhang, Chengguang Tang, Dading Chong, Ke Shi, Guohua Tang, Feng Jiang, Haizhou Li
Abstract: Mainstream approaches to aligning large language models (LLMs) heavily rely on human preference data, particularly when models require periodic updates. The standard process for iterative alignment of LLMs involves collecting new human feedback for each update. However, the data collection process is costly and challenging to scale. To address this issue, we introduce the "TS-Align" framework, which fine-tunes a policy model using pairwise feedback data automatically mined from its outputs. This automatic mining process is efficiently accomplished through the collaboration between a large-scale teacher model and a small-scale student model. The policy fine-tuning process can be iteratively repeated using on-policy generations within our proposed teacher-student collaborative framework. Through extensive experiments, we demonstrate that our final aligned policy outperforms the base policy model with an average win rate of 69.7% across seven conversational or instruction-following datasets. Furthermore, we show that the ranking capability of the teacher is effectively distilled into the student through our pipeline, resulting in a small-scale yet effective reward model for policy model alignment.
Authors: Franz Louis Cesista, Rui Aguiar, Jason Kim, Paolo Acilo
Abstract: Business Document Information Extraction (BDIE) is the problem of transforming a blob of unstructured information (raw text, scanned documents, etc.) into a structured format that downstream systems can parse and use. It has two main tasks: Key-Information Extraction (KIE) and Line Items Recognition (LIR). In this paper, we argue that BDIE is best modeled as a Tool Use problem, where the tools are these downstream systems. We then present Retrieval Augmented Structured Generation (RASG), a novel general framework for BDIE that achieves state of the art (SOTA) results on both KIE and LIR tasks on BDIE benchmarks. The contributions of this paper are threefold: (1) We show, with ablation benchmarks, that Large Language Models (LLMs) with RASG are already competitive with or surpasses current SOTA Large Multimodal Models (LMMs) without RASG on BDIE benchmarks. (2) We propose a new metric class for Line Items Recognition, General Line Items Recognition Metric (GLIRM), that is more aligned with practical BDIE use cases compared to existing metrics, such as ANLS*, DocILE, and GriTS. (3) We provide a heuristic algorithm for backcalculating bounding boxes of predicted line items and tables without the need for vision encoders. Finally, we claim that, while LMMs might sometimes offer marginal performance benefits, LLMs + RASG is oftentimes superior given real-world applications and constraints of BDIE.
Authors: Yuchi Liu, Jaskirat Singh, Gaowen Liu, Ali Payani, Liang Zheng
Abstract: Large language models (LLMs) have shown great progress in responding to user questions, allowing for a multitude of diverse applications. Yet, the quality of LLM outputs heavily depends on the prompt design, where a good prompt might enable the LLM to answer a very challenging question correctly. Therefore, recent works have developed many strategies for improving the prompt, including both manual crafting and in-domain optimization. However, their efficacy in unrestricted scenarios remains questionable, as the former depends on human design for specific questions and the latter usually generalizes poorly to unseen scenarios. To address these problems, we give LLMs the freedom to design the best prompts according to themselves. Specifically, we include a hierarchy of LLMs, first constructing a prompt with precise instructions and accurate wording in a hierarchical manner, and then using this prompt to generate the final answer to the user query. We term this pipeline Hierarchical Multi-Agent Workflow, or HMAW. In contrast with prior works, HMAW imposes no human restriction and requires no training, and is completely task-agnostic while capable of adjusting to the nuances of the underlying task. Through both quantitative and qualitative experiments across multiple benchmarks, we verify that despite its simplicity, the proposed approach can create detailed and suitable prompts, further boosting the performance of current LLMs.
Authors: Andy Liu, Mona Diab, Daniel Fried
Abstract: The task of persona-steered text generation requires large language models (LLMs) to generate text that reflects the distribution of views that an individual fitting a persona could have. People have multifaceted personas, but prior work on bias in LLM-generated opinions has only explored multiple-choice settings or one-dimensional personas. We define an incongruous persona as a persona with multiple traits where one trait makes its other traits less likely in human survey data, e.g. political liberals who support increased military spending. We find that LLMs are 9.7% less steerable towards incongruous personas than congruous ones, sometimes generating the stereotypical stance associated with its demographic rather than the target stance. Models that we evaluate that are fine-tuned with Reinforcement Learning from Human Feedback (RLHF) are more steerable, especially towards stances associated with political liberals and women, but present significantly less diverse views of personas. We also find variance in LLM steerability that cannot be predicted from multiple-choice opinion evaluation. Our results show the importance of evaluating models in open-ended text generation, as it can surface new LLM opinion biases. Moreover, such a setup can shed light on our ability to steer models toward a richer and more diverse range of viewpoints.
Authors: Ruochen Zhao, Wenxuan Zhang, Yew Ken Chia, Deli Zhao, Lidong Bing
Abstract: As LLMs evolve on a daily basis, there is an urgent need for a trustworthy evaluation method that can provide robust evaluation results in a timely fashion. Currently, as static benchmarks are prone to contamination concerns, users tend to trust human voting platforms, such as Chatbot Arena. However, human annotations require extensive manual efforts. To provide an automatic, robust, and trustworthy evaluation framework, we innovatively propose the Auto-Arena of LLMs, which automates the entire evaluation process with LLM agents. Firstly, an examiner LLM devises queries. Then, a pair of candidate LLMs engage in a multi-round peer-battle around the query, during which the LLM's true performance gaps become visible. Finally, a committee of LLM judges collectively discuss and determine the winner, which alleviates bias and promotes fairness. In our extensive experiment on the 17 newest LLMs, Auto-Arena shows the highest correlation with human preferences, providing a promising alternative to human evaluation platforms.
Authors: Guy Mor-Lan, Effi Levi, Tamir Sheafer, Shaul R. Shenhav
Abstract: We introduce IsraParlTweet, a new linked corpus of Hebrew-language parliamentary discussions from the Knesset (Israeli Parliament) between the years 1992-2023 and Twitter posts made by Members of the Knesset between the years 2008-2023, containing a total of 294.5 million Hebrew tokens. In addition to raw text, the corpus contains comprehensive metadata on speakers and Knesset sessions as well as several linguistic annotations. As a result, IsraParlTweet can be used to conduct a wide variety of quantitative and qualitative analyses and provide valuable insights into political discourse in Israel.
Authors: Siva Uday Sampreeth Chebolu, Franck Dernoncourt, Nedim Lipka, Thamar Solorio
Abstract: Aspect-Based Sentiment Analysis (ABSA) has experienced tremendous expansion and diversity due to various shared tasks spanning several languages and fields and organized via SemEval workshops and Germeval. Nonetheless, a few shortcomings still need to be addressed, such as the lack of low-resource language evaluations and the emphasis on sentence-level analysis. To thoroughly assess ABSA techniques in the context of complete reviews, this research presents a novel task, Review-Level Opinion Aspect Sentiment Target (ROAST). ROAST seeks to close the gap between sentence-level and text-level ABSA by identifying every ABSA constituent at the review level. We extend the available datasets to enable ROAST, addressing the drawbacks noted in previous research by incorporating low-resource languages, numerous languages, and a variety of topics. Through this effort, ABSA research will be able to cover more ground and get a deeper comprehension of the task and its practical application in a variety of languages and domains (https://github.com/RiTUAL-UH/ROAST-ABSA).
Authors: Panagiotis C. Theocharopoulos, Spiros V. Georgakopoulos, Sotiris K. Tasoulis, Vassilis P. Plagianakos
Abstract: With the increasing use of Artificial Intelligence in Natural Language Processing, concerns have been raised regarding the detection of AI-generated text in various domains. This study aims to investigate this issue by proposing a methodology to accurately distinguish AI-generated and human-written book reviews. Our approach utilizes transfer learning, enabling the model to identify generated text across different topics while improving its ability to detect variations in writing style and vocabulary. To evaluate the effectiveness of the proposed methodology, we developed a dataset consisting of real book reviews and AI-generated reviews using the recently proposed Vicuna open-source language model. The experimental results demonstrate that it is feasible to detect the original source of text, achieving an accuracy rate of 96.86%. Our efforts are oriented toward the exploration of the capabilities and limitations of Large Language Models in the context of text identification. Expanding our knowledge in these aspects will be valuable for effectively navigating similar models in the future and ensuring the integrity and authenticity of human-generated content.
Authors: Shyam Sundhar Ramesh, Yifan Hu, Iason Chaimalas, Viraj Mehta, Pier Giuseppe Sessa, Haitham Bou Ammar, Ilija Bogunovic
Abstract: Adapting large language models (LLMs) for specific tasks usually involves fine-tuning through reinforcement learning with human feedback (RLHF) on preference data. While these data often come from diverse labelers' groups (e.g., different demographics, ethnicities, company teams, etc.), traditional RLHF approaches adopt a "one-size-fits-all" approach, i.e., they indiscriminately assume and optimize a single preference model, thus not being robust to unique characteristics and needs of the various groups. To address this limitation, we propose a novel Group Robust Preference Optimization (GRPO) method to align LLMs to individual groups' preferences robustly. Our approach builds upon reward-free direct preference optimization methods, but unlike previous approaches, it seeks a robust policy which maximizes the worst-case group performance. To achieve this, GRPO adaptively and sequentially weights the importance of different groups, prioritizing groups with worse cumulative loss. We theoretically study the feasibility of GRPO and analyze its convergence for the log-linear policy class. By fine-tuning LLMs with GRPO using diverse group-based global opinion data, we significantly improved performance for the worst-performing groups, reduced loss imbalances across groups, and improved probability accuracies compared to non-robust baselines.
Authors: Wei Zhong, Manasa Bharadwaj
Abstract: Speculative decoding (SD) has attracted a significant amount of research attention due to the substantial speedup it can achieve for LLM inference. However, despite the high speedups they offer, speculative decoding methods often achieve optimal performance on high-end devices or with a substantial GPU memory overhead. Given limited memory and the necessity of quantization, a high-performing model on a high-end GPU can slow down by up to 7 times. To this end, we propose Skippy Simultaneous Speculative Decoding (or S3D), a cost-effective self-speculative SD method based on simultaneous multi-token decoding and mid-layer skipping. When compared against recent effective open-source SD systems, our method has achieved one of the top performance-memory ratios while requiring minimal architecture changes and training data. Leveraging our memory efficiency, we created a smaller yet more effective SD model based on Phi-3. It is 1.4 to 2 times faster than the quantized EAGLE model and operates in half-precision while using less VRAM.
Authors: Ziwei Ji, Yuzhe Gu, Wenwei Zhang, Chengqi Lyu, Dahua Lin, Kai Chen
Abstract: Reducing the `$\textit{hallucination}$' problem of Large Language Models (LLMs) is crucial for their wide applications. A comprehensive and fine-grained measurement of the hallucination is the first key step for the governance of this issue but is under-explored in the community. Thus, we present $\textbf{ANAH}$, a bilingual dataset that offers $\textbf{AN}$alytical $\textbf{A}$nnotation of $\textbf{H}$allucinations in LLMs within Generative Question Answering. Each answer sentence in our dataset undergoes rigorous annotation, involving the retrieval of a reference fragment, the judgment of the hallucination type, and the correction of hallucinated content. ANAH consists of ~12k sentence-level annotations for ~4.3k LLM responses covering over 700 topics, constructed by a human-in-the-loop pipeline. Thanks to the fine granularity of the hallucination annotations, we can quantitatively confirm that the hallucinations of LLMs progressively accumulate in the answer and use ANAH to train and evaluate hallucination annotators. We conduct extensive experiments on studying generative and discriminative annotators and show that, although current open-source LLMs have difficulties in fine-grained hallucination annotation, the generative annotator trained with ANAH can surpass all open-source LLMs and GPT-3.5, obtain performance competitive with GPT-4, and exhibits better generalization ability on unseen questions.
Authors: Roberto Ceraolo, Dmitrii Kharlapenko, Am\'elie Reymond, Rada Mihalcea, Mrinmaya Sachan, Bernhard Sch\"olkopf, Zhijing Jin
Abstract: Humans have an innate drive to seek out causality. Whether fuelled by curiosity or specific goals, we constantly question why things happen, how they are interconnected, and many other related phenomena. To develop AI agents capable of addressing this natural human quest for causality, we urgently need a comprehensive dataset of natural causal questions. Unfortunately, existing datasets either contain only artificially-crafted questions that do not reflect real AI usage scenarios or have limited coverage of questions from specific sources. To address this gap, we present CausalQuest, a dataset of 13,500 naturally occurring questions sourced from social networks, search engines, and AI assistants. We formalize the definition of causal questions and establish a taxonomy for finer-grained classification. Through a combined effort of human annotators and large language models (LLMs), we carefully label the dataset. We find that 42% of the questions humans ask are indeed causal, with the majority seeking to understand the causes behind given effects. Using this dataset, we train efficient classifiers (up to 2.85B parameters) for the binary task of identifying causal questions, achieving high performance with F1 scores of up to 0.877. We conclude with a rich set of future research directions that can build upon our data and models.
Authors: Bolin Ni, JingCheng Hu, Yixuan Wei, Houwen Peng, Zheng Zhang, Gaofeng Meng, Han Hu
Abstract: In this work, we present Xwin-LM, a comprehensive suite of alignment methodologies for large language models (LLMs). This suite encompasses several key techniques, including supervised finetuning (SFT), reward modeling (RM), rejection sampling finetuning (RS), and direct preference optimization (DPO). The key components are as follows: (1) Xwin-LM-SFT, models initially finetuned with high-quality instruction data; (2) Xwin-Pair, a large-scale, multi-turn preference dataset meticulously annotated using GPT-4; (3) Xwin-RM, reward models trained on Xwin-Pair, developed at scales of 7B, 13B, and 70B parameters; (4) Xwin-Set, a multiwise preference dataset in which each prompt is linked to 64 unique responses generated by Xwin-LM-SFT and scored by Xwin-RM; (5) Xwin-LM-RS, models finetuned with the highest-scoring responses from Xwin-Set; (6) Xwin-LM-DPO, models further optimized on Xwin-Set using the DPO algorithm. Our evaluations on AlpacaEval and MT-bench demonstrate consistent and significant improvements across the pipeline, demonstrating the strength and scalability of Xwin-LM. The repository https://github.com/Xwin-LM/Xwin-LM will be continually updated to foster community research.
Authors: Chlo\'e Sekkat, Fanny Leroy, Salima Mdhaffar, Blake Perry Smith, Yannick Est\`eve, Joseph Dureau, Alice Coucke
Abstract: Recent works demonstrate that voice assistants do not perform equally well for everyone, but research on demographic robustness of speech technologies is still scarce. This is mainly due to the rarity of large datasets with controlled demographic tags. This paper introduces the Sonos Voice Control Bias Assessment Dataset, an open dataset composed of voice assistant requests for North American English in the music domain (1,038 speakers, 166 hours, 170k audio samples, with 9,040 unique labelled transcripts) with a controlled demographic diversity (gender, age, dialectal region and ethnicity). We also release a statistical demographic bias assessment methodology, at the univariate and multivariate levels, tailored to this specific use case and leveraging spoken language understanding metrics rather than transcription accuracy, which we believe is a better proxy for user experience. To demonstrate the capabilities of this dataset and statistical method to detect demographic bias, we consider a pair of state-of-the-art Automatic Speech Recognition and Spoken Language Understanding models. Results show statistically significant differences in performance across age, dialectal region and ethnicity. Multivariate tests are crucial to shed light on mixed effects between dialectal region, gender and age.
Authors: Andrew Katumba, Sudi Murindanyi, John Trevor Kasule, Elvis Mugume
Abstract: The advent of Internet of Things (IoT) technology has generated massive interest in voice-controlled smart homes. While many voice-controlled smart home systems are designed to understand and support widely spoken languages like English, speakers of low-resource languages like Luganda may need more support. This research project aimed to develop a Luganda speech intent classification system for IoT applications to integrate local languages into smart home environments. The project uses hardware components such as Raspberry Pi, Wio Terminal, and ESP32 nodes as microcontrollers. The Raspberry Pi processes Luganda voice commands, the Wio Terminal is a display device, and the ESP32 nodes control the IoT devices. The ultimate objective of this work was to enable voice control using Luganda, which was accomplished through a natural language processing (NLP) model deployed on the Raspberry Pi. The NLP model utilized Mel Frequency Cepstral Coefficients (MFCCs) as acoustic features and a Convolutional Neural Network (Conv2D) architecture for speech intent classification. A dataset of Luganda voice commands was curated for this purpose and this has been made open-source. This work addresses the localization challenges and linguistic diversity in IoT applications by incorporating Luganda voice commands, enabling users to interact with smart home devices without English proficiency, especially in regions where local languages are predominant.
Authors: Angelica Chen, Sadhika Malladi, Lily H. Zhang, Xinyi Chen, Qiuyi Zhang, Rajesh Ranganath, Kyunghyun Cho
Abstract: Preference learning algorithms (e.g., RLHF and DPO) are frequently used to steer LLMs to produce generations that are more preferred by humans, but our understanding of their inner workings is still limited. In this work, we study the conventional wisdom that preference learning trains models to assign higher likelihoods to more preferred outputs than less preferred outputs, measured via $\textit{ranking accuracy}$. Surprisingly, we find that most state-of-the-art preference-tuned models achieve a ranking accuracy of less than 60% on common preference datasets. We furthermore derive the $\textit{idealized ranking accuracy}$ that a preference-tuned LLM would achieve if it optimized the DPO or RLHF objective perfectly. We demonstrate that existing models exhibit a significant $\textit{alignment gap}$ -- $\textit{i.e.}$, a gap between the observed and idealized ranking accuracies. We attribute this discrepancy to the DPO objective, which is empirically and theoretically ill-suited to fix even mild ranking errors in the reference model, and derive a simple and efficient formula for quantifying the difficulty of learning a given preference datapoint. Finally, we demonstrate that ranking accuracy strongly correlates with the empirically popular win rate metric when the model is close to the reference model used in the objective, shedding further light on the differences between on-policy (e.g., RLHF) and off-policy (e.g., DPO) preference learning algorithms.
Authors: Venkat Venkatasubramanian, Arijit Chakraborty
Abstract: The startling success of ChatGPT and other large language models (LLMs) using transformer-based generative neural network architecture in applications such as natural language processing and image synthesis has many researchers excited about potential opportunities in process systems engineering (PSE). The almost human-like performance of LLMs in these areas is indeed very impressive, surprising, and a major breakthrough. Their capabilities are very useful in certain tasks, such as writing first drafts of documents, code writing assistance, text summarization, etc. However, their success is limited in highly scientific domains as they cannot yet reason, plan, or explain due to their lack of in-depth domain knowledge. This is a problem in domains such as chemical engineering as they are governed by fundamental laws of physics and chemistry (and biology), constitutive relations, and highly technical knowledge about materials, processes, and systems. Although purely data-driven machine learning has its immediate uses, the long-term success of AI in scientific and engineering domains would depend on developing hybrid AI systems that use first principles and technical knowledge effectively. We call these hybrid AI systems Large Knowledge Models (LKMs), as they will not be limited to only NLP-based techniques or NLP-like applications. In this paper, we discuss the challenges and opportunities in developing such systems in chemical engineering.
Authors: Lucas Monteiro Paes, Dennis Wei, Flavio P. Calmon
Abstract: Feature attribution methods explain black-box machine learning (ML) models by assigning importance scores to input features. These methods can be computationally expensive for large ML models. To address this challenge, there has been increasing efforts to develop amortized explainers, where a machine learning model is trained to predict feature attribution scores with only one inference. Despite their efficiency, amortized explainers can produce inaccurate predictions and misleading explanations. In this paper, we propose selective explanations, a novel feature attribution method that (i) detects when amortized explainers generate low-quality explanations and (ii) improves these explanations using a technique called explanations with initial guess. Our selective explanation method allows practitioners to specify the fraction of samples that receive explanations with initial guess, offering a principled way to bridge the gap between amortized explainers and their high-quality counterparts.
Authors: Zhenmei Shi, Junyi Wei, Zhuoyan Xu, Yingyu Liang
Abstract: Large language models (LLM) have emerged as a powerful tool for AI, with the key ability of in-context learning (ICL), where they can perform well on unseen tasks based on a brief series of task examples without necessitating any adjustments to the model parameters. One recent interesting mysterious observation is that models of different scales may have different ICL behaviors: larger models tend to be more sensitive to noise in the test context. This work studies this observation theoretically aiming to improve the understanding of LLM and ICL. We analyze two stylized settings: (1) linear regression with one-layer single-head linear transformers and (2) parity classification with two-layer multiple attention heads transformers (non-linear data and non-linear model). In both settings, we give closed-form optimal solutions and find that smaller models emphasize important hidden features while larger ones cover more hidden features; thus, smaller models are more robust to noise while larger ones are more easily distracted, leading to different ICL behaviors. This sheds light on where transformers pay attention to and how that affects ICL. Preliminary experimental results on large base and chat models provide positive support for our analysis.
Authors: Vijay Lingam, Atula Tejaswi, Aditya Vavre, Aneesh Shetty, Gautham Krishna Gudur, Joydeep Ghosh, Alex Dimakis, Eunsol Choi, Aleksandar Bojchevski, Sujay Sanghavi
Abstract: Popular parameter-efficient fine-tuning (PEFT) methods, such as LoRA and its variants, freeze pre-trained model weights \(W\) and inject learnable matrices \(\Delta W\). These \(\Delta W\) matrices are structured for efficient parameterization, often using techniques like low-rank approximations or scaling vectors. However, these methods typically show a performance gap compared to full fine-tuning. Although recent PEFT methods have narrowed this gap, they do so at the cost of additional learnable parameters. We propose SVFT, a simple approach that fundamentally differs from existing methods: the structure imposed on \(\Delta W\) depends on the specific weight matrix \(W\). Specifically, SVFT updates \(W\) as a sparse combination of outer products of its singular vectors, training only the coefficients (scales) of these sparse combinations. This approach allows fine-grained control over expressivity through the number of coefficients. Extensive experiments on language and vision benchmarks show that SVFT recovers up to 96% of full fine-tuning performance while training only 0.006 to 0.25% of parameters, outperforming existing methods that only recover up to 85% performance using 0.03 to 0.8% of the trainable parameter budget.
Authors: Sean Williams, James Huckle
Abstract: We introduce a comprehensive Linguistic Benchmark designed to evaluate the limitations of Large Language Models (LLMs) in domains such as logical reasoning, spatial intelligence, and linguistic understanding, among others. Through a series of straightforward questions, it uncovers the significant limitations of well-regarded models to perform tasks that humans manage with ease. It also highlights the potential of prompt engineering to mitigate some errors and underscores the necessity for better training methodologies. Our findings stress the importance of grounding LLMs with human reasoning and common sense, emphasising the need for human-in-the-loop for enterprise applications. We hope this work paves the way for future research to enhance the usefulness and reliability of new models.
Authors: Yihe Deng, Pan Lu, Fan Yin, Ziniu Hu, Sheng Shen, James Zou, Kai-Wei Chang, Wei Wang
Abstract: Large vision language models (LVLMs) integrate large language models (LLMs) with pre-trained vision encoders, thereby activating the perception capability of the model to understand image inputs for different queries and conduct subsequent reasoning. Improving this capability requires high-quality vision-language data, which is costly and labor-intensive to acquire. Self-training approaches have been effective in single-modal settings to alleviate the need for labeled data by leveraging model's own generation. However, effective self-training remains a challenge regarding the unique visual perception and reasoning capability of LVLMs. To address this, we introduce Self-Training on Image Comprehension (STIC), which emphasizes a self-training approach specifically for image comprehension. First, the model self-constructs a preference dataset for image descriptions using unlabeled images. Preferred responses are generated through a step-by-step prompt, while dis-preferred responses are generated from either corrupted images or misleading prompts. To further self-improve reasoning on the extracted visual information, we let the model reuse a small portion of existing instruction-tuning data and append its self-generated image descriptions to the prompts. We validate the effectiveness of STIC across seven different benchmarks, demonstrating substantial performance gains of 4.0% on average while using 70% less supervised fine-tuning data than the current method. Further studies investigate various components of STIC and highlight its potential to leverage vast quantities of unlabeled images for self-training. Code and data are made publicly available.
Authors: Zixian Guo, Ming Liu, Zhilong Ji, Jinfeng Bai, Yiwen Guo, Wangmeng Zuo
Abstract: Learning a skill generally relies on both practical experience by doer and insightful high-level guidance by instructor. Will this strategy also work well for solving complex non-convex optimization problems? Here, a common gradient-based optimizer acts like a disciplined doer, making locally optimal update at each step. Recent methods utilize large language models (LLMs) to optimize solutions for concrete problems by inferring from natural language instructions, akin to a high-level instructor. In this paper, we show that these two optimizers are complementary to each other, suggesting a collaborative optimization approach. The gradient-based optimizer and LLM-based optimizer are combined in an interleaved manner. We instruct LLMs using task descriptions and timely optimization trajectories recorded during gradient-based optimization. Inferred results from LLMs are used as restarting points for the next stage of gradient optimization. By leveraging both the locally rigorous gradient-based optimizer and the high-level deductive LLM-based optimizer, our combined optimization method consistently yields improvements over competitive baseline prompt tuning methods. Our results demonstrate the synergistic effect of conventional gradient-based optimization and the inference ability of LLMs. The code is released at https://github.com/guozix/LLM-catalyst.
Authors: Wei Cheng, Yuhan Wu, Wei Hu
Abstract: Recent years have witnessed the deployment of code language models (LMs) in various code intelligence tasks such as code completion. Yet, it is challenging for pre-trained LMs to generate correct completions in private repositories. Previous studies retrieve cross-file context based on import relations or text similarity, which is insufficiently relevant to completion targets. In this paper, we propose a dataflow-guided retrieval augmentation approach, called DraCo, for repository-level code completion. DraCo parses a private repository into code entities and establishes their relations through an extended dataflow analysis, forming a repo-specific context graph. Whenever triggering code completion, DraCo precisely retrieves relevant background knowledge from the repo-specific context graph and generates well-formed prompts to query code LMs. Furthermore, we construct a large Python dataset, ReccEval, with more diverse completion targets. Our experiments demonstrate the superior accuracy and applicable efficiency of DraCo, improving code exact match by 3.43% and identifier F1-score by 3.27% on average compared to the state-of-the-art approach.
Authors: Arto Bendiken
Abstract: We present KNOW--the Knowledge Navigator Ontology for the World--the first ontology designed to capture everyday knowledge to augment large language models (LLMs) in real-world generative AI use cases such as personal AI assistants. Our domain is human life, both its everyday concerns and its major milestones. We have limited the initial scope of the modeled concepts to only established human universals: spacetime (places, events) plus social (people, groups, organizations). The inclusion criteria for modeled concepts are pragmatic, beginning with universality and utility. We compare and contrast previous work such as Schema.org and Cyc--as well as attempts at a synthesis of knowledge graphs and language models--noting how LLMs already encode internally much of the commonsense tacit knowledge that took decades to capture in the Cyc project. We also make available code-generated software libraries for the 12 most popular programming languages, enabling the direct use of ontology concepts in software engineering. We emphasize simplicity and developer experience in promoting AI interoperability.
Authors: Ehud Malul, Yair Meidan, Dudu Mimran, Yuval Elovici, Asaf Shabtai
Abstract: A key challenge associated with Kubernetes configuration files (KCFs) is that they are often highly complex and error-prone, leading to security vulnerabilities and operational setbacks. Rule-based (RB) tools for KCF misconfiguration detection rely on static rule sets, making them inherently limited and unable to detect newly-discovered misconfigurations. RB tools also suffer from misdetection, since mistakes are likely when coding the detection rules. Recent methods for detecting and remediating KCF misconfigurations are limited in terms of their scalability and detection coverage, or due to the fact that they have high expertise requirements and do not offer automated remediation along with misconfiguration detection. Novel approaches that employ LLMs in their pipeline rely on API-based, general-purpose, and mainly commercial models. Thus, they pose security challenges, have inconsistent classification performance, and can be costly. In this paper, we propose GenKubeSec, a comprehensive and adaptive, LLM-based method, which, in addition to detecting a wide variety of KCF misconfigurations, also identifies the exact location of the misconfigurations and provides detailed reasoning about them, along with suggested remediation. When empirically compared with three industry-standard RB tools, GenKubeSec achieved equivalent precision (0.990) and superior recall (0.999). When a random sample of KCFs was examined by a Kubernetes security expert, GenKubeSec's explanations as to misconfiguration localization, reasoning and remediation were 100% correct, informative and useful. To facilitate further advancements in this domain, we share the unique dataset we collected, a unified misconfiguration index we developed for label standardization, our experimentation code, and GenKubeSec itself as an open-source tool.
Authors: Alexander Nikitin, Jannik Kossen, Yarin Gal, Pekka Marttinen
Abstract: Uncertainty quantification in Large Language Models (LLMs) is crucial for applications where safety and reliability are important. In particular, uncertainty can be used to improve the trustworthiness of LLMs by detecting factually incorrect model responses, commonly called hallucinations. Critically, one should seek to capture the model's semantic uncertainty, i.e., the uncertainty over the meanings of LLM outputs, rather than uncertainty over lexical or syntactic variations that do not affect answer correctness. To address this problem, we propose Kernel Language Entropy (KLE), a novel method for uncertainty estimation in white- and black-box LLMs. KLE defines positive semidefinite unit trace kernels to encode the semantic similarities of LLM outputs and quantifies uncertainty using the von Neumann entropy. It considers pairwise semantic dependencies between answers (or semantic clusters), providing more fine-grained uncertainty estimates than previous methods based on hard clustering of answers. We theoretically prove that KLE generalizes the previous state-of-the-art method called semantic entropy and empirically demonstrate that it improves uncertainty quantification performance across multiple natural language generation datasets and LLM architectures.
Authors: Ihab Asaad, Maxime Jacquelin, Olivier Perrotin, Laurent Girin, Thomas Hueber
Abstract: Most speech self-supervised learning (SSL) models are trained with a pretext task which consists in predicting missing parts of the input signal, either future segments (causal prediction) or segments masked anywhere within the input (non-causal prediction). Learned speech representations can then be efficiently transferred to downstream tasks (e.g., automatic speech or speaker recognition). In the present study, we investigate the use of a speech SSL model for speech inpainting, that is reconstructing a missing portion of a speech signal from its surrounding context, i.e., fulfilling a downstream task that is very similar to the pretext task. To that purpose, we combine an SSL encoder, namely HuBERT, with a neural vocoder, namely HiFiGAN, playing the role of a decoder. In particular, we propose two solutions to match the HuBERT output with the HiFiGAN input, by freezing one and fine-tuning the other, and vice versa. Performance of both approaches was assessed in single- and multi-speaker settings, for both informed and blind inpainting configurations (i.e., the position of the mask is known or unknown, respectively), with different objective metrics and a perceptual evaluation. Performances show that if both solutions allow to correctly reconstruct signal portions up to the size of 200ms (and even 400ms in some cases), fine-tuning the SSL encoder provides a more accurate signal reconstruction in the single-speaker setting case, while freezing it (and training the neural vocoder instead) is a better strategy when dealing with multi-speaker data.
Authors: Alaa Nfissi, Wassim Bouachir, Nizar Bouguila, Brian Mishara
Abstract: In speech emotion recognition (SER), using predefined features without considering their practical importance may lead to high dimensional datasets, including redundant and irrelevant information. Consequently, high-dimensional learning often results in decreasing model accuracy while increasing computational complexity. Our work underlines the importance of carefully considering and analyzing features in order to build efficient SER systems. We present a new supervised SER method based on an efficient feature engineering approach. We pay particular attention to the explainability of results to evaluate feature relevance and refine feature sets. This is performed iteratively through feature evaluation loop, using Shapley values to boost feature selection and improve overall framework performance. Our approach allows thus to balance the benefits between model performance and transparency. The proposed method outperforms human-level performance (HLP) and state-of-the-art machine learning methods in emotion recognition on the TESS dataset.
Authors: Vijay Jaisankar, Sambaran Bandyopadhyay, Kalp Vyas, Varre Chaitanya, Shwetha Somasundaram
Abstract: A poster from a long input document can be considered as a one-page easy-to-read multimodal (text and images) summary presented on a nice template with good design elements. Automatic transformation of a long document into a poster is a very less studied but challenging task. It involves content summarization of the input document followed by template generation and harmonization. In this work, we propose a novel deep submodular function which can be trained on ground truth summaries to extract multimodal content from the document and explicitly ensures good coverage, diversity and alignment of text and images. Then, we use an LLM based paraphraser and propose to generate a template with various design aspects conditioned on the input content. We show the merits of our approach through extensive automated and human evaluations.
Authors: Massimo Bini, Karsten Roth, Zeynep Akata, Anna Khoreva
Abstract: Parameter-efficient finetuning (PEFT) has become ubiquitous to adapt foundation models to downstream task requirements while retaining their generalization ability. However, the amount of additionally introduced parameters and compute for successful adaptation and hyperparameter searches can explode quickly, especially when deployed at scale to serve numerous individual requests. To ensure effective, parameter-efficient, and hyperparameter-robust adaptation, we propose the ETHER transformation family, which performs Efficient fineTuning via HypErplane Reflections. By design, ETHER transformations require a minimal number of parameters, are less likely to deteriorate model performance, and exhibit robustness to hyperparameter and learning rate choices. In particular, we introduce ETHER and its relaxation ETHER+, which match or outperform existing PEFT methods with significantly fewer parameters ($\sim$$10$-$100$ times lower than LoRA or OFT) across multiple image synthesis and natural language tasks without exhaustive hyperparameter tuning. Finally, we investigate the recent emphasis on Hyperspherical Energy retention for adaptation and raise questions on its practical utility. The code is available at https://github.com/mwbini/ether.
Authors: Ajay Patel, Markus Hofmarcher, Claudiu Leoveanu-Condrei, Marius-Constantin Dinu, Chris Callison-Burch, Sepp Hochreiter
Abstract: Training models to act as agents that can effectively navigate and perform actions in a complex environment, such as a web browser, has typically been challenging due to lack of training data. Large language models (LLMs) have recently demonstrated some capability to navigate novel environments as agents in a zero-shot or few-shot fashion, purely guided by natural language instructions as prompts. Recent research has also demonstrated LLMs have the capability to exceed their base performance through self-improvement, i.e. fine-tuning on data generated by the model itself. In this work, we explore the extent to which LLMs can self-improve their performance as agents in long-horizon tasks in a complex environment using the WebArena benchmark. In WebArena, an agent must autonomously navigate and perform actions on web pages to achieve a specified objective. We explore fine-tuning on three distinct synthetic training data mixtures and achieve a 31\% improvement in task completion rate over the base model on the WebArena benchmark through a self-improvement procedure. We additionally contribute novel evaluation metrics for assessing the performance, robustness, capabilities, and quality of trajectories of our fine-tuned agent models to a greater degree than simple, aggregate-level benchmark scores currently used to measure self-improvement.
Authors: Tal Reiss, George Kour, Naama Zwerdling, Ateret Anaby-Tavor, Yedid Hoshen
Abstract: When first deploying an anomaly detection system, e.g., to detect out-of-scope queries in chatbots, there are no observed data, making data-driven approaches ineffective. Zero-shot anomaly detection methods offer a solution to such "cold-start" cases, but unfortunately they are often not accurate enough. This paper studies the realistic but underexplored cold-start setting where an anomaly detection model is initialized using zero-shot guidance, but subsequently receives a small number of contaminated observations (namely, that may include anomalies). The goal is to make efficient use of both the zero-shot guidance and the observations. We propose ColdFusion, a method that effectively adapts the zero-shot anomaly detector to contaminated observations. To support future development of this new setting, we propose an evaluation suite consisting of evaluation protocols and metrics.
Authors: Ryan Cotterell, Georg Heigold
Abstract: Even for common NLP tasks, sufficient supervision is not available in many languages -- morphological tagging is no exception. In the work presented here, we explore a transfer learning scheme, whereby we train character-level recurrent neural taggers to predict morphological taggings for high-resource languages and low-resource languages together. Learning joint character representations among multiple related languages successfully enables knowledge transfer from the high-resource languages to the low-resource ones, improving accuracy by up to 30% over a monolingual model.
Authors: Yinpei Dai, Yichi Zhang, Hong Liu, Zhijian Ou, Yi Huang, Junlan Feng
Abstract: Slot filling is a crucial component in task-oriented dialog systems that is used to parse (user) utterances into semantic concepts called slots. An ontology is defined by the collection of slots and the values that each slot can take. The most widely used practice of treating slot filling as a sequence labeling task suffers from two main drawbacks. First, the ontology is usually pre-defined and fixed and therefore is not able to detect new labels for unseen slots. Second, the one-hot encoding of slot labels ignores the correlations between slots with similar semantics, which makes it difficult to share knowledge learned across different domains. To address these problems, we propose a new model called elastic conditional random field (eCRF), where each slot is represented by the embedding of its natural language description and modeled by a CRF layer. New slot values can be detected by eCRF whenever a language description is available for the slot. In our experiment, we show that eCRFs outperform existing models in both in-domain and cross-domain tasks, especially in predicting unseen slots and values.
Authors: Tom De Smedt
Abstract: We created a fine-grained AI system for the detection of antisemitism. This Explainable AI will identify English and German anti-Semitic expressions of dehumanization, verbal aggression and conspiracies in online social media messages across platforms, to support high-level decision making.
Authors: Zhuoyuan Mao, Chenhui Chu, Sadao Kurohashi
Abstract: Massively multilingual sentence representation models, e.g., LASER, SBERT-distill, and LaBSE, help significantly improve cross-lingual downstream tasks. However, the use of a large amount of data or inefficient model architectures results in heavy computation to train a new model according to our preferred languages and domains. To resolve this issue, we introduce efficient and effective massively multilingual sentence embedding (EMS), using cross-lingual token-level reconstruction (XTR) and sentence-level contrastive learning as training objectives. Compared with related studies, the proposed model can be efficiently trained using significantly fewer parallel sentences and GPU computation resources. Empirical results showed that the proposed model significantly yields better or comparable results with regard to cross-lingual sentence retrieval, zero-shot cross-lingual genre classification, and sentiment classification. Ablative analyses demonstrated the efficiency and effectiveness of each component of the proposed model. We release the codes for model training and the EMS pre-trained sentence embedding model, which supports 62 languages ( https://github.com/Mao-KU/EMS ).
Authors: Shih-Cheng Huang, Shih-Heng Wang, Min-Han Shih, Saurav Sahay, Hung-yi Lee
Abstract: Parameter-efficient (PE) methods (like Prompts or Adapters) for adapting pre-trained language models (PLM) to downstream tasks have been popular recently. However, hindrances still prevent these methods from reaching their full potential. For example, two significant challenges are few-shot adaptation and cross-task generalization. To tackle these issues, we propose a general PE priming framework to enhance and explore the few-shot adaptation and generalization ability of PE methods. In this framework, PLMs are primed with PE methods for rapidly adapting to various target tasks. To evaluate the generalization ability of these PE methods, we conduct experiments on a few-shot cross-domain benchmark containing 160 diverse NLP tasks. Our experiment not only reveals the best priming strategy but also verifies that priming facilitates the adaptation to target tasks.
Authors: Jiuzhou Han, Nigel Collier, Wray Buntine, Ehsan Shareghi
Abstract: Large language models (LLMs) have shown great abilities of solving various natural language tasks in different domains. Due to the training objective of LLMs and their pre-training data, LLMs are not very well equipped for tasks involving structured data generation. We propose a framework, Prompting with Iterative Verification (PiVe), to improve graph-based generative capability of LLMs. We show how a small language model could be trained to act as a verifier module for the output of an LLM~(i.e., ChatGPT, GPT-4), and to iteratively improve its performance via fine-grained corrective instructions. We also show how the verifier module could apply iterative corrections offline for a more cost-effective solution to the text-to-graph generation task. Experiments on three graph-based datasets show consistent improvement gained via PiVe. Additionally, we create GenWiki-HIQ and highlight that the verifier module can be used as a data augmentation tool to help improve the quality of automatically generated parallel text-graph datasets.
Authors: Joe Stacey, Marek Rei
Abstract: Knowledge distillation optimises a smaller student model to behave similarly to a larger teacher model, retaining some of the performance benefits. While this method can improve results on in-distribution examples, it does not necessarily generalise to out-of-distribution (OOD) settings. We investigate two complementary methods for improving the robustness of the resulting student models on OOD domains. The first approach augments the distillation with generated unlabelled examples that match the target distribution. The second method upsamples data points among the training set that are similar to the target distribution. When applied on the task of natural language inference (NLI), our experiments on MNLI show that distillation with these modifications outperforms previous robustness solutions. We also find that these methods improve performance on OOD domains even beyond the target domain.
Authors: Jiaxi Cui, Munan Ning, Zongjian Li, Bohua Chen, Yang Yan, Hao Li, Bin Ling, Yonghong Tian, Li Yuan
Abstract: AI legal assistants based on Large Language Models (LLMs) can provide accessible legal consulting services, but the hallucination problem poses potential legal risks. This paper presents Chatlaw, an innovative legal assistant utilizing a Mixture-of-Experts (MoE) model and a multi-agent system to enhance the reliability and accuracy of AI-driven legal services. By integrating knowledge graphs with artificial screening, we construct a high-quality legal dataset to train the MoE model. This model utilizes different experts to address various legal issues, optimizing the accuracy of legal responses. Additionally, Standardized Operating Procedures (SOP), modeled after real law firm workflows, significantly reduce errors and hallucinations in legal services. Our MoE model outperforms GPT-4 in the Lawbench and Unified Qualification Exam for Legal Professionals by 7.73% in accuracy and 11 points, respectively, and also surpasses other models in multiple dimensions during real-case consultations, demonstrating our robust capability for legal consultation.
Authors: Abhisek Tiwari, Muhammed Sinan, Kaushik Roy, Amit Sheth, Sriparna Saha, Pushpak Bhattacharyya
Abstract: Over the past two decades, dialogue modeling has made significant strides, moving from simple rule-based responses to personalized and persuasive response generation. However, despite these advancements, the objective functions and evaluation metrics for dialogue generation have remained stagnant. These lexical-based metrics, e.g., cross-entropy and BLEU, have two key limitations: (a) word-to-word matching without semantic consideration: It assigns the same credit for failure to generate "nice" and "rice" for "good", (b) missing context attribute for evaluating the generated response: Even if a generated response is relevant to the ongoing dialogue context, it may still be penalized for not matching the gold utterance provided in the corpus. In this paper, we first investigate these limitations comprehensively and propose a new loss function called Semantic Infused Contextualized diaLogue (SemTextualLogue) loss function. We also formulate an evaluation metric called Dialuation, incorporating both context and semantic relevance. We experimented with both non-pretrained and pre-trained models on two dialogue corpora, encompassing task-oriented and open-domain scenarios. We found that the dialogue generation models trained with SemTextualLogueloss attained superior performance compared to the traditional cross-entropy loss function. The findings establish that the effective training of a dialogue generation model hinges significantly on incorporating semantics and context. This pattern is also mirrored in the introduced Dialuation metric, where the consideration of both context and semantics correlates more strongly with human evaluation compared to traditional metrics.
Authors: Jiaan Wang, Yunlong Liang, Zengkui Sun, Yuxuan Cao, Jiarong Xu, Fandong Meng
Abstract: Knowledge editing aims to change language models' performance on several special cases (i.e., editing scope) by infusing the corresponding expected knowledge into them. With the recent advancements in large language models (LLMs), knowledge editing has been shown as a promising technique to adapt LLMs to new knowledge without retraining from scratch. However, most of the previous studies neglect the multi-lingual nature of some main-stream LLMs (e.g., LLaMA, ChatGPT and GPT-4), and typically focus on monolingual scenarios, where LLMs are edited and evaluated in the same language. As a result, it is still unknown the effect of source language editing on a different target language. In this paper, we aim to figure out this cross-lingual effect in knowledge editing. Specifically, we first collect a large-scale cross-lingual synthetic dataset by translating ZsRE from English to Chinese. Then, we conduct English editing on various knowledge editing methods covering different paradigms, and evaluate their performance in Chinese, and vice versa. To give deeper analyses of the cross-lingual effect, the evaluation includes four aspects, i.e., reliability, generality, locality and portability. Furthermore, we analyze the inconsistent behaviors of the edited models and discuss their specific challenges. Data and codes are available at https://github.com/krystalan/Bi_ZsRE
Authors: Haoran Luo, Haihong E, Zichen Tang, Shiyao Peng, Yikai Guo, Wentai Zhang, Chenghao Ma, Guanting Dong, Meina Song, Wei Lin, Yifan Zhu, Luu Anh Tuan
Abstract: Knowledge Base Question Answering (KBQA) aims to answer natural language questions over large-scale knowledge bases (KBs), which can be summarized into two crucial steps: knowledge retrieval and semantic parsing. However, three core challenges remain: inefficient knowledge retrieval, mistakes of retrieval adversely impacting semantic parsing, and the complexity of previous KBQA methods. To tackle these challenges, we introduce ChatKBQA, a novel and simple generate-then-retrieve KBQA framework, which proposes first generating the logical form with fine-tuned LLMs, then retrieving and replacing entities and relations with an unsupervised retrieval method, to improve both generation and retrieval more directly. Experimental results show that ChatKBQA achieves new state-of-the-art performance on standard KBQA datasets, WebQSP, and CWQ. This work can also be regarded as a new paradigm for combining LLMs with knowledge graphs (KGs) for interpretable and knowledge-required question answering. Our code is publicly available.
Authors: Franz Nowak, Anej Svete, Li Du, Ryan Cotterell
Abstract: This work investigates the computational expressivity of language models (LMs) based on recurrent neural networks (RNNs). Siegelmann and Sontag (1992) famously showed that RNNs with rational weights and hidden states and unbounded computation time are Turing complete. However, LMs define weightings over strings in addition to just (unweighted) language membership and the analysis of the computational power of RNN LMs (RLMs) should reflect this. We extend the Turing completeness result to the probabilistic case, showing how a rationally weighted RLM with unbounded computation time can simulate any deterministic probabilistic Turing machine (PTM) with rationally weighted transitions. Since, in practice, RLMs work in real-time, processing a symbol at every time step, we treat the above result as an upper bound on the expressivity of RLMs. We also provide a lower bound by showing that under the restriction to real-time computation, such models can simulate deterministic real-time rational PTMs.
Authors: Haoran Li, Dadi Guo, Donghao Li, Wei Fan, Qi Hu, Xin Liu, Chunkit Chan, Duanyi Yao, Yuan Yao, Yangqiu Song
Abstract: The rapid development of language models (LMs) brings unprecedented accessibility and usage for both models and users. On the one hand, powerful LMs achieve state-of-the-art performance over numerous downstream NLP tasks. On the other hand, more and more attention is paid to unrestricted model accesses that may bring malicious privacy risks of data leakage. To address these issues, many recent works propose privacy-preserving language models (PPLMs) with differential privacy (DP). Unfortunately, different DP implementations make it challenging for a fair comparison among existing PPLMs. In this paper, we present PrivLM-Bench, a multi-perspective privacy evaluation benchmark to empirically and intuitively quantify the privacy leakage of LMs. Instead of only reporting DP parameters, PrivLM-Bench sheds light on the neglected inference data privacy during actual usage. PrivLM-Bench first clearly defines multi-faceted privacy objectives. Then, PrivLM-Bench constructs a unified pipeline to perform private fine-tuning. Lastly, PrivLM-Bench performs existing privacy attacks on LMs with pre-defined privacy objectives as the empirical evaluation results. The empirical attack results are used to fairly and intuitively evaluate the privacy leakage of various PPLMs. We conduct extensive experiments on three datasets of GLUE for mainstream LMs.
Authors: Lanrui Wang, Jiangnan Li, Chenxu Yang, Zheng Lin, Hongyin Tang, Huan Liu, Xiaolei Huang, Yanan Cao, Jingang Wang, Weiping Wang
Abstract: Recently, there has been a heightened interest in building chatbots based on Large Language Models (LLMs) to emulate human-like qualities in dialogues, including expressing empathy and offering emotional support. Despite having access to commonsense knowledge to better understand the psychological aspects and causality of dialogue context, even these powerful LLMs struggle to achieve the goals of empathy and emotional support. As current approaches do not adequately anticipate dialogue future, they may mislead language models to ignore complex dialogue goals of empathy and emotional support, resulting in unsupportive responses lacking empathy. To address this issue, we present an innovative framework named Sensible Empathetic Dialogue Generation with Visionary Commonsense Knowledge (Sibyl). Designed to concentrate on the imminent dialogue future, this paradigm directs LLMs toward the implicit requirements of the conversation, aiming to provide more sensible responses. Experimental results demonstrate that incorporating our paradigm for acquiring commonsense knowledge into LLMs comprehensively enhances the quality of their responses.
Authors: Kung-Hsiang Huang, Mingyang Zhou, Hou Pong Chan, Yi R. Fung, Zhenhailong Wang, Lingyu Zhang, Shih-Fu Chang, Heng Ji
Abstract: Recent advancements in large vision-language models (LVLMs) have led to significant progress in generating natural language descriptions for visual content and thus enhancing various applications. One issue with these powerful models is that they sometimes produce texts that are factually inconsistent with the visual input. While there has been some effort to mitigate such inconsistencies in natural image captioning, the factuality of generated captions for structured document images, such as charts, has not received as much scrutiny, posing a potential threat to information reliability in critical applications. This work delves into the factuality aspect by introducing a comprehensive typology of factual errors in generated chart captions. A large-scale human annotation effort provides insight into the error patterns and frequencies in captions crafted by various chart captioning models, ultimately forming the foundation of a novel dataset, CHOCOLATE. Our analysis reveals that even state-of-the-art models, including GPT-4V, frequently produce captions laced with factual inaccuracies. In response to this challenge, we establish the new task of Chart Caption Factual Error Correction and introduce CHARTVE, a model for visual entailment that outperforms proprietary and open-source LVLMs in evaluating factual consistency. Furthermore, we propose C2TFEC, an interpretable two-stage framework that excels at correcting factual errors. This work inaugurates a new domain in factual error correction for chart captions, presenting a novel evaluation mechanism, and demonstrating an effective approach to ensuring the factuality of generated chart captions. The code and data as well as the continuously updated benchmark can be found at: https://khuangaf.github.io/CHOCOLATE/.
Authors: Chenghao Fan, Wei Wei, Xiaoye Qu, Zhenyi Lu, Wenfeng Xie, Yu Cheng, Dangyang Chen
Abstract: Recently, prompt-tuning with pre-trained language models (PLMs) has demonstrated the significantly enhancing ability of relation extraction (RE) tasks. However, in low-resource scenarios, where the available training data is scarce, previous prompt-based methods may still perform poorly for prompt-based representation learning due to a superficial understanding of the relation. To this end, we highlight the importance of learning high-quality relation representation in low-resource scenarios for RE, and propose a novel prompt-based relation representation method, named MVRE (\underline{M}ulti-\underline{V}iew \underline{R}elation \underline{E}xtraction), to better leverage the capacity of PLMs to improve the performance of RE within the low-resource prompt-tuning paradigm. Specifically, MVRE decouples each relation into different perspectives to encompass multi-view relation representations for maximizing the likelihood during relation inference. Furthermore, we also design a Global-Local loss and a Dynamic-Initialization method for better alignment of the multi-view relation-representing virtual words, containing the semantics of relation labels during the optimization learning process and initialization. Extensive experiments on three benchmark datasets show that our method can achieve state-of-the-art in low-resource settings.
Authors: Chengyue Wu, Yukang Gan, Yixiao Ge, Zeyu Lu, Jiahao Wang, Ye Feng, Ying Shan, Ping Luo
Abstract: Humans generally acquire new skills without compromising the old; however, the opposite holds for Large Language Models (LLMs), e.g., from LLaMA to CodeLLaMA. To this end, we propose a new post-pretraining method for LLMs with an expansion of Transformer blocks. We tune the expanded blocks using only new corpus, efficiently and effectively improving the model's knowledge without catastrophic forgetting. In this paper, we experiment on the corpus of code and math, yielding LLaMA Pro-8.3B, a versatile foundation model initialized from LLaMA2-7B, excelling in general tasks, programming, and mathematics. LLaMA Pro and its instruction-following counterpart (LLaMA Pro-Instruct) achieve advanced performance among various benchmarks, demonstrating superiority over existing open models in the LLaMA family and the immense potential of reasoning and addressing diverse tasks as an intelligent agent. Our findings provide valuable insights into integrating natural and programming languages, laying a solid foundation for developing advanced language agents that operate effectively in various environments.
Authors: Asma Ghandeharioun, Avi Caciularu, Adam Pearce, Lucas Dixon, Mor Geva
Abstract: Understanding the internal representations of large language models (LLMs) can help explain models' behavior and verify their alignment with human values. Given the capabilities of LLMs in generating human-understandable text, we propose leveraging the model itself to explain its internal representations in natural language. We introduce a framework called Patchscopes and show how it can be used to answer a wide range of questions about an LLM's computation. We show that many prior interpretability methods based on projecting representations into the vocabulary space and intervening on the LLM computation can be viewed as instances of this framework. Moreover, several of their shortcomings such as failure in inspecting early layers or lack of expressivity can be mitigated by Patchscopes. Beyond unifying prior inspection techniques, Patchscopes also opens up new possibilities such as using a more capable model to explain the representations of a smaller model, and multihop reasoning error correction.
Authors: Saibo Geng, Berkay D\"oner, Chris Wendler, Martin Josifoski, Robert West
Abstract: Constrained decoding, a technique for enforcing constraints on language model outputs, offers a way to control text generation without retraining or architectural modifications. Its application is, however, typically restricted to models that give users access to next-token distributions (usually via softmax logits), which poses a limitation with blackbox large language models (LLMs). This paper introduces sketch-guided constrained decoding (SGCD), a novel approach to constrained decoding for blackbox LLMs, which operates without access to the logits of the blackbox LLM. SGCD utilizes a locally hosted auxiliary model to refine the output of an unconstrained blackbox LLM, effectively treating this initial output as a "sketch" for further elaboration. This approach is complementary to traditional logit-based techniques and enables the application of constrained decoding in settings where full model transparency is unavailable. We demonstrate the efficacy of SGCD through experiments in closed information extraction and constituency parsing, showing how it enhances the utility and flexibility of blackbox LLMs for complex NLP tasks.
Authors: Qingyun Wang, Zixuan Zhang, Hongxiang Li, Xuan Liu, Jiawei Han, Huimin Zhao, Heng Ji
Abstract: Fine-grained few-shot entity extraction in the chemical domain faces two unique challenges. First, compared with entity extraction tasks in the general domain, sentences from chemical papers usually contain more entities. Moreover, entity extraction models usually have difficulty extracting entities of long-tailed types. In this paper, we propose Chem-FINESE, a novel sequence-to-sequence (seq2seq) based few-shot entity extraction approach, to address these two challenges. Our Chem-FINESE has two components: a seq2seq entity extractor to extract named entities from the input sentence and a seq2seq self-validation module to reconstruct the original input sentence from extracted entities. Inspired by the fact that a good entity extraction system needs to extract entities faithfully, our new self-validation module leverages entity extraction results to reconstruct the original input sentence. Besides, we design a new contrastive loss to reduce excessive copying during the extraction process. Finally, we release ChemNER+, a new fine-grained chemical entity extraction dataset that is annotated by domain experts with the ChemNER schema. Experiments in few-shot settings with both ChemNER+ and CHEMET datasets show that our newly proposed framework has contributed up to 8.26% and 6.84% absolute F1-score gains respectively.
Authors: Jiuzhou Han, Wray Buntine, Ehsan Shareghi
Abstract: While Language Agents have achieved promising success by placing Large Language Models at the core of a more versatile design that dynamically interacts with the external world, the existing approaches neglect the notion of uncertainty during these interactions. We present the Uncertainty-Aware Language Agent (UALA), a framework that orchestrates the interaction between the agent and the external world using uncertainty quantification. Compared with other well-known counterparts like ReAct, our extensive experiments across 3 representative tasks (HotpotQA, StrategyQA, MMLU) and various LLM sizes demonstrate that UALA brings a significant improvement of performance, while having a substantially lower reliance on the external world (i.e., reduced number of tool calls and tokens). Our analyses provide various insights including the great potential of UALA compared with agent fine-tuning, and underscore the unreliability of verbalised confidence of LLMs as a proxy for uncertainty.
Authors: Haochun Wang, Sendong Zhao, Zewen Qiang, Nuwa Xi, Bing Qin, Ting Liu
Abstract: In the field of natural language processing (NLP), Large Language Models (LLMs) have precipitated a paradigm shift, markedly enhancing performance in natural language generation tasks. Despite these advancements, the comprehensive evaluation of LLMs remains an inevitable challenge for the community. Recently, the utilization of Multiple Choice Question Answering (MCQA) as a benchmark for LLMs has gained considerable traction. This study first investigates the limitations of MCQA as an evaluation method for LLMs and then analyzes the fundamental reason for the limitations of MCQA, that while LLMs may select the correct answers, it is possible that they also recognize other wrong options as correct. Finally, we propose a dataset augmenting method for Multiple-Choice Questions (MCQs), MCQA+, that can more accurately reflect the performance of the model, which underscores the need for more robust evaluation mechanisms in assessing the performance of LLMs.
Authors: Zhiyuan Hu, Chumin Liu, Xidong Feng, Yilun Zhao, See-Kiong Ng, Anh Tuan Luu, Junxian He, Pang Wei Koh, Bryan Hooi
Abstract: In the face of uncertainty, the ability to *seek information* is of fundamental importance. In many practical applications, such as medical diagnosis and troubleshooting, the information needed to solve the task is not initially given and has to be actively sought by asking follow-up questions (for example, a doctor asking a patient for more details about their symptoms). In this work, we introduce Uncertainty of Thoughts (UoT), an algorithm to augment large language models with the ability to actively seek information by asking effective questions. UoT combines 1) an *uncertainty-aware simulation approach* which enables the model to simulate possible future scenarios and how likely they are to occur, 2) *uncertainty-based rewards* motivated by information gain which incentivizes the model to seek information, and 3) a *reward propagation scheme* to select the optimal question to ask in a way that maximizes the expected reward. In experiments on medical diagnosis, troubleshooting, and the `20 Questions` game, UoT achieves an average performance improvement of 38.1% in the rate of successful task completion across multiple LLMs compared with direct prompting and also improves efficiency (i.e., the number of questions needed to complete the task). Our code has been released [here](https://github.com/zhiyuanhubj/UoT)
Authors: Jian Wang, Chak Tou Leong, Jiashuo Wang, Dongding Lin, Wenjie Li, Xiao-Yong Wei
Abstract: Tuning language models for dialogue generation has been a prevalent paradigm for building capable dialogue agents. Yet, traditional tuning narrowly views dialogue generation as resembling other language generation tasks, ignoring the role disparities between two speakers and the multi-round interactive process that dialogues ought to be. Such a manner often leads to unsatisfactory chat consistency for the built agent. In this work, we emphasize the interactive, communicative nature of dialogue and argue that it is more feasible to model the speaker roles of agent and user separately, enabling the agent to adhere to its role consistently. With this in mind, we propose an efficient Multi-round Interactive Dialogue Tuning (Midi-Tuning) framework. It models the agent and user individually with two adapters built upon large language models. The adapters make use of respective utterances round by round in alternating order and they are tuned via a round-level memory caching mechanism. Extensive experiments demonstrate that, our framework performs superior to traditional fine-tuning and harbors the tremendous potential for improving dialogue consistency.
Authors: Zekun Li, Zhiyu Zoey Chen, Mike Ross, Patrick Huber, Seungwhan Moon, Zhaojiang Lin, Xin Luna Dong, Adithya Sagar, Xifeng Yan, Paul A. Crook
Abstract: Large language models (LLMs) are increasingly prevalent in conversational systems due to their advanced understanding and generative capabilities in general contexts. However, their effectiveness in task-oriented dialogues (TOD), which requires not only response generation but also effective dialogue state tracking (DST) within specific tasks and domains, remains less satisfying. In this work, we propose a novel approach FnCTOD for solving DST with LLMs through function calling. This method improves zero-shot DST, allowing adaptation to diverse domains without extensive data collection or model tuning. Our experimental results demonstrate that our approach achieves exceptional performance with both modestly sized open-source and also proprietary LLMs: with in-context prompting it enables various 7B or 13B parameter models to surpass the previous state-of-the-art (SOTA) achieved by ChatGPT, and improves ChatGPT's performance beating the SOTA by 5.6% average joint goal accuracy (JGA). Individual model results for GPT-3.5 and GPT-4 are boosted by 4.8% and 14%, respectively. We also show that by fine-tuning on a small collection of diverse task-oriented dialogues, we can equip modestly sized models, specifically a 13B parameter LLaMA2-Chat model, with function-calling capabilities and DST performance comparable to ChatGPT while maintaining their chat capabilities. We have made the code publicly available at https://github.com/facebookresearch/FnCTOD
Authors: Jiamu Bai, Daoyuan Chen, Bingchen Qian, Liuyi Yao, Yaliang Li
Abstract: Federated Learning (FL) has recently been applied to the parameter-efficient fine-tuning of Large Language Models (LLMs). While promising, it raises significant challenges due to the heterogeneous resources and data distributions of clients. This study introduces FlexLoRA, a simple yet effective aggregation scheme for LLM fine-tuning, which mitigates the ``bucket effect'' in traditional FL that restricts the potential of clients with ample resources by tying them to the capabilities of the least-resourced participants. FlexLoRA allows for dynamic adjustment of local LoRA ranks, fostering the development of a global model imbued with broader, less task-specific knowledge. By synthesizing a full-size LoRA weight from individual client contributions and employing Singular Value Decomposition (SVD) for weight redistribution, FlexLoRA fully leverages heterogeneous client resources. Involving thousands of clients performing heterogeneous NLP tasks and client resources, our experiments validate the efficacy of FlexLoRA, with the federated global model achieving consistently better improvement over SOTA FL methods in downstream NLP task performance across various heterogeneous distributions. FlexLoRA's practicality is further underscored by our theoretical analysis and its seamless integration with existing LoRA-based FL methods, offering a path toward cross-device, privacy-preserving federated tuning for LLMs.
Authors: Jiejun Tan, Zhicheng Dou, Yutao Zhu, Peidong Guo, Kun Fang, Ji-Rong Wen
Abstract: The integration of large language models (LLMs) and search engines represents a significant evolution in knowledge acquisition methodologies. However, determining the knowledge that an LLM already possesses and the knowledge that requires the help of a search engine remains an unresolved issue. Most existing methods solve this problem through the results of preliminary answers or reasoning done by the LLM itself, but this incurs excessively high computational costs. This paper introduces a novel collaborative approach, namely SlimPLM, that detects missing knowledge in LLMs with a slim proxy model, to enhance the LLM's knowledge acquisition process. We employ a proxy model which has far fewer parameters, and take its answers as heuristic answers. Heuristic answers are then utilized to predict the knowledge required to answer the user question, as well as the known and unknown knowledge within the LLM. We only conduct retrieval for the missing knowledge in questions that the LLM does not know. Extensive experimental results on five datasets with two LLMs demonstrate a notable improvement in the end-to-end performance of LLMs in question-answering tasks, achieving or surpassing current state-of-the-art models with lower LLM inference costs.
Authors: Jiajie Jin, Yutao Zhu, Yujia Zhou, Zhicheng Dou
Abstract: Retrieval-augmented large language models (LLMs) have demonstrated efficacy in knowledge-intensive tasks such as open-domain QA, addressing inherent challenges in knowledge update and factual inadequacy. However, inconsistencies between retrieval knowledge and the necessary knowledge for LLMs, leading to a decline in LLM's answer quality. This paper introduces BIDER, an approach that refines retrieval documents into Key Supporting Evidence (KSE) through knowledge synthesis, supervised fine-tuning (SFT), and preference alignment. We train BIDER by learning from crafting KSE, while maximizing its output to align with LLM's information acquisition preferences through reinforcement learning. Evaluations across five datasets show BIDER boosts LLMs' answer quality by 7% while reducing input content length in retrieval documents by 80%, outperforming existing methods. The proposed KSE simulation effectively equips LLMs with essential information for accurate question answering.
Authors: Guan-Ting Lin, Cheng-Han Chiang, Hung-yi Lee
Abstract: In spoken dialogue, even if two current turns are the same sentence, their responses might still differ when they are spoken in different styles. The spoken styles, containing paralinguistic and prosodic information, mark the most significant difference between text and speech modality. When using text-only LLMs to model spoken dialogue, text-only LLMs cannot give different responses based on the speaking style of the current turn. In this paper, we focus on enabling LLMs to listen to the speaking styles and respond properly. Our goal is to teach the LLM that "even if the sentences are identical if they are spoken in different styles, their corresponding responses might be different". Since there is no suitable dataset for achieving this goal, we collect a speech-to-speech dataset, StyleTalk, with the following desired characteristics: when two current speeches have the same content but are spoken in different styles, their responses will be different. To teach LLMs to understand and respond properly to the speaking styles, we propose the Spoken-LLM framework that can model the linguistic content and the speaking styles. We train Spoken-LLM using the StyleTalk dataset and devise a two-stage training pipeline to help the Spoken-LLM better learn the speaking styles. Based on extensive experiments, we show that Spoken-LLM outperforms text-only baselines and prior speech LLMs methods.
Authors: Yueqi Xie, Minghong Fang, Renjie Pi, Neil Gong
Abstract: Large Language Models (LLMs) face threats from jailbreak prompts. Existing methods for detecting jailbreak prompts are primarily online moderation APIs or finetuned LLMs. These strategies, however, often require extensive and resource-intensive data collection and training processes. In this study, we propose GradSafe, which effectively detects jailbreak prompts by scrutinizing the gradients of safety-critical parameters in LLMs. Our method is grounded in a pivotal observation: the gradients of an LLM's loss for jailbreak prompts paired with compliance response exhibit similar patterns on certain safety-critical parameters. In contrast, safe prompts lead to different gradient patterns. Building on this observation, GradSafe analyzes the gradients from prompts (paired with compliance responses) to accurately detect jailbreak prompts. We show that GradSafe, applied to Llama-2 without further training, outperforms Llama Guard, despite its extensive finetuning with a large dataset, in detecting jailbreak prompts. This superior performance is consistent across both zero-shot and adaptation scenarios, as evidenced by our evaluations on ToxicChat and XSTest. The source code is available at https://github.com/xyq7/GradSafe.
Authors: Zhihao Zhang, Jun Zhao, Qi Zhang, Tao Gui, Xuanjing Huang
Abstract: Large Language Models (LLMs) have demonstrated considerable cross-lingual alignment and generalization ability. Current research primarily focuses on improving LLMs' cross-lingual generalization capabilities. However, there is still a lack of research on the intrinsic mechanisms of how LLMs achieve cross-lingual alignment. From the perspective of region partitioning, this paper conducts several investigations on the linguistic competence of LLMs. We discover a core region in LLMs that corresponds to linguistic competence, accounting for approximately 1% of the total model parameters. Removing this core region by setting parameters to zero results in a significant performance decrease across 30 different languages. Furthermore, this core region exhibits significant dimensional dependence, perturbations to even a single parameter on specific dimensions leading to a loss of linguistic competence. Moreover, we discover that distinct monolingual regions exist for different languages, and disruption to these specific regions substantially reduces the LLMs' proficiency in those corresponding languages. Our research also indicates that freezing the core linguistic region during further pre-training can mitigate the issue of catastrophic forgetting (CF), a common phenomenon observed during further pre-training of LLMs. Overall, exploring the LLMs' functional regions provides insights into the foundation of their intelligence.
Authors: Xudong Lu, Qi Liu, Yuhui Xu, Aojun Zhou, Siyuan Huang, Bo Zhang, Junchi Yan, Hongsheng Li
Abstract: A pivotal advancement in the progress of large language models (LLMs) is the emergence of the Mixture-of-Experts (MoE) LLMs. Compared to traditional LLMs, MoE LLMs can achieve higher performance with fewer parameters, but it is still hard to deploy them due to their immense parameter sizes. Different from previous weight pruning methods that rely on specifically designed hardware, this paper mainly aims to enhance the deployment efficiency of MoE LLMs by introducing plug-and-play expert-level sparsification techniques. Specifically, we propose, for the first time to our best knowledge, post-training approaches for task-agnostic and task-specific expert pruning and skipping of MoE LLMs, tailored to improve deployment efficiency while maintaining model performance across a wide range of tasks. Extensive experiments show that our proposed methods can simultaneously reduce model sizes and increase the inference speed, while maintaining satisfactory performance. Data and code will be available at https://github.com/Lucky-Lance/Expert_Sparsity.
Authors: Lei Zhu, Xinjiang Wang, Wayne Zhang, Rynson W. H. Lau
Abstract: A practical large language model (LLM) service may involve a long system prompt, which specifies the instructions, examples, and knowledge documents of the task and is reused across requests. However, the long system prompt causes throughput/latency bottlenecks as the cost of generating the next token grows w.r.t. the sequence length. This paper aims to improve the efficiency of LLM services that involve long system prompts. Our key observation is that handling these system prompts requires heavily redundant memory accesses in existing causal attention computation algorithms. Specifically, for batched requests, the cached hidden states (\ie, key-value pairs) of system prompts are transferred from off-chip DRAM to on-chip SRAM multiple times, each corresponding to an individual request. To eliminate such a redundancy, we propose RelayAttention, an attention algorithm that allows reading these hidden states from DRAM exactly once for a batch of input tokens. RelayAttention is a free lunch: it maintains the generation quality while requiring no model retraining, as it is based on a mathematical reformulation of causal attention. We have observed significant performance improvements to a production-level system, vLLM, through integration with RelayAttention. The improvements are even more profound with longer system prompts.
Authors: Jin Yao, Eli Chien, Minxin Du, Xinyao Niu, Tianhao Wang, Zezhou Cheng, Xiang Yue
Abstract: This study investigates the concept of the `right to be forgotten' within the context of large language models (LLMs). We explore machine unlearning as a pivotal solution, with a focus on pre-trained models--a notably under-researched area. Our research delineates a comprehensive framework for machine unlearning in pre-trained LLMs, encompassing a critical analysis of seven diverse unlearning methods. Through rigorous evaluation using curated datasets from arXiv, books, and GitHub, we establish a robust benchmark for unlearning performance, demonstrating that these methods are over $10^5$ times more computationally efficient than retraining. Our results show that integrating gradient ascent with gradient descent on in-distribution data improves hyperparameter robustness. We also provide detailed guidelines for efficient hyperparameter tuning in the unlearning process. Our findings advance the discourse on ethical AI practices, offering substantive insights into the mechanics of machine unlearning for pre-trained LLMs and underscoring the potential for responsible AI development.
Authors: Luiza Pozzobon, Patrick Lewis, Sara Hooker, Beyza Ermis
Abstract: To date, toxicity mitigation in language models has almost entirely been focused on single-language settings. As language models embrace multilingual capabilities, it's crucial our safety measures keep pace. Recognizing this research gap, our approach expands the scope of conventional toxicity mitigation to address the complexities presented by multiple languages. In the absence of sufficient annotated datasets across languages, we employ translated data to evaluate and enhance our mitigation techniques. We also compare finetuning mitigation approaches against retrieval-augmented techniques under both static and continual toxicity mitigation scenarios. This allows us to examine the effects of translation quality and the cross-lingual transfer on toxicity mitigation. We also explore how model size and data quantity affect the success of these mitigation efforts. Covering nine languages, our study represents a broad array of linguistic families and levels of resource availability, ranging from high to mid-resource languages. Through comprehensive experiments, we provide insights into the complexities of multilingual toxicity mitigation, offering valuable insights and paving the way for future research in this increasingly important field. Code and data are available at https://github.com/for-ai/goodtriever.
Authors: Yanming Liu, Xinyue Peng, Jiannan Cao, Le Dai, Xingzu Liu, Weihao Liu, Mingbang Wang
Abstract: Large language models(LLMs) have shown its outperforming ability on various tasks and question answering. However, LLMs require substantial memory storage on low-resource devices. More critically, the computational speed on these devices is also severely limited. In this paper, we propose SPA(Side Plugin Adaption), a lightweight architecture for fast on-devices inference on the constraints of strict on-devices computation and memory constraints. Compared with other on-devices seq2seq generation, SPA could make a fast and stable inference on low-resource constraints, allowing it to obtain cost effiency. Our method establish an interaction between a pretrained LLMs on-cloud and additive parameters on-devices, which could provide the knowledge on both pretrained LLMs and featured personal feature. Further more, SPA provides a framework to keep feature-base parameters on low computational devices while leave the parameters containing general information on the high computational devices.
Authors: Aonan Zhang, Chong Wang, Yi Wang, Xuanyu Zhang, Yunfei Cheng
Abstract: In this paper, we introduce an improved approach of speculative decoding aimed at enhancing the efficiency of serving large language models. Our method capitalizes on the strengths of two established techniques: the classic two-model speculative decoding approach, and the more recent single-model approach, Medusa. Drawing inspiration from Medusa, our approach adopts a single-model strategy for speculative decoding. However, our method distinguishes itself by employing a single, lightweight draft head with a recurrent dependency design, akin in essence to the small, draft model uses in classic speculative decoding, but without the complexities of the full transformer architecture. And because of the recurrent dependency, we can use beam search to swiftly filter out undesired candidates with the draft head. The outcome is a method that combines the simplicity of single-model design and avoids the need to create a data-dependent tree attention structure only for inference in Medusa. We empirically demonstrate the effectiveness of the proposed method on several popular open source language models, along with a comprehensive analysis of the trade-offs involved in adopting this approach.
Authors: Korbinian Randl, John Pavlopoulos, Aron Henriksson, Tony Lindgren
Abstract: Contaminated or adulterated food poses a substantial risk to human health. Given sets of labeled web texts for training, Machine Learning and Natural Language Processing can be applied to automatically detect such risks. We publish a dataset of 7,546 short texts describing public food recall announcements. Each text is manually labeled, on two granularity levels (coarse and fine), for food products and hazards that the recall corresponds to. We describe the dataset and benchmark naive, traditional, and Transformer models. Based on our analysis, Logistic Regression based on a tf-idf representation outperforms RoBERTa and XLM-R on classes with low support. Finally, we discuss different prompting strategies and present an LLM-in-the-loop framework, based on Conformal Prediction, which boosts the performance of the base classifier while reducing energy consumption compared to normal prompting.
Authors: Alina Petukhova, Jo\~ao P. Matos-Carvalho, Nuno Fachada
Abstract: Text clustering is an important approach for organising the growing amount of digital content, helping to structure and find hidden patterns in uncategorised data. However, the effectiveness of text clustering heavily relies on the choice of textual embeddings and clustering algorithms. We argue that recent advances in large language models (LLMs) can potentially improve this task. In this research, we investigated how different textual embeddings -- particularly those used in LLMs -- and clustering algorithms affect how text datasets are clustered. A series of experiments were conducted to assess how embeddings influence clustering results, the role played by dimensionality reduction through summarisation, and model size adjustment. Findings reveal that LLM embeddings excel at capturing subtleties in structured language, while BERT leads the lightweight options in performance. In addition, we observe that increasing model dimensionality and employing summarization techniques do not consistently lead to improvements in clustering efficiency, suggesting that these strategies require careful analysis to use in real-life models. These results highlight a complex balance between the need for refined text representation and computational feasibility in text clustering applications. This study extends traditional text clustering frameworks by incorporating embeddings from LLMs, providing a path for improved methodologies, while informing new avenues for future research in various types of textual analysis.
Authors: Shijia Zhou, Leonie Weissweiler, Taiqi He, Hinrich Sch\"utze, David R. Mortensen, Lori Levin
Abstract: In this paper, we make a contribution that can be understood from two perspectives: from an NLP perspective, we introduce a small challenge dataset for NLI with large lexical overlap, which minimises the possibility of models discerning entailment solely based on token distinctions, and show that GPT-4 and Llama 2 fail it with strong bias. We then create further challenging sub-tasks in an effort to explain this failure. From a Computational Linguistics perspective, we identify a group of constructions with three classes of adjectives which cannot be distinguished by surface features. This enables us to probe for LLM's understanding of these constructions in various ways, and we find that they fail in a variety of ways to distinguish between them, suggesting that they don't adequately represent their meaning or capture the lexical properties of phrasal heads.
Authors: Ali Mahboub, Muhy Eddin Za'ter, Bashar Al-Rfooh, Yazan Estaitia, Adnan Jaljuli, Asma Hakouz
Abstract: The latest advancements in machine learning and deep learning have brought forth the concept of semantic similarity, which has proven immensely beneficial in multiple applications and has largely replaced keyword search. However, evaluating semantic similarity and conducting searches for a specific query across various documents continue to be a complicated task. This complexity is due to the multifaceted nature of the task, the lack of standard benchmarks, whereas these challenges are further amplified for Arabic language. This paper endeavors to establish a straightforward yet potent benchmark for semantic search in Arabic. Moreover, to precisely evaluate the effectiveness of these metrics and the dataset, we conduct our assessment of semantic search within the framework of retrieval augmented generation (RAG).
Authors: Jinwei Yao, Kaiqi Chen, Kexun Zhang, Jiaxuan You, Binhang Yuan, Zeke Wang, Tao Lin
Abstract: Given the increasing demand for tree-structured interactions with LLMs, we introduce DeFT (Decoding with Flash Tree-Attention), an IO-aware tree attention algorithm tailored for tree-structured inference. Unlike traditional sequence-based decoding, tree-structured decoding better accommodates modern task requirements, including self-consistency, few-shot prompting, multi-step reasoning, and multi-model/head coordination. However, existing sequence-based inference systems are ill-suited for tree-structured decoding, resulting in redundancy in computation, memory footprints, and memory access, thereby undermining inference efficiency. To address this challenge, DeFT maintains memory-efficient attention calculation with low memory footprints through two key stages: (1) QKV Preparation: We propose a KV-Guided Grouping Strategy with Tree Split to intelligently group QKV, optimizing GPU resource utilization while minimizing memory reads/writes for KV cache between GPU global memory and on-chip shared memory; (2)Attention Calculation: We compute partial attention of each QKV group in a fused kernel and employ a Tree-topology-aware Global Reduction strategy to obtain final attention. By reducing 73-99% KV cache IO and nearly 100% IO for partial results during attention calculation (e.g., Softmax), DeFT achieves up to 2.52/3.82x speedup in the end-to-end/attention latency across three practical tree-based workloads: namely, few-shot prompting, multi-step reasoning, and speculative decoding, over state-of-the-art attention algorithms.
Authors: Poulami Ghosh, Shikhar Vashishth, Raj Dabre, Pushpak Bhattacharyya
Abstract: Contemporary deep learning models effectively handle languages with diverse morphology despite not being directly integrated into them. Morphology and word order are closely linked, with the latter incorporated into transformer-based models through positional encodings. This prompts a fundamental inquiry: Is there a correlation between the morphological complexity of a language and the utilization of positional encoding in pre-trained language models? In pursuit of an answer, we present the first study addressing this question, encompassing 22 languages and 5 downstream tasks. Our findings reveal that the importance of positional encoding diminishes with increasing morphological complexity in languages. Our study motivates the need for a deeper understanding of positional encoding, augmenting them to better reflect the different languages under consideration.
Authors: Yichong Huang, Xiaocheng Feng, Baohang Li, Yang Xiang, Hui Wang, Bing Qin, Ting Liu
Abstract: Large language models (LLMs) exhibit complementary strengths in various tasks, motivating the research of LLM ensembling. However, existing work focuses on training an extra reward model or fusion model to select or combine all candidate answers, posing a great challenge to the generalization on unseen data distributions. Besides, prior methods use textual responses as communication media, ignoring the valuable information in the internal representations. In this work, we propose a training-free ensemble framework DeePEn, fusing the informative probability distributions yielded by different LLMs at each decoding step. Unfortunately, the vocabulary discrepancy between heterogeneous LLMs directly makes averaging the distributions unfeasible due to the token misalignment. To address this challenge, DeePEn maps the probability distribution of each model from its own probability space to a universal relative space based on the relative representation theory, and performs aggregation. Next, we devise a search-based inverse transformation to transform the aggregated result back to the probability space of one of the ensembling LLMs (main model), in order to determine the next token. We conduct extensive experiments on ensembles of different number of LLMs, ensembles of LLMs with different architectures, and ensembles between the LLM and the specialist model. Experimental results show that (i) DeePEn achieves consistent improvements across six benchmarks covering subject examination, reasoning, and knowledge, (ii) a well-performing specialist model can benefit from a less effective LLM through distribution fusion, and (iii) DeePEn has complementary strengths with other ensemble methods such as voting.
Authors: Bowen Chen, Namgi Han, Yusuke Miyao
Abstract: Large Language Models (LLMs), trained on massive corpora with billions of parameters, show unprecedented performance in various fields. Though surprised by their excellent performances, researchers also noticed some special behaviors of those LLMs. One of those behaviors is memorization, in which LLMs can generate the same content used to train them. Though previous research has discussed memorization, the memorization of LLMs still lacks explanation, especially the cause of memorization and the dynamics of generating them. In this research, we comprehensively discussed memorization from various perspectives and extended the discussion scope to not only just the memorized content but also less and unmemorized content. Through various studies, we found that: (1) Through experiments, we revealed the relation of memorization between model size, continuation size, and context size. Further, we showed how unmemorized sentences transition to memorized sentences. (2) Through embedding analysis, we showed the distribution and decoding dynamics across model size in embedding space for sentences with different memorization scores. The n-gram statistics analysis presents d (3) An analysis over n-gram and entropy decoding dynamics discovered a boundary effect when the model starts to generate memorized sentences or unmemorized sentences. (4)We trained a Transformer model to predict the memorization of different models, showing that it is possible to predict memorizations by context.
Authors: Haoze He, Juncheng Billy Li, Xuan Jiang, Heather Miller
Abstract: LoRA and its variants have become popular parameter-efficient fine-tuning (PEFT) methods due to their ability to avoid excessive computational costs. However, an accuracy gap often exists between PEFT methods and full fine-tuning (FT), and this gap has yet to be systematically studied. In this work, we introduce a method for selecting sparse sub-matrices that aim to minimize the performance gap between PEFT vs. full fine-tuning (FT) while also reducing both fine-tuning computational cost and memory cost. Our Sparse Matrix Tuning (SMT) method begins by identifying the most significant sub-matrices in the gradient update, updating only these blocks during the fine-tuning process. In our experiments, we demonstrate that SMT consistently surpasses other PEFT baseline (e.g. LoRA and DoRA) in fine-tuning popular large language models such as LLaMA across a broad spectrum of tasks, while reducing the GPU memory footprint by 67% compared to FT. We also examine how the performance of LoRA and DoRA tends to plateau and decline as the number of trainable parameters increases, in contrast, our SMT method does not suffer from such issue.
Authors: Kun Zhao, Bohao Yang, Chen Tang, Chenghua Lin, Liang Zhan
Abstract: The long-standing one-to-many problem of gold standard responses in open-domain dialogue systems presents challenges for automatic evaluation metrics. Though prior works have demonstrated some success by applying powerful Large Language Models (LLMs), existing approaches still struggle with the one-to-many problem, and exhibit subpar performance in domain-specific scenarios. We assume the commonsense reasoning biases within LLMs may hinder their performance in domainspecific evaluations. To address both issues, we propose a novel framework SLIDE (Small and Large Integrated for Dialogue Evaluation), that leverages both a small, specialised model (SLM), and LLMs for the evaluation of open domain dialogues. Our approach introduces several techniques: (1) Contrastive learning to differentiate between robust and non-robust response embeddings; (2) A novel metric for semantic sensitivity that combines embedding cosine distances with similarity learned through neural networks, and (3) a strategy for incorporating the evaluation results from both the SLM and LLMs. Our empirical results demonstrate that our approach achieves state-of-the-art performance in both the classification and evaluation tasks, and additionally the SLIDE evaluator exhibits better correlation with human judgements. Our code is available at https:// github.com/hegehongcha/SLIDE-ACL2024.
Authors: Yuhao Chen, Zhimu Wang, Bo Wen, Farhana Zulkernine
Abstract: Unstructured text in medical notes and dialogues contains rich information. Recent advancements in Large Language Models (LLMs) have demonstrated superior performance in question answering and summarization tasks on unstructured text data, outperforming traditional text analysis approaches. However, there is a lack of scientific studies in the literature that methodically evaluate and report on the performance of different LLMs, specifically for domain-specific data such as medical chart notes. We propose an evaluation approach to analyze the performance of open-source LLMs such as Llama2 and Mistral for medical summarization tasks, using GPT-4 as an assessor. Our innovative approach to quantitative evaluation of LLMs can enable quality control, support the selection of effective LLMs for specific tasks, and advance knowledge discovery in digital health.
Authors: Jiahuan Cao, Yongxin Shi, Dezhi Peng, Yang Liu, Lianwen Jin
Abstract: Classical Chinese Understanding (CCU) holds significant value in preserving and exploration of the outstanding traditional Chinese culture. Recently, researchers have attempted to leverage the potential of Large Language Models (LLMs) for CCU by capitalizing on their remarkable comprehension and semantic capabilities. However, no comprehensive benchmark is available to assess the CCU capabilities of LLMs. To fill this gap, this paper introduces C$^{3}$bench, a Comprehensive Classical Chinese understanding benchmark, which comprises 50,000 text pairs for five primary CCU tasks, including classification, retrieval, named entity recognition, punctuation, and translation. Furthermore, the data in C$^{3}$bench originates from ten different domains, covering most of the categories in classical Chinese. Leveraging the proposed C$^{3}$bench, we extensively evaluate the quantitative performance of 15 representative LLMs on all five CCU tasks. Our results not only establish a public leaderboard of LLMs' CCU capabilities but also gain some findings. Specifically, existing LLMs are struggle with CCU tasks and still inferior to supervised models. Additionally, the results indicate that CCU is a task that requires special attention. We believe this study could provide a standard benchmark, comprehensive baselines, and valuable insights for the future advancement of LLM-based CCU research. The evaluation pipeline and dataset are available at \url{https://github.com/SCUT-DLVCLab/C3bench}.
Authors: Zhengyang Tang, Chenyu Huang, Xin Zheng, Shixi Hu, Zizhuo Wang, Dongdong Ge, Benyou Wang
Abstract: Large Language Models (LLMs) have emerged as powerful tools for tackling complex Operations Research (OR) problem by providing the capacity in automating optimization modeling. However, current methodologies heavily rely on prompt engineering (e.g., multi-agent cooperation) with proprietary LLMs, raising data privacy concerns that could be prohibitive in industry applications. To tackle this issue, we propose training open-source LLMs for optimization modeling. We identify four critical requirements for the training dataset of OR LLMs, design and implement OR-Instruct, a semi-automated process for creating synthetic data tailored to specific requirements. We also introduce the IndustryOR benchmark, the first industrial benchmark for testing LLMs on solving real-world OR problems. We apply the data from OR-Instruct to various open-source LLMs of 7b size (termed as ORLMs), resulting in a significantly improved capability for optimization modeling. Our best-performing ORLM achieves state-of-the-art performance on the NL4OPT, MAMO, and IndustryOR benchmarks. Our code and data are available at \url{https://github.com/Cardinal-Operations/ORLM}.
Authors: Changle Qu, Sunhao Dai, Xiaochi Wei, Hengyi Cai, Shuaiqiang Wang, Dawei Yin, Jun Xu, Ji-Rong Wen
Abstract: Recently, tool learning with large language models (LLMs) has emerged as a promising paradigm for augmenting the capabilities of LLMs to tackle highly complex problems. Despite growing attention and rapid advancements in this field, the existing literature remains fragmented and lacks systematic organization, posing barriers to entry for newcomers. This gap motivates us to conduct a comprehensive survey of existing works on tool learning with LLMs. In this survey, we focus on reviewing existing literature from the two primary aspects (1) why tool learning is beneficial and (2) how tool learning is implemented, enabling a comprehensive understanding of tool learning with LLMs. We first explore the "why" by reviewing both the benefits of tool integration and the inherent benefits of the tool learning paradigm from six specific aspects. In terms of "how", we systematically review the literature according to a taxonomy of four key stages in the tool learning workflow: task planning, tool selection, tool calling, and response generation. Additionally, we provide a detailed summary of existing benchmarks and evaluation methods, categorizing them according to their relevance to different stages. Finally, we discuss current challenges and outline potential future directions, aiming to inspire both researchers and industrial developers to further explore this emerging and promising area. We also maintain a GitHub repository to continually keep track of the relevant papers and resources in this rising area at \url{https://github.com/quchangle1/LLM-Tool-Survey}.
Authors: Olga Golovneva, Tianlu Wang, Jason Weston, Sainbayar Sukhbaatar
Abstract: The attention mechanism is a critical component of Large Language Models (LLMs) that allows tokens in a sequence to interact with each other, but is order-invariant. Incorporating position encoding (PE) makes it possible to address by position, such as attending to the i-th token. However, current PE methods use token counts to derive position, and thus cannot generalize to higher levels of abstraction, such as attending to the i-th sentence. In this paper, we propose a new position encoding method, Contextual Position Encoding (CoPE), that allows positions to be conditioned on context by incrementing position only on certain tokens determined by the model. This allows more general position addressing such as attending to the $i$-th particular word, noun, or sentence. We show that CoPE can solve the selective copy, counting and Flip-Flop tasks where popular position embeddings fail, and improves perplexity on language modeling and coding tasks.
Authors: Ge Zhang, Scott Qu, Jiaheng Liu, Chenchen Zhang, Chenghua Lin, Chou Leuang Yu, Danny Pan, Esther Cheng, Jie Liu, Qunshu Lin, Raven Yuan, Tuney Zheng, Wei Pang, Xinrun Du, Yiming Liang, Yinghao Ma, Yizhi Li, Ziyang Ma, Bill Lin, Emmanouil Benetos, Huan Yang, Junting Zhou, Kaijing Ma, Minghao Liu, Morry Niu, Noah Wang, Quehry Que, Ruibo Liu, Sine Liu, Shawn Guo, Soren Gao, Wangchunshu Zhou, Xinyue Zhang, Yizhi Zhou, Yubo Wang, Yuelin Bai, Yuhan Zhang, Yuxiang Zhang, Zenith Wang, Zhenzhu Yang, Zijian Zhao, Jiajun Zhang, Wanli Ouyang, Wenhao Huang, Wenhu Chen
Abstract: Large Language Models (LLMs) have made great strides in recent years to achieve unprecedented performance across different tasks. However, due to commercial interest, the most competitive models like GPT, Gemini, and Claude have been gated behind proprietary interfaces without disclosing the training details. Recently, many institutions have open-sourced several strong LLMs like LLaMA-3, comparable to existing closed-source LLMs. However, only the model's weights are provided with most details (e.g., intermediate checkpoints, pre-training corpus, and training code, etc.) being undisclosed. To improve the transparency of LLMs, the research community has formed to open-source truly open LLMs (e.g., Pythia, Amber, OLMo), where more details (e.g., pre-training corpus and training code) are being provided. These models have greatly advanced the scientific study of these large models including their strengths, weaknesses, biases and risks. However, we observe that the existing truly open LLMs on reasoning, knowledge, and coding tasks are still inferior to existing state-of-the-art LLMs with similar model sizes. To this end, we open-source MAP-Neo, a highly capable and transparent bilingual language model with 7B parameters trained from scratch on 4.5T high-quality tokens. Our MAP-Neo is the first fully open-sourced bilingual LLM with comparable performance compared to existing state-of-the-art LLMs. Moreover, we open-source all details to reproduce our MAP-Neo, where the cleaned pre-training corpus, data cleaning pipeline, checkpoints, and well-optimized training/evaluation framework are provided. Finally, we hope our MAP-Neo will enhance and strengthen the open research community and inspire more innovations and creativities to facilitate the further improvements of LLMs.
Authors: Yihao Xue, Ali Payani, Yu Yang, Baharan Mirzasoleiman
Abstract: Pretrained machine learning models need to be adapted to distribution shifts when deployed in new target environments. When obtaining labeled data from the target distribution is expensive, few-shot adaptation with only a few examples from the target distribution becomes essential. In this work, we propose MixPro, a lightweight and highly data-efficient approach for few-shot adaptation. MixPro first generates a relatively large dataset by mixing (linearly combining) pre-trained embeddings of large source data with those of the few target examples. This process preserves important features of both source and target distributions, while mitigating the specific noise in the small target data. Then, it trains a linear classifier on the mixed embeddings to effectively adapts the model to the target distribution without overfitting the small target data. Theoretically, we demonstrate the advantages of MixPro over previous methods. Our experiments, conducted across various model architectures on 8 datasets featuring different types of distribution shifts, reveal that MixPro can outperform baselines by up to 7\%, with only 2-4 target examples.
Authors: Yinpei Dai, Run Peng, Sikai Li, Joyce Chai
Abstract: Zero-Shot Object Navigation (ZSON) enables agents to navigate towards open-vocabulary objects in unknown environments. The existing works of ZSON mainly focus on following individual instructions to find generic object classes, neglecting the utilization of natural language interaction and the complexities of identifying user-specific objects. To address these limitations, we introduce Zero-shot Interactive Personalized Object Navigation (ZIPON), where robots need to navigate to personalized goal objects while engaging in conversations with users. To solve ZIPON, we propose a new framework termed Open-woRld Interactive persOnalized Navigation (ORION), which uses Large Language Models (LLMs) to make sequential decisions to manipulate different modules for perception, navigation and communication. Experimental results show that the performance of interactive agents that can leverage user feedback exhibits significant improvement. However, obtaining a good balance between task completion and the efficiency of navigation and interaction remains challenging for all methods. We further provide more findings on the impact of diverse user feedback forms on the agents' performance. Code is available at https://github.com/sled-group/navchat.
Authors: Yupei Liu, Yuqi Jia, Runpeng Geng, Jinyuan Jia, Neil Zhenqiang Gong
Abstract: A prompt injection attack aims to inject malicious instruction/data into the input of an LLM-Integrated Application such that it produces results as an attacker desires. Existing works are limited to case studies. As a result, the literature lacks a systematic understanding of prompt injection attacks and their defenses. We aim to bridge the gap in this work. In particular, we propose a framework to formalize prompt injection attacks. Existing attacks are special cases in our framework. Moreover, based on our framework, we design a new attack by combining existing ones. Using our framework, we conduct a systematic evaluation on 5 prompt injection attacks and 10 defenses with 10 LLMs and 7 tasks. Our work provides a common benchmark for quantitatively evaluating future prompt injection attacks and defenses. To facilitate research on this topic, we make our platform public at https://github.com/liu00222/Open-Prompt-Injection.
Authors: Reyna Abhyankar, Zijian He, Vikranth Srivatsa, Hao Zhang, Yiying Zhang
Abstract: Large language models are increasingly integrated with external environments, tools, and agents like ChatGPT plugins to extend their capability beyond language-centric tasks. However, today's LLM inference systems are designed for standalone LLMs. They treat each external interaction as the end of LLM generation and form a new request when the interaction finishes, causing unnecessary recomputation of already computed contexts, which accounts for 37-40% of total model forwarding time. This paper presents InferCept, the first LLM inference framework targeting augmented LLMs and supporting the efficient interception of LLM generation. InferCept minimizes the GPU resource waste caused by LLM interceptions and dedicates saved memory for serving more requests. InferCept improves the overall serving throughput by 1.6x-2x and completes 2x more requests per second compared to the state-of-the-art LLM inference systems.
Authors: Cheng Zhang, Jianyi Cheng, George A. Constantinides, Yiren Zhao
Abstract: Post-training quantization of Large Language Models (LLMs) is challenging. In this work, we introduce Low-rank Quantization Error Reduction (LQER), which combines quantization and low-rank approximation to recover the model capability. LQER leverages an activation-induced scale matrix to drive the singular value distribution of quantization error towards a desirable distribution, which enables nearly-lossless W4A8 quantization on various LLMs and downstream tasks without the need for knowledge distillation, grid search, or gradient-base iterative optimization. Unlike existing methods, the computation pattern of LQER eliminates the need for specialized Scatter and Gather processes to collect high-precision weights from irregular memory locations. Our W4A8 LLMs achieve near-lossless performance on six popular downstream tasks, while using 1.36$\times$ fewer hardware resources than the leading state-of-the-art method. We open-source our framework at https://github.com/ChengZhang-98/lqer
Authors: Junhong Shen, Neil Tenenholtz, James Brian Hall, David Alvarez-Melis, Nicolo Fusi
Abstract: Large Language Models (LLMs) have demonstrated remarkable proficiency in understanding and generating natural language. However, their capabilities wane in highly specialized domains underrepresented in the pretraining corpus, such as physical and biomedical sciences. This work explores how to repurpose general LLMs into effective task solvers for specialized domains. We introduce a novel, model-agnostic framework for learning custom input tags, which are parameterized as continuous vectors appended to the LLM's embedding layer, to condition the LLM. We design two types of input tags: domain tags are used to delimit specialized representations (e.g., chemical formulas) and provide domain-relevant context; function tags are used to represent specific functions (e.g., predicting molecular properties) and compress function-solving instructions. We develop a three-stage protocol to learn these tags using auxiliary data and domain knowledge. By explicitly disentangling task domains from task functions, our method enables zero-shot generalization to unseen problems through diverse combinations of the input tags. It also boosts LLM's performance in various specialized domains, such as predicting protein or chemical properties and modeling drug-target interactions, outperforming expert models tailored to these tasks.
Authors: Akbir Khan, John Hughes, Dan Valentine, Laura Ruis, Kshitij Sachan, Ansh Radhakrishnan, Edward Grefenstette, Samuel R. Bowman, Tim Rockt\"aschel, Ethan Perez
Abstract: Common methods for aligning large language models (LLMs) with desired behaviour heavily rely on human-labelled data. However, as models grow increasingly sophisticated, they will surpass human expertise, and the role of human evaluation will evolve into non-experts overseeing experts. In anticipation of this, we ask: can weaker models assess the correctness of stronger models? We investigate this question in an analogous setting, where stronger models (experts) possess the necessary information to answer questions and weaker models (non-experts) lack this information. The method we evaluate is debate, where two LLM experts each argue for a different answer, and a non-expert selects the answer. We find that debate consistently helps both non-expert models and humans answer questions, achieving 76% and 88% accuracy respectively (naive baselines obtain 48% and 60%). Furthermore, optimising expert debaters for persuasiveness in an unsupervised manner improves non-expert ability to identify the truth in debates. Our results provide encouraging empirical evidence for the viability of aligning models with debate in the absence of ground truth.
Authors: Siddharth Karamcheti, Suraj Nair, Ashwin Balakrishna, Percy Liang, Thomas Kollar, Dorsa Sadigh
Abstract: Visually-conditioned language models (VLMs) have seen growing adoption in applications such as visual dialogue, scene understanding, and robotic task planning; adoption that has fueled a wealth of new models such as LLaVa, InstructBLIP, and PaLI-3. Despite the volume of new releases, key design decisions around image preprocessing, architecture, and optimization are under-explored, making it challenging to understand what factors account for model performance $-$ a challenge further complicated by the lack of objective, consistent evaluations. To address these gaps, we first compile a suite of standardized evaluations spanning visual question answering, object localization, and challenge sets that probe properties such as hallucination; evaluations that provide fine-grained insight VLM capabilities. Second, we rigorously investigate VLMs along key design axes, including pretrained visual representations and training from base vs. instruct-tuned language models, amongst others. We couple our analysis with three resource contributions: (1) a unified framework for evaluating VLMs, (2) optimized, flexible training code, and (3) checkpoints for all models, including a family of VLMs at the 7-13B scale that strictly outperform InstructBLIP and LLaVa v1.5, the state-of-the-art in open VLMs.
Authors: Wentao Zhu, Zhining Zhang, Yizhou Wang
Abstract: Understanding and attributing mental states, known as Theory of Mind (ToM), emerges as a fundamental capability for human social reasoning. While Large Language Models (LLMs) appear to possess certain ToM abilities, the mechanisms underlying these capabilities remain elusive. In this study, we discover that it is possible to linearly decode the belief status from the perspectives of various agents through neural activations of language models, indicating the existence of internal representations of self and others' beliefs. By manipulating these representations, we observe dramatic changes in the models' ToM performance, underscoring their pivotal role in the social reasoning process. Additionally, our findings extend to diverse social reasoning tasks that involve different causal inference patterns, suggesting the potential generalizability of these representations.
Authors: Mehran Hosseini, Peyman Hosseini
Abstract: Scaled Dot Product Attention (SDPA) is the backbone of many modern deep-learning models. It is so versatile that it has been used in natural language, vision, and multi-modal domains with very little change compared to its original formulation. This paper discusses why the current formulation is inefficient by delving into the mathematical details of the attention mechanism. We propose three improvements to mitigate these inefficiencies, thereby, introducing three enhanced attention mechanisms: Optimised, Efficient, and Super Attention. Optimised and Efficient Attention have one and two matrix multiplications fewer per head, respectively, and 25% and 50% fewer parameters, respectively, than standard SDPA, but perform similarly to standard SDPA in both vision and natural language tasks. They can be used in all applications where SDPA is used while offering smaller model sizes and faster training and inference without noticeable loss in performance. Super Attention introduces a new linear transformation on the values, transforming them from the left. It outperforms standard SPDA on vision and natural language tasks by up to 17% while having one fewer matrix multiplication per head and 25% fewer parameters than standard SDPA. Consequently, it is also faster than standard SDPA. Super Attention is ideal in applications where the attention layer's context length is fixed, such as Vision Transformers. In addition to providing mathematical reasoning, we evaluate the presented attention mechanisms on several datasets including MNIST, CIFAR100, ImageNet, IMDB Movie Reviews, and Amazon Reviews datasets, as well as combined Europarl and Anki English-Spanish datasets for neural machine translation.
Authors: Qusai Abo Obaidah, Muhy Eddin Za'ter, Adnan Jaljuli, Ali Mahboub, Asma Hakouz, Bashar Al-Rfooh, Yazan Estaitia
Abstract: This work is an attempt to introduce a comprehensive benchmark for Arabic speech recognition, specifically tailored to address the challenges of telephone conversations in Arabic language. Arabic, characterized by its rich dialectal diversity and phonetic complexity, presents a number of unique challenges for automatic speech recognition (ASR) systems. These challenges are further amplified in the domain of telephone calls, where audio quality, background noise, and conversational speech styles negatively affect recognition accuracy. Our work aims to establish a robust benchmark that not only encompasses the broad spectrum of Arabic dialects but also emulates the real-world conditions of call-based communications. By incorporating diverse dialectical expressions and accounting for the variable quality of call recordings, this benchmark seeks to provide a rigorous testing ground for the development and evaluation of ASR systems capable of navigating the complexities of Arabic speech in telephonic contexts. This work also attempts to establish a baseline performance evaluation using state-of-the-art ASR technologies.
Authors: Zhijun Guo, Alvina Lai, Johan Hilge Thygesen, Joseph Farrington, Thomas Keen, Kezhi Li
Abstract: Large language models (LLMs) have attracted significant attention for potential applications in digital health, while their application in mental health is subject to ongoing debate. This systematic review aims to evaluate the usage of LLMs in mental health, focusing on their strengths and limitations in early screening, digital interventions, and clinical applications. Adhering to PRISMA guidelines, we searched PubMed, IEEE Xplore, Scopus, and the JMIR using keywords: 'mental health OR mental illness OR mental disorder OR psychiatry' AND 'large language models'. We included articles published between January 1, 2017, and December 31, 2023, excluding non-English articles. 30 articles were evaluated, which included research on mental illness and suicidal ideation detection through text (n=12), usage of LLMs for mental health conversational agents (CAs) (n=5), and other applications and evaluations of LLMs in mental health (n=13). LLMs exhibit substantial effectiveness in detecting mental health issues and providing accessible, de-stigmatized eHealth services. However, the current risks associated with the clinical use might surpass their benefits. The study identifies several significant issues: the lack of multilingual datasets annotated by experts, concerns about the accuracy and reliability of the content generated, challenges in interpretability due to the 'black box' nature of LLMs, and persistent ethical dilemmas. These include the lack of a clear ethical framework, concerns about data privacy, and the potential for over-reliance on LLMs by both therapists and patients, which could compromise traditional medical practice. Despite these issues, the rapid development of LLMs underscores their potential as new clinical aids, emphasizing the need for continued research and development in this area.
Authors: Tianbao Xie, Danyang Zhang, Jixuan Chen, Xiaochuan Li, Siheng Zhao, Ruisheng Cao, Toh Jing Hua, Zhoujun Cheng, Dongchan Shin, Fangyu Lei, Yitao Liu, Yiheng Xu, Shuyan Zhou, Silvio Savarese, Caiming Xiong, Victor Zhong, Tao Yu
Abstract: Autonomous agents that accomplish complex computer tasks with minimal human interventions have the potential to transform human-computer interaction, significantly enhancing accessibility and productivity. However, existing benchmarks either lack an interactive environment or are limited to environments specific to certain applications or domains, failing to reflect the diverse and complex nature of real-world computer use, thereby limiting the scope of tasks and agent scalability. To address this issue, we introduce OSWorld, the first-of-its-kind scalable, real computer environment for multimodal agents, supporting task setup, execution-based evaluation, and interactive learning across various operating systems such as Ubuntu, Windows, and macOS. OSWorld can serve as a unified, integrated computer environment for assessing open-ended computer tasks that involve arbitrary applications. Building upon OSWorld, we create a benchmark of 369 computer tasks involving real web and desktop apps in open domains, OS file I/O, and workflows spanning multiple applications. Each task example is derived from real-world computer use cases and includes a detailed initial state setup configuration and a custom execution-based evaluation script for reliable, reproducible evaluation. Extensive evaluation of state-of-the-art LLM/VLM-based agents on OSWorld reveals significant deficiencies in their ability to serve as computer assistants. While humans can accomplish over 72.36% of the tasks, the best model achieves only 12.24% success, primarily struggling with GUI grounding and operational knowledge. Comprehensive analysis using OSWorld provides valuable insights for developing multimodal generalist agents that were not possible with previous benchmarks. Our code, environment, baseline models, and data are publicly available at https://os-world.github.io.
Authors: Shu-wen Yang, Heng-Jui Chang, Zili Huang, Andy T. Liu, Cheng-I Lai, Haibin Wu, Jiatong Shi, Xuankai Chang, Hsiang-Sheng Tsai, Wen-Chin Huang, Tzu-hsun Feng, Po-Han Chi, Yist Y. Lin, Yung-Sung Chuang, Tzu-Hsien Huang, Wei-Cheng Tseng, Kushal Lakhotia, Shang-Wen Li, Abdelrahman Mohamed, Shinji Watanabe, Hung-yi Lee
Abstract: The foundation model paradigm leverages a shared foundation model to achieve state-of-the-art (SOTA) performance for various tasks, requiring minimal downstream-specific modeling and data annotation. This approach has proven crucial in the field of Natural Language Processing (NLP). However, the speech processing community lacks a similar setup to explore the paradigm systematically. In this work, we establish the Speech processing Universal PERformance Benchmark (SUPERB) to study the effectiveness of the paradigm for speech. We propose a unified multi-tasking framework to address speech processing tasks in SUPERB using a frozen foundation model followed by task-specialized, lightweight prediction heads. Combining our results with community submissions, we verify that the foundation model paradigm is promising for speech, and our multi-tasking framework is simple yet effective, as the best-performing foundation model shows competitive generalizability across most SUPERB tasks. For reproducibility and extensibility, we have developed a long-term maintained platform that enables deterministic benchmarking, allows for result sharing via an online leaderboard, and promotes collaboration through a community-driven benchmark database to support new development cycles. Finally, we conduct a series of analyses to offer an in-depth understanding of SUPERB and speech foundation models, including information flows across tasks inside the models, the correctness of the weighted-sum benchmarking protocol and the statistical significance and robustness of the benchmark.
Authors: Federico Moiraghi, Matteo Palmonari, Davide Allavena, Federico Morando
Abstract: Classifying public tenders is a useful task for both companies that are invited to participate and for inspecting fraudulent activities. To facilitate the task for both participants and public administrations, the European Union presented a common taxonomy (Common Procurement Vocabulary, CPV) which is mandatory for tenders of certain importance; however, the contracts in which a CPV label is mandatory are the minority compared to all the Public Administrations activities. Classifying over a real-world taxonomy introduces some difficulties that can not be ignored. First of all, some fine-grained classes have an insufficient (if any) number of observations in the training set, while other classes are far more frequent (even thousands of times) than the average. To overcome those difficulties, we present a zero-shot approach, based on a pre-trained language model that relies only on label description and respects the label taxonomy. To train our proposed model, we used industrial data, which comes from contrattipubblici.org, a service by SpazioDati s.r.l. that collects public contracts stipulated in Italy in the last 25 years. Results show that the proposed model achieves better performance in classifying low-frequent classes compared to three different baselines, and is also able to predict never-seen classes.
Authors: Zhenwei Shao, Zhou Yu, Jun Yu, Xuecheng Ouyang, Lihao Zheng, Zhenbiao Gai, Mingyang Wang, Jiajun Ding
Abstract: By harnessing the capabilities of large language models (LLMs), recent large multimodal models (LMMs) have shown remarkable versatility in open-world multimodal understanding. Nevertheless, they are usually parameter-heavy and computation-intensive, thus hindering their applicability in resource-constrained scenarios. To this end, several lightweight LMMs have been proposed successively to maximize the capabilities under constrained scale (e.g., 3B). Despite the encouraging results achieved by these methods, most of them only focus on one or two aspects of the design space, and the key design choices that influence model capability have not yet been thoroughly investigated. In this paper, we conduct a systematic study for lightweight LMMs from the aspects of model architecture, training strategy, and training data. Based on our findings, we obtain Imp -- a family of highly capable LMMs at the 2B-4B scales. Notably, our Imp-3B model steadily outperforms all the existing lightweight LMMs of similar size, and even surpasses the state-of-the-art LMMs at the 13B scale. With low-bit quantization and resolution reduction techniques, our Imp model can be deployed on a Qualcomm Snapdragon 8Gen3 mobile chip with a high inference speed of about 13 tokens/s.
Authors: Cong Lu, Shengran Hu, Jeff Clune
Abstract: Go-Explore is a powerful family of algorithms designed to solve hard-exploration problems, built on the principle of archiving discovered states, and iteratively returning to and exploring from the most promising states. This approach has led to superhuman performance across a wide variety of challenging problems including Atari games and robotic control, but requires manually designing heuristics to guide exploration, which is time-consuming and infeasible in general. To resolve this, we propose Intelligent Go-Explore (IGE) which greatly extends the scope of the original Go-Explore by replacing these heuristics with the intelligence and internalized human notions of interestingness captured by giant foundation models (FMs). This provides IGE with a human-like ability to instinctively identify how interesting or promising any new state is (e.g. discovering new objects, locations, or behaviors), even in complex environments where heuristics are hard to define. Moreover, IGE offers the exciting and previously impossible opportunity to recognize and capitalize on serendipitous discoveries that cannot be predicted ahead of time. We evaluate IGE on a range of language-based tasks that require search and exploration. In Game of 24, a multistep mathematical reasoning problem, IGE reaches 100% success rate 70.8% faster than the best classic graph search baseline. Next, in BabyAI-Text, a challenging partially observable gridworld, IGE exceeds the previous SOTA with orders of magnitude fewer online samples. Finally, in TextWorld, we show the unique ability of IGE to succeed in settings requiring long-horizon exploration where prior SOTA FM agents like Reflexion completely fail. Overall, IGE combines the tremendous strengths of FMs and the powerful Go-Explore algorithm, opening up a new frontier of research into creating more generally capable agents with impressive exploration capabilities.
Authors: Xiyao Wang, Jiuhai Chen, Zhaoyang Wang, Yuhang Zhou, Yiyang Zhou, Huaxiu Yao, Tianyi Zhou, Tom Goldstein, Parminder Bhatia, Furong Huang, Cao Xiao
Abstract: Large vision-language models (LVLMs) have achieved impressive results in various visual question-answering and reasoning tasks through vision instruction tuning on specific datasets. However, there is still significant room for improvement in the alignment between visual and language modalities. Previous methods to enhance this alignment typically require external models or data, heavily depending on their capabilities and quality, which inevitably sets an upper bound on performance. In this paper, we propose SIMA, a framework that enhances visual and language modality alignment through self-improvement, eliminating the needs for external models or data. SIMA leverages prompts from existing vision instruction tuning datasets to self-generate responses and employs an in-context self-critic mechanism to select response pairs for preference tuning. The key innovation is the introduction of three vision metrics during the in-context self-critic process, which can guide the LVLM in selecting responses that enhance image comprehension. Through experiments across 14 hallucination and comprehensive benchmarks, we demonstrate that SIMA not only improves model performance across all benchmarks but also achieves superior modality alignment, outperforming previous approaches.
Authors: Cong Zhang, Derrick Goh Xin Deik, Dexun Li, Hao Zhang, Yong Liu
Abstract: The rapid advancement of neural language models has sparked a new surge of intelligent agent research. Unlike traditional agents, large language model-based agents (LLM agents) have emerged as a promising paradigm for achieving artificial general intelligence (AGI) due to their superior reasoning and generalization capabilities. Effective planning is crucial for the success of LLM agents in real-world tasks, making it a highly pursued topic in the community. Current planning methods typically translate tasks into executable action sequences. However, determining a feasible or optimal sequence for complex tasks at fine granularity, which often requires compositing long chains of heterogeneous actions, remains challenging. This paper introduces Meta-Task Planning (MTP), a zero-shot methodology for collaborative LLM-based multi-agent systems that simplifies complex task planning by decomposing it into a hierarchy of subordinate tasks, or meta-tasks. Each meta-task is then mapped into executable actions. MTP was assessed on two rigorous benchmarks, TravelPlanner and API-Bank. Notably, MTP achieved an average $\sim40\%$ success rate on TravelPlanner, significantly higher than the state-of-the-art (SOTA) baseline ($2.92\%$), and outperforming $LLM_{api}$-4 with ReAct on API-Bank by $\sim14\%$, showing the immense potential of integrating LLM with multi-agent systems.
Authors: Hao Tang, Keya Hu, Jin Peng Zhou, Sicheng Zhong, Wei-Long Zheng, Xujie Si, Kevin Ellis
Abstract: Iteratively improving and repairing source code with large language models (LLMs), known as refinement, has emerged as a popular way of generating programs that would be too complex to construct in one shot. Given a bank of test cases, together with a candidate program, an LLM can improve that program by being prompted with failed test cases. But it remains an open question how to best iteratively refine code, with prior work employing simple greedy or breadth-first strategies. We show here that refinement exposes an explore-exploit tradeoff: exploit by refining the program that passes the most test cases, or explore by refining a lesser considered program. We frame this as an arm-acquiring bandit problem, which we solve with Thompson Sampling. The resulting LLM-based program synthesis algorithm is broadly applicable: Across loop invariant synthesis, visual reasoning puzzles, and competition programming problems, we find that our new method can solve more problems using fewer language model calls.