Authors: Qifan Zhang, Xiaobin Hong, Jianheng Tang, Nuo Chen, Yuhan Li, Wenzhong Li, Jing Tang, Jia Li
Abstract: Large Language Models (LLMs) have demonstrated strong reasoning abilities, making them suitable for complex tasks such as graph computation. Traditional reasoning steps paradigm for graph problems is hindered by unverifiable steps, limited long-term reasoning, and poor generalization to graph variations. To overcome these limitations, we introduce GCoder, a code-based LLM designed to enhance problem-solving in generalized graph computation problems. Our method involves constructing an extensive training dataset, GraphWild, featuring diverse graph formats and algorithms. We employ a multi-stage training process, including Supervised Fine-Tuning (SFT) and Reinforcement Learning from Compiler Feedback (RLCF), to refine model capabilities. For unseen tasks, a hybrid retrieval technique is used to augment performance. Experiments demonstrate that GCoder outperforms GPT-4o, with an average accuracy improvement of 16.42% across various graph computational problems. Furthermore, GCoder efficiently manages large-scale graphs with millions of nodes and diverse input formats, overcoming the limitations of previous models focused on the reasoning steps paradigm. This advancement paves the way for more intuitive and effective graph problem-solving using LLMs. Code and data are available at here: https://github.com/Bklight999/WWW25-GCoder/tree/master.
URLs: https://github.com/Bklight999/WWW25-GCoder/tree/master.
Authors: Dylan Wilson
Abstract: This project aims to investigate a novel sequence generation method inspired by the AlphaGo paradigm, adapting it for use with large language models (LLMs). The proposed approach involves creating search trees of different possible completions and evaluating these completions based on model confidence. By considering various paths in the search tree and scoring them according to the model's confidence in each completion, we can generate diverse and high-quality sequences. This research explores the implementation of this paradigm by using confidence as a proxy for response quality akin to beam search \citep{vijayakumar2016diverse}. The primary goal of this paper is to outline the paradigm and demonstrate its potential, rather than focusing on achieving perfect results. The paper will outline the reasons why we believe this paradigm has the potential to improve LLMs in the following manners: 1) increase output quality, 2) decrease errors, 3) eliminate or reduce the compound error problems, 4) generate diverse and creative completions, 5) allow for iterative problem-solving, and 6) self-training. We expect this approach to yield a set of diverse and coherent sequences, offering insights into balancing exploration and exploitation in sequence generation. Potential applications include creative text generation tasks, such as storytelling and content creation, as well as other natural language processing domains, like machine translation and automated summarization. The goal is that the model will be far more effective as it will be able to consider many possible variations allowing it to find the ideal completion. This research aims to contribute to the understanding of effective search strategies in sequence generation and their impact on generating high-quality, varied textual outputs.
Authors: Ruisi Cai, Yeonju Ro, Geon-Woo Kim, Peihao Wang, Babak Ehteshami Bejnordi, Aditya Akella, Zhangyang Wang
Abstract: The proliferation of large language models (LLMs) has led to the adoption of Mixture-of-Experts (MoE) architectures that dynamically leverage specialized subnetworks for improved efficiency and performance. Despite their benefits, MoE models face significant challenges during inference, including inefficient memory management and suboptimal batching, due to misaligned design choices between the model architecture and the system policies. Furthermore, the conventional approach of training MoEs from scratch is increasingly prohibitive in terms of cost. In this paper, we propose a novel framework Read-ME that transforms pre-trained dense LLMs into smaller MoE models (in contrast to "upcycling" generalist MoEs), avoiding the high costs of ground-up training. Our approach employs activation sparsity to extract experts. To compose experts, we examine the widely-adopted layer-wise router design and show its redundancy, and thus we introduce the pre-gating router decoupled from the MoE backbone that facilitates system-friendly pre-computing and lookahead scheduling, enhancing expert-aware batching and caching. Our codesign therefore addresses critical gaps on both the algorithmic and system fronts, establishing a scalable and efficient alternative for LLM inference in resource-constrained settings. Read-ME outperforms other popular open-source dense models of similar scales, achieving improvements of up to 10.1% on MMLU, and improving mean end-to-end latency up to 6.1%. Codes are available at: https://github.com/VITA-Group/READ-ME.
Authors: Guimin Hu
Abstract: Large language models (LLMs) have garnered significant attention in recent years due to their impressive performance. While considerable research has evaluated these models from various perspectives, the extent to which LLMs can perform implicit and explicit emotion retrieval remains largely unexplored. To address this gap, this study investigates LLMs' emotion retrieval capabilities in commonsense. Through extensive experiments involving multiple models, we systematically evaluate the ability of LLMs on emotion retrieval. Specifically, we propose a supervised contrastive probing method to verify LLMs' performance for implicit and explicit emotion retrieval, as well as the diversity of the emotional events they retrieve. The results offer valuable insights into the strengths and limitations of LLMs in handling emotion retrieval.
Authors: Lester James V. Miranda, Yizhong Wang, Yanai Elazar, Sachin Kumar, Valentina Pyatkin, Faeze Brahman, Noah A. Smith, Hannaneh Hajishirzi, Pradeep Dasigi
Abstract: Learning from human feedback has enabled the alignment of language models (LMs) with human preferences. However, directly collecting human preferences can be expensive, time-consuming, and can have high variance. An appealing alternative is to distill preferences from LMs as a source of synthetic annotations as they are more consistent, cheaper, and scale better than human annotation; however, they are also prone to biases and errors. In this work, we introduce a routing framework that combines inputs from humans and LMs to achieve better annotation quality, while reducing the total cost of human annotation. The crux of our approach is to identify preference instances that will benefit from human annotations. We formulate this as an optimization problem: given a preference dataset and an evaluation metric, we train a performance prediction model to predict a reward model's performance on an arbitrary combination of human and LM annotations and employ a routing strategy that selects a combination that maximizes predicted performance. We train the performance prediction model on MultiPref, a new preference dataset with 10K instances paired with human and LM labels. We show that the selected hybrid mixture of LM and direct human preferences using our routing framework achieves better reward model performance compared to using either one exclusively. We simulate selective human preference collection on three other datasets and show that our method generalizes well to all three. We analyze features from the routing model to identify characteristics of instances that can benefit from human feedback, e.g., prompts with a moderate safety concern or moderate intent complexity. We release the dataset, annotation platform, and source code used in this study to foster more efficient and accurate preference collection in the future.
Authors: Ziqi Liang, Haoxiang Shi, Hanhui Chen
Abstract: Speech Emotion Captioning (SEC) has gradually become an active research task. The emotional content conveyed through human speech are often complex, and classifying them into fixed categories may not be enough to fully capture speech emotions. Describing speech emotions through natural language may be a more effective approach. However, existing SEC methods often produce hallucinations and lose generalization on unseen speech. To overcome these problems, we propose AlignCap, which Aligning Speech Emotion Captioning to Human Preferences based on large language model (LLM) with two properties: 1) Speech-Text Alignment, which minimizing the divergence between the LLM's response prediction distributions for speech and text inputs using knowledge distillation (KD) Regularization. 2) Human Preference Alignment, where we design Preference Optimization (PO) Regularization to eliminate factuality and faithfulness hallucinations. We also extract emotional clues as a prompt for enriching fine-grained information under KD-Regularization. Experiments demonstrate that AlignCap presents stronger performance to other state-of-the-art methods on Zero-shot SEC task.
Authors: Mohit Chandra, Siddharth Sriraman, Gaurav Verma, Harneet Singh Khanuja, Jose Suarez Campayo, Zihang Li, Michael L. Birnbaum, Munmun De Choudhury
Abstract: Adverse Drug Reactions (ADRs) from psychiatric medications are the leading cause of hospitalizations among mental health patients. With healthcare systems and online communities facing limitations in resolving ADR-related issues, Large Language Models (LLMs) have the potential to fill this gap. Despite the increasing capabilities of LLMs, past research has not explored their capabilities in detecting ADRs related to psychiatric medications or in providing effective harm reduction strategies. To address this, we introduce the Psych-ADR benchmark and the Adverse Drug Reaction Response Assessment (ADRA) framework to systematically evaluate LLM performance in detecting ADR expressions and delivering expert-aligned mitigation strategies. Our analyses show that LLMs struggle with understanding the nuances of ADRs and differentiating between types of ADRs. While LLMs align with experts in terms of expressed emotions and tone of the text, their responses are more complex, harder to read, and only 70.86% aligned with expert strategies. Furthermore, they provide less actionable advice by a margin of 12.32% on average. Our work provides a comprehensive benchmark and evaluation framework for assessing LLMs in strategy-driven tasks within high-risk domains.
Authors: Israel Fama, B\'arbara Bueno, Alexandre Alcoforado, Thomas Palmeira Ferraz, Arnold Moya, Anna Helena Reali Costa
Abstract: In a context where the Brazilian judiciary system, the largest in the world, faces a crisis due to the slow processing of millions of cases, it becomes imperative to develop efficient methods for analyzing legal texts. We introduce uBERT, a hybrid model that combines Transformer and Recurrent Neural Network architectures to effectively handle long legal texts. Our approach processes the full text regardless of its length while maintaining reasonable computational overhead. Our experiments demonstrate that uBERT achieves superior performance compared to BERT+LSTM when overlapping input is used and is significantly faster than ULMFiT for processing long legal documents.
Authors: Bruno Croso Cunha da Silva, Thomas Palmeira Ferraz, Roseli De Deus Lopes
Abstract: Disinformation on social media poses both societal and technical challenges. While previous studies have integrated textual information into propagation networks, they have yet to fully leverage the advancements in Transformer-based language models for high-quality contextual text representations. This work investigates the impact of incorporating textual features into Graph Neural Networks (GNNs) for fake news detection. Our experiments demonstrate that contextual representations improve performance by 9.3% in Macro F1 over static ones and 33.8% over GNNs without textual features. However, noisy data augmentation degrades performance and increases instability. We expect our methodology to open avenues for further research, and all code is made publicly available.
Authors: Yue Li, Zhixue Zhao, Carolina Scarton
Abstract: In-context learning (ICL) performance is known to be sensitive to the prompt design, yet the impact of class label options in zero-shot classification has been largely overlooked. This study presents the first comprehensive empirical study investigating how label option (e.g., lexical choice, order, and elaboration) influences zero-shot ICL classification performance. Our findings reveal that lexical choices for label names (e.g., agree vs.support in stance classification) play an important role, with effects also linked to label orders. An analysis of the model internal states further shows that optimal label names tend to activate fewer outlier neurons in the feed forward network. Based on this observation, we propose Label set Optimization via Activation Distribution kurtosiS (LOADS), a post-hoc approach requiring no gradient propagation. LOADS not only demonstrates effectiveness with only 100 unlabelled samples across different model types and sizes, but also shows cross-lingual transferability.
Authors: Vahid Sadiri Javadi, Johanne R. Trippas, Yash Kumar Lal, Lucie Flek
Abstract: Narratives are widely recognized as a powerful tool for structuring information and facilitating comprehension of complex ideas in various domains such as science communication. This paper investigates whether incorporating narrative elements can assist Large Language Models (LLMs) in solving complex problems more effectively. We propose a novel approach, Story of Thought (SoT), integrating narrative structures into prompting techniques for problem-solving. This approach involves constructing narratives around problem statements and creating a framework to identify and organize relevant information. Our experiments show that using various LLMs with SoT consistently surpasses using them with other techniques on physics, chemistry, math, and biology questions in both the GPQA and JEEBench datasets. The narrative-based information curation process in SoT enhances problem comprehension by contextualizing critical in-domain information and highlighting causal relationships within the problem space.
Authors: Menna Fateen, Tsunenori Mine
Abstract: Recent advances in large language models (LLMs) have shown promise for scalable educational applications, but their use in dialog-based tutoring systems remains challenging due to the need for effective pedagogical strategies and the high costs associated with expert-curated datasets. Our study explores the use of smaller, more affordable LLMs for one-on-one tutoring in the context of solving reading comprehension problems. We developed a synthetic tutoring dialog dataset, evaluated by human teachers, and fine-tuned a smaller LLM using this dataset. Furthermore, we conducted an interactive experiment comparing the performance of the fine-tuned model with a larger model in real-world tutoring scenarios. Our results show that the fine-tuned model performs on par with the larger model but at a lower cost, demonstrating a viable, cost-effective approach for implementing LLM-based tutoring systems in educational settings.
Authors: Xinyu Wang, Wenbo Zhang, Sai Koneru, Hangzhi Guo, Bonam Mingole, S. Shyam Sundar, Sarah Rajtmajer, Amulya Yadav
Abstract: With the rise of AI-generated content spewed at scale from large language models (LLMs), genuine concerns about the spread of fake news have intensified. The perceived ability of LLMs to produce convincing fake news at scale poses new challenges for both human and automated fake news detection systems. To address this gap, this work presents the findings from a university-level competition which aimed to explore how LLMs can be used by humans to create fake news, and to assess the ability of human annotators and AI models to detect it. A total of 110 participants used LLMs to create 252 unique fake news stories, and 84 annotators participated in the detection tasks. Our findings indicate that LLMs are ~68% more effective at detecting real news than humans. However, for fake news detection, the performance of LLMs and humans remains comparable (~60% accuracy). Additionally, we examine the impact of visual elements (e.g., pictures) in news on the accuracy of detecting fake news stories. Finally, we also examine various strategies used by fake news creators to enhance the credibility of their AI-generated content. This work highlights the increasing complexity of detecting AI-generated fake news, particularly in collaborative human-AI settings.
Authors: Yu Fu, Zefan Cai, Abedelkadir Asi, Wayne Xiong, Yue Dong, Wen Xiao
Abstract: Key-Value (KV) caching is a common technique to enhance the computational efficiency of Large Language Models (LLMs), but its memory overhead grows rapidly with input length. Prior work has shown that not all tokens are equally important for text generation, proposing layer-level KV cache compression to selectively retain key information. Recognizing the distinct roles of attention heads in generation, we propose HeadKV, a head-level KV cache compression method, and HeadKV-R2, which leverages a novel contextual reasoning ability estimation for compression. Our approach operates at the level of individual heads, estimating their importance for contextual QA tasks that require both retrieval and reasoning capabilities. Extensive experiments across diverse benchmarks (LongBench, LooGLE), model architectures (e.g., Llama-3-8B-Instruct, Mistral-7B-Instruct), and long-context abilities tests demonstrate that our head-level KV cache compression significantly outperforms strong baselines, particularly in low-resource settings (KV size = 64 & 128). Notably, our method retains just 1.5% of the KV cache while achieving 97% of the performance of the full KV cache on the contextual question answering benchmark.
Authors: Yujian Liu, Shiyu Chang, Tommi Jaakkola, Yang Zhang
Abstract: Recent studies have identified one aggravating factor of LLM hallucinations as the knowledge inconsistency between pre-training and fine-tuning, where unfamiliar fine-tuning data mislead the LLM to fabricate plausible but wrong outputs. In this paper, we propose a novel fine-tuning strategy called Prereq-Tune to address this knowledge inconsistency and reduce hallucinations. Fundamentally, Prereq-Tune disentangles the learning of skills and knowledge, so the model learns only the task skills without being impacted by the knowledge inconsistency. To achieve this, Prereq-Tune introduces an additional prerequisite learning stage to learn the necessary knowledge for SFT, allowing subsequent SFT to focus only on task skills. Prereq-Tune can also be combined with fictitious synthetic data to enhance the grounding of LLM outputs to their internal knowledge. Experiments show that Prereq-Tune outperforms existing baselines in improving LLM's factuality across short QA and long-form generation tasks. It also opens new possibilities for knowledge-controlled generation in LLMs. Our code is available at https://github.com/UCSB-NLP-Chang/Prereq_tune.git.
Authors: Abhijnan Nath, Videep Venkatesha, Mariah Bradford, Avyakta Chelle, Austin Youngren, Carlos Mabrey, Nathaniel Blanchard, Nikhil Krishnaswamy
Abstract: Question-asking in collaborative dialogue has long been established as key to knowledge construction, both in internal and collaborative problem solving. In this work, we examine probing questions in collaborative dialogues: questions that explicitly elicit responses from the speaker's interlocutors. Specifically, we focus on modeling the causal relations that lead directly from utterances earlier in the dialogue to the emergence of the probing question. We model these relations using a novel graph-based framework of deliberation chains, and reframe the problem of constructing such chains as a coreference-style clustering problem. Our framework jointly models probing and causal utterances and the links between them, and we evaluate on two challenging collaborative task datasets: the Weights Task and DeliData. Our results demonstrate the effectiveness of our theoretically-grounded approach compared to both baselines and stronger coreference approaches, and establish a standard of performance in this novel task.
Authors: Zhiting Fan, Ruizhe Chen, Tianxiang Hu, Zuozhu Liu
Abstract: The growing use of large language model (LLM)-based chatbots has raised concerns about fairness. Fairness issues in LLMs can lead to severe consequences, such as bias amplification, discrimination, and harm to marginalized communities. While existing fairness benchmarks mainly focus on single-turn dialogues, multi-turn scenarios, which in fact better reflect real-world conversations, present greater challenges due to conversational complexity and potential bias accumulation. In this paper, we propose a comprehensive fairness benchmark for LLMs in multi-turn dialogue scenarios, \textbf{FairMT-Bench}. Specifically, we formulate a task taxonomy targeting LLM fairness capabilities across three stages: context understanding, user interaction, and instruction trade-offs, with each stage comprising two tasks. To ensure coverage of diverse bias types and attributes, we draw from existing fairness datasets and employ our template to construct a multi-turn dialogue dataset, \texttt{FairMT-10K}. For evaluation, GPT-4 is applied, alongside bias classifiers including Llama-Guard-3 and human validation to ensure robustness. Experiments and analyses on \texttt{FairMT-10K} reveal that in multi-turn dialogue scenarios, current LLMs are more likely to generate biased responses, and there is significant variation in performance across different tasks and models. Based on this, we curate a challenging dataset, \texttt{FairMT-1K}, and test 15 current state-of-the-art (SOTA) LLMs on this dataset. The results show the current state of fairness in LLMs and showcase the utility of this novel approach for assessing fairness in more realistic multi-turn dialogue contexts, calling for future work to focus on LLM fairness improvement and the adoption of \texttt{FairMT-1K} in such efforts.
Authors: Wei Han, Pan Zhou, Soujanya Poria, Shuicheng Yan
Abstract: The limited context window of contemporary large language models (LLMs) remains a huge barrier to their broader application across various domains. While continual pre-training on long-context data is a straightforward and effective solution, it incurs substantial costs in terms of data acquisition and computational resources. To alleviate this issue, we propose SharedLLM, a novel approach grounded in the design philosophy of multi-grained context compression and query-aware information retrieval. SharedLLM is composed of two short-context LLMs such as LLaMA-2, termed upper model and lower model. The lower model functions as a compressor while the upper model acts as a decoder. The upper model receives compressed, multi-grained context information from the lower model and performs context-aware modeling on the running text. Information transfer between the compressor and decoder occurs only at the lowest layers to refrain from long forward paths in the lower model and redundant cross-attention modules in the upper model. Based on this architecture, we introduce a specialized tree-style data structure to efficiently encode, store and retrieve multi-grained contextual information for text chunks. This structure, combined with a search algorithm, enables rapid encoding and retrieval of relevant information from various levels of the tree based on the input query. This entire process, wherein the sender and receiver are derived from the same LLM layer, is referred to as self-injection.
Authors: Xinyi Mou, Jingcong Liang, Jiayu Lin, Xinnong Zhang, Xiawei Liu, Shiyue Yang, Rong Ye, Lei Chen, Haoyu Kuang, Xuanjing Huang, Zhongyu Wei
Abstract: Large language models (LLMs) are increasingly leveraged to empower autonomous agents to simulate human beings in various fields of behavioral research. However, evaluating their capacity to navigate complex social interactions remains a challenge. Previous studies face limitations due to insufficient scenario diversity, complexity, and a single-perspective focus. To this end, we introduce AgentSense: Benchmarking Social Intelligence of Language Agents through Interactive Scenarios. Drawing on Dramaturgical Theory, AgentSense employs a bottom-up approach to create 1,225 diverse social scenarios constructed from extensive scripts. We evaluate LLM-driven agents through multi-turn interactions, emphasizing both goal completion and implicit reasoning. We analyze goals using ERG theory and conduct comprehensive experiments. Our findings highlight that LLMs struggle with goals in complex social scenarios, especially high-level growth needs, and even GPT-4o requires improvement in private information reasoning.
Authors: Marvin Alberts, Gianmarco Gabrieli, Irina Espejo Morales
Abstract: Integrating text and numbers effectively is a crucial step towards enhancing Large Language Models (LLMs) capabilities in assisting in scientific tasks. While most current approaches rely on discrete tokenization of numbers, for instance, conversion to scientific notation or base 10-decomposition, a recent approach proposed a continuous numerical encoding as an inductive bias. In this paper, we build upon this approach by introducing more expressive numerical embeddings. Our method addresses key shortcomings, including the elimination of numerical artefacts and the ability to handle a wide range of magnitudes without clipping. Our work presents two key contributions. First, we employ an MLP to assign distinct directions in the embedding space to different numbers. Our second contribution is the introduction of a routing layer that differentiates between numerical and text embeddings. We hypothesise that this combined approach enables the model to distinguish between text and number distributions while maintaining its capacity for arithmetic operations. Using only a 45 M parameter encoder-decoder architecture our method achieves a $R^2$=0.9988 over a wide range of magnitude ($10^{-3},10^{8}$). In addition, we empirically observe a reduction of the numerical artefacts and biases observed compared to the baselines.
Authors: Liam Barkley, Brink van der Merwe
Abstract: Large Language Models (LLMs) are powerful computational models trained on extensive corpora of human-readable text, enabling them to perform general-purpose language understanding and generation. LLMs have garnered significant attention in both industry and academia due to their exceptional performance across various natural language processing (NLP) tasks. Despite these successes, LLMs often produce inaccuracies, commonly referred to as hallucinations. Prompt engineering, the process of designing and formulating instructions for LLMs to perform specific tasks, has emerged as a key approach to mitigating hallucinations. This paper provides a comprehensive empirical evaluation of different prompting strategies and frameworks aimed at reducing hallucinations in LLMs. Various prompting techniques are applied to a broad set of benchmark datasets to assess the accuracy and hallucination rate of each method. Additionally, the paper investigates the influence of tool-calling agents (LLMs augmented with external tools to enhance their capabilities beyond language generation) on hallucination rates in the same benchmarks. The findings demonstrate that the optimal prompting technique depends on the type of problem, and that simpler techniques often outperform more complex methods in reducing hallucinations. Furthermore, it is shown that LLM agents can exhibit significantly higher hallucination rates due to the added complexity of external tool usage.
Authors: Hamna, Deepthi Sudharsan, Agrima Seth, Ritvik Budhiraja, Deepika Khullar, Vyshak Jain, Kalika Bali, Aditya Vashistha, Sameer Segal
Abstract: Large Language Models (LLMs) and Text-To-Image (T2I) models have demonstrated the ability to generate compelling text and visual stories. However, their outputs are predominantly aligned with the sensibilities of the Global North, often resulting in an outsider's gaze on other cultures. As a result, non-Western communities have to put extra effort into generating culturally specific stories. To address this challenge, we developed a visual storytelling pipeline called KAHANI that generates culturally grounded visual stories for non-Western cultures. Our pipeline leverages off-the-shelf models GPT-4 Turbo and Stable Diffusion XL (SDXL). By using Chain of Thought (CoT) and T2I prompting techniques, we capture the cultural context from user's prompt and generate vivid descriptions of the characters and scene compositions. To evaluate the effectiveness of KAHANI, we conducted a comparative user study with ChatGPT-4 (with DALL-E3) in which participants from different regions of India compared the cultural relevance of stories generated by the two tools. Results from the qualitative and quantitative analysis performed on the user study showed that KAHANI was able to capture and incorporate more Culturally Specific Items (CSIs) compared to ChatGPT-4. In terms of both its cultural competence and visual story generation quality, our pipeline outperformed ChatGPT-4 in 27 out of the 36 comparisons.
Authors: Yirui Chen, Qinyu Xiao, Jia Yi, Jing Chen, Mengyang Wang
Abstract: This paper explores the application of prompt engineering to enhance the performance of large language models (LLMs) in the domain of Traditional Chinese Medicine (TCM). We propose TCM-Prompt, a framework that integrates various pre-trained language models (PLMs), templates, tokenization, and verbalization methods, allowing researchers to easily construct and fine-tune models for specific TCM-related tasks. We conducted experiments on disease classification, syndrome identification, herbal medicine recommendation, and general NLP tasks, demonstrating the effectiveness and superiority of our approach compared to baseline methods. Our findings suggest that prompt engineering is a promising technique for improving the performance of LLMs in specialized domains like TCM, with potential applications in digitalization, modernization, and personalized medicine.
Authors: Hengyuan Zhang, Chenming Shang, Sizhe Wang, Dongdong Zhang, Feng Yao, Renliang Sun, Yiyao Yu, Yujiu Yang, Furu Wei
Abstract: Although fine-tuning Large Language Models (LLMs) with multilingual data can rapidly enhance the multilingual capabilities of LLMs, they still exhibit a performance gap between the dominant language (e.g., English) and non-dominant ones due to the imbalance of training data across languages. To further enhance the performance of non-dominant languages, we propose ShifCon, a Shift-based Contrastive framework that aligns the internal forward process of other languages toward that of the dominant one. Specifically, it shifts the representations of non-dominant languages into the dominant language subspace, allowing them to access relatively rich information encoded in the model parameters. The enriched representations are then shifted back into their original language subspace before generation. Moreover, we introduce a subspace distance metric to pinpoint the optimal layer area for shifting representations and employ multilingual contrastive learning to further enhance the alignment of representations within this area. Experiments demonstrate that our ShifCon framework significantly enhances the performance of non-dominant languages, particularly for low-resource ones. Further analysis offers extra insights to verify the effectiveness of ShifCon and propel future research
Authors: Ray Li, Tanishka Bagade, Kevin Martinez, Flora Yasmin, Grant Ayala, Michael Lam, Kevin Zhu
Abstract: Large language models (LLMs) have achieved a degree of success in generating coherent and contextually relevant text, yet they remain prone to a significant challenge known as hallucination: producing information that is not substantiated by the input or external knowledge. Previous efforts to mitigate hallucinations have focused on techniques such as fine-tuning models on high-quality datasets, incorporating fact-checking mechanisms, and developing adversarial training methods. While these approaches have shown some promise, they often address the issue at the level of individual model outputs, leaving unexplored the effects of inter-model interactions on hallucination. This study investigates the phenomenon of hallucination in LLMs through a novel experimental framework where multiple instances of GPT-4o-Mini models engage in a debate-like interaction prompted with questions from the TruthfulQA dataset. One model is deliberately instructed to generate plausible but false answers while the other models are asked to respond truthfully. The experiment is designed to assess whether the introduction of misinformation by one model can challenge the truthful majority to better justify their reasoning, improving performance on the TruthfulQA benchmark. The findings suggest that inter-model interactions can offer valuable insights into improving the accuracy and robustness of LLM outputs, complementing existing mitigation strategies.
Authors: Christos Xypolopoulos, Guokan Shang, Xiao Fei, Giannis Nikolentzos, Hadi Abdine, Iakovos Evdaimon, Michail Chatzianastasis, Giorgos Stamou, Michalis Vazirgiannis
Abstract: Large language models have evolved to process multiple modalities beyond text, such as images and audio, which motivates us to explore how to effectively leverage them for graph machine learning tasks. The key question, therefore, is how to transform graphs into linear sequences of tokens, a process we term graph linearization, so that LLMs can handle graphs naturally. We consider that graphs should be linearized meaningfully to reflect certain properties of natural language text, such as local dependency and global alignment, in order to ease contemporary LLMs, trained on trillions of textual tokens, better understand graphs. To achieve this, we developed several graph linearization methods based on graph centrality, degeneracy, and node relabeling schemes. We then investigated their effect on LLM performance in graph reasoning tasks. Experimental results on synthetic graphs demonstrate the effectiveness of our methods compared to random linearization baselines. Our work introduces novel graph representations suitable for LLMs, contributing to the potential integration of graph machine learning with the trend of multi-modal processing using a unified transformer model.
Authors: Anthony Cui, Pranav Nandyalam, Kevin Zhu
Abstract: Momentum-Aided Prompt Optimization (MAPO) enhances the efficiency and efficacy of prompt optimization for Large Language Models (LLMs). Building on ProTeGi, MAPO uses positive natural language "gradients" and a momentum-based extension to refine prompts effectively. By tracking gradient history, MAPO avoids local minima and oscillations. It also utilizes beam search and an Upper Confidence Bound (UCB) algorithm for balanced candidate expansion and selection. Benchmark testing shows that MAPO achieves faster convergence time with fewer API calls and higher F1 scores than ProTeGi, proving it as a robust and scalable solution for automated prompt engineering in LLMs.
Authors: Jahyun Koo, Yerin Hwang, Yongil Kim, Taegwan Kang, Hyunkyung Bae, Kyomin Jung
Abstract: Despite the success of Large Language Models (LLMs), they still face challenges related to high inference costs and memory requirements. To address these issues, Knowledge Distillation (KD) has emerged as a popular method for model compression, with student-generated outputs (SGOs) being particularly notable for reducing the mismatch between training and inference. However, SGOs often produce noisy and biased sequences, which can lead to misguidance from the teacher model, especially in long sequences. To mitigate these challenges, we propose SWITCH (Studying WIth TeaCHer for Knowledge Distillation), a novel approach that strategically incorporates the teacher model during the student's sequence generation. SWITCH identifies discrepancies between the token probabilities of the teacher and student models, allowing the teacher to intervene selectively, particularly in long sequences that are more prone to teacher misguidance. Extensive experimental results across three model families and five instruction-following datasets show that SWITCH surpasses traditional KD methods, particularly excelling in the generation of long sequential data.
Authors: Muhammad Zain Ali, Yuxia Wang, Bernhard Pfahringer, Tony Smith
Abstract: The rise of social media has amplified the spread of fake news, now further complicated by large language models (LLMs) like ChatGPT, which ease the generation of highly convincing, error-free misinformation, making it increasingly challenging for the public to discern truth from falsehood. Traditional fake news detection methods relying on linguistic cues also becomes less effective. Moreover, current detectors primarily focus on binary classification and English texts, often overlooking the distinction between machine-generated true vs. fake news and the detection in low-resource languages. To this end, we updated detection schema to include machine-generated news with focus on the Urdu language. We further propose a hierarchical detection strategy to improve the accuracy and robustness. Experiments show its effectiveness across four datasets in various settings.
Authors: Ritvik Aggarwal Ishneet Sukhvinder Singh Ibrahim Allahverdiyev, Muhammad Taha, Aslihan Akalin, Kevin Zhu
Abstract: Retrieval-Augmented Generation (RAG) systems using large language models (LLMs) often generate inaccurate responses due to the retrieval of irrelevant or loosely related information. Existing methods, which operate at the document level, fail to effectively filter out such content. We propose LLM-driven chunk filtering, ChunkRAG, a framework that enhances RAG systems by evaluating and filtering retrieved information at the chunk level. Our approach employs semantic chunking to divide documents into coherent sections and utilizes LLM-based relevance scoring to assess each chunk's alignment with the user's query. By filtering out less pertinent chunks before the generation phase, we significantly reduce hallucinations and improve factual accuracy. Experiments show that our method outperforms existing RAG models, achieving higher accuracy on tasks requiring precise information retrieval. This advancement enhances the reliability of RAG systems, making them particularly beneficial for applications like fact-checking and multi-hop reasoning.
Authors: Hongliang He, Wenlin Yao, Kaixin Ma, Wenhao Yu, Hongming Zhang, Tianqing Fang, Zhenzhong Lan, Dong Yu
Abstract: The rapid development of large language and multimodal models has sparked significant interest in using proprietary models, such as GPT-4o, to develop autonomous agents capable of handling real-world scenarios like web navigation. Although recent open-source efforts have tried to equip agents with the ability to explore environments and continuously improve over time, they are building text-only agents in synthetic environments where the reward signals are clearly defined. Such agents struggle to generalize to realistic settings that require multimodal perception abilities and lack ground-truth signals. In this paper, we introduce an open-source framework designed to facilitate the development of multimodal web agent that can autonomously conduct real-world exploration and improve itself. We first train the base model with imitation learning to gain the basic abilities. We then let the agent explore the open web and collect feedback on its trajectories. After that, it further improves its policy by learning from well-performing trajectories judged by another general-purpose model. This exploration-feedback-optimization cycle can continue for several iterations. Experimental results show that our web agent successfully improves itself after each iteration, demonstrating strong performance across multiple test sets.
Authors: Riccardo Rende, Federica Gerace, Alessandro Laio, Sebastian Goldt
Abstract: The remarkable capability of over-parameterised neural networks to generalise effectively has been explained by invoking a ``simplicity bias'': neural networks prevent overfitting by initially learning simple classifiers before progressing to more complex, non-linear functions. While simplicity biases have been described theoretically and experimentally in feed-forward networks for supervised learning, the extent to which they also explain the remarkable success of transformers trained with self-supervised techniques remains unclear. In our study, we demonstrate that transformers, trained on natural language data, also display a simplicity bias. Specifically, they sequentially learn many-body interactions among input tokens, reaching a saturation point in the prediction error for low-degree interactions while continuing to learn high-degree interactions. To conduct this analysis, we develop a procedure to generate \textit{clones} of a given natural language data set, which rigorously capture the interactions between tokens up to a specified order. This approach opens up the possibilities of studying how interactions of different orders in the data affect learning, in natural language processing and beyond.
Authors: Abhay Kumar, Vigneshwaran Shankaran, Rajesh Sharma
Abstract: In the recent years online social media platforms has been flooded with hateful remarks such as racism, sexism, homophobia etc. As a result, there have been many measures taken by various social media platforms to mitigate the spread of hate-speech over the internet. One particular concept within the domain of hate speech is instigating hate, which involves provoking hatred against a particular community, race, colour, gender, religion or ethnicity. In this work, we introduce \textit{ProvocationProbe} - a dataset designed to explore what distinguishes instigating hate speech from general hate speech. For this study, we collected around twenty thousand tweets from Twitter, encompassing a total of nine global controversies. These controversies span various themes including racism, politics, and religion. In this paper, i) we present an annotated dataset after comprehensive examination of all the controversies, ii) we also highlight the difference between hate speech and instigating hate speech by identifying distinguishing features, such as targeted identity attacks and reasons for hate.
Authors: Clemencia Siro, Yifei Yuan, Mohammad Aliannejadi, Maarten de Rijke
Abstract: Generating diverse and effective clarifying questions is crucial for improving query understanding and retrieval performance in open-domain conversational search (CS) systems. We propose AGENT-CQ (Automatic GENeration, and evaluaTion of Clarifying Questions), an end-to-end LLM-based framework addressing the challenges of scalability and adaptability faced by existing methods that rely on manual curation or template-based approaches. AGENT-CQ consists of two stages: a generation stage employing LLM prompting strategies to generate clarifying questions, and an evaluation stage (CrowdLLM) that simulates human crowdsourcing judgments using multiple LLM instances to assess generated questions and answers based on comprehensive quality metrics. Extensive experiments on the ClariQ dataset demonstrate CrowdLLM's effectiveness in evaluating question and answer quality. Human evaluation and CrowdLLM show that the AGENT-CQ - generation stage, consistently outperforms baselines in various aspects of question and answer quality. In retrieval-based evaluation, LLM-generated questions significantly enhance retrieval effectiveness for both BM25 and cross-encoder models compared to human-generated questions.
Authors: Yifei Zhang, Hao Zhu, Aiwei Liu, Han Yu, Piotr Koniusz, Irwin King
Abstract: Fine-tuning Large Language Models (LLMs) has become a crucial technique for adapting pre-trained models to downstream tasks. However, the enormous size of LLMs poses significant challenges in terms of computational complexity and resource requirements. Low-Rank Adaptation (LoRA) has emerged as a promising solution. However, there exists a gap between the practical performance of low-rank adaptations and its theoretical optimum. In this work, we propose eXtreme Gradient Boosting LoRA (XGBLoRA), a novel framework that bridges this gap by leveraging the power of ensemble learning. Inspired by gradient boosting, XGBLoRA iteratively learns and merges a sequence of LoRA adaptations to refine model predictions. It achieves better performance than the standard LoRA, while enjoying the computational efficiency of rank-1 adaptations. We provide theoretical analysis to show the convergence and optimality of our approach, and conduct extensive experiments on a range of natural language processing tasks. The results demonstrate that XGBLoRA consistently outperforms standard LoRA and achieves performance comparable to full fine-tuning with significantly fewer trainable parameters. This work advances parameter-efficient fine-tuning for LLMs, and offers a promising solution for adapting LLMs to downstream tasks while optimizing performance and efficiency.
Authors: Shilong Li, Yancheng He, Hui Huang, Xingyuan Bu, Jiaheng Liu, Hangyu Guo, Weixun Wang, Jihao Gu, Wenbo Su, Bo Zheng
Abstract: Recent advancements in Direct Preference Optimization (DPO) have significantly enhanced the alignment of Large Language Models (LLMs) with human preferences, owing to its simplicity and effectiveness. However, existing methods typically optimize a scalar score or ranking reward, thereby overlooking the multi-dimensional nature of human preferences. In this work, we propose to extend the preference of DPO to two dimensions: segments and aspects. We first introduce a 2D supervision dataset called HelpSteer-2D. For the segment dimension, we divide the response into sentences and assign scores to each segment. For the aspect dimension, we meticulously design several criteria covering the response quality rubrics. With the 2-dimensional signals as feedback, we develop a 2D-DPO framework, decomposing the overall objective into multi-segment and multi-aspect objectives. Extensive experiments on popular benchmarks demonstrate that 2D-DPO performs better than methods that optimize for scalar or 1-dimensional preferences.
Authors: Xiang Zhang, Juntai Cao, Chenyu You
Abstract: Transformers, the backbone of modern large language models (LLMs), face inherent architectural limitations that impede their reasoning capabilities. Unlike recurrent networks, Transformers lack recurrent connections, confining them to constant-depth computation. This restriction places them in the complexity class TC$^0$, making them theoretically incapable of solving tasks that demand increasingly deep reasoning as input length grows. Counting, a fundamental component of many reasoning tasks, also requires reasoning depth to grow linearly to be performed inductively. While previous studies have established the upper limits of counting ability in Transformer-based expert models (i.e., models specifically trained for counting tasks), these findings do not directly extend to general-purpose LLMs due to differences in reasoning mechanisms. Recent work has highlighted how Chain of Thought (CoT) reasoning can help alleviate some of the architectural limitations of Transformers in counting tasks. However, little attention has been paid to the role of tokenization in these models. Unlike expert models that often use character-level tokenization, LLMs typically rely on byte-level (BPE) tokenizers, which fundamentally alters the way reasoning is processed. Our work investigates the impact of tokenization on the counting abilities of LLMs, uncovering substantial performance variations based on input tokenization differences. We provide both theoretical and experimental analyses, offering insights into how tokenization choices can undermine models' theoretical computability, thereby inspiring the design of new tokenization methods to enhance reasoning in LLMs.
Authors: Yucheng Zhou, Zhi Rao, Jun Wan, Jianbing Shen
Abstract: Large Vision-Language Models (LVLMs) excel in cross-model tasks but experience performance declines in long-context reasoning due to overreliance on textual information and reduced visual dependency. In this study, we empirically analyze LVLMs in long-context reasoning, revealing that increased context length leads to a higher dependence on language at the expense of visual dependency. To address this issue, we propose a novel training-free context pruning method that selectively removes less critical textual information. Our approach enhances visual dependency and reduces textual noise, thereby improving LVLM performance in long-context reasoning. We validate our method by constructing a long-context dataset, demonstrating its effectiveness across various LVLMs. Moreover, further analysis confirms the robustness of different token pruning strategies and preliminary explores scaling laws between pruning rates and context length.
Authors: Shawn Tan, Yikang Shen, Songlin Yang, Aaron Courville, Rameswar Panda
Abstract: The self-attention mechanism traditionally relies on the softmax operator, necessitating positional embeddings like RoPE, or position biases to account for token order. But current methods using still face length generalisation challenges. We propose an alternative attention mechanism based on the stick-breaking process: For each token before the current, we determine a break point $\beta_{i,j}$, which represents the proportion of the remaining stick to allocate to the current token. We repeat the process until the stick is fully allocated, resulting in a sequence of attention weights. This process naturally incorporates recency bias, which has linguistic motivations for grammar parsing (Shen et. al., 2017). We study the implications of replacing the conventional softmax-based attention mechanism with stick-breaking attention. We then discuss implementation of numerically stable stick-breaking attention and adapt Flash Attention to accommodate this mechanism. When used as a drop-in replacement for current softmax+RoPE attention systems, we find that stick-breaking attention performs competitively with current methods on length generalisation and downstream tasks. Stick-breaking also performs well at length generalisation, allowing a model trained with $2^{11}$ context window to perform well at $2^{14}$ with perplexity improvements.
Authors: Yiwei Qin, Xuefeng Li, Haoyang Zou, Yixiu Liu, Shijie Xia, Zhen Huang, Yixin Ye, Weizhe Yuan, Hector Liu, Yuanzhi Li, Pengfei Liu
Abstract: This paper introduces a pioneering approach to artificial intelligence research, embodied in our O1 Replication Journey. In response to the announcement of OpenAI's groundbreaking O1 model, we embark on a transparent, real-time exploration to replicate its capabilities while reimagining the process of conducting and communicating AI research. Our methodology addresses critical challenges in modern AI research, including the insularity of prolonged team-based projects, delayed information sharing, and the lack of recognition for diverse contributions. By providing comprehensive, real-time documentation of our replication efforts, including both successes and failures, we aim to foster open science, accelerate collective advancement, and lay the groundwork for AI-driven scientific discovery. Our research progress report diverges significantly from traditional research papers, offering continuous updates, full process transparency, and active community engagement throughout the research journey. Technologically, we proposed the journey learning paradigm, which encourages models to learn not just shortcuts, but the complete exploration process, including trial and error, reflection, and backtracking. With only 327 training samples and without any additional tricks, journey learning outperformed conventional supervised learning by over 8\% on the MATH dataset, demonstrating its extremely powerful potential. We believe this to be the most crucial component of O1 technology that we have successfully decoded. We share valuable resources including technical hypotheses and insights, cognitive exploration maps, custom-developed tools, etc at https://github.com/GAIR-NLP/O1-Journey.
Authors: Revanth Gangi Reddy, Sagnik Mukherjee, Jeonghwan Kim, Zhenhailong Wang, Dilek Hakkani-Tur, Heng Ji
Abstract: Despite seemingly performant web agents on the task-completion benchmarks, most existing methods evaluate the agents based on a presupposition: the web navigation task consists of linear sequence of actions with an end state that marks task completion. In contrast, our work focuses on web navigation for information aggregation, wherein the agent must explore different websites to gather information for a complex query. We consider web information aggregation from two different perspectives: (i) Direct API-driven Access relies on a text-only view of the Web, leveraging external tools such as Google Search API to navigate the web and a scraper to extract website contents. (ii) Interactive Visual Access uses screenshots of the webpages and requires interaction with the browser to navigate and access information. Motivated by these diverse information access settings, we introduce Infogent, a novel modular framework for web information aggregation involving three distinct components: Navigator, Extractor and Aggregator. Experiments on different information access settings demonstrate Infogent beats an existing SOTA multi-agent search framework by 7% under Direct API-Driven Access on FRAMES, and improves over an existing information-seeking web agent by 4.3% under Interactive Visual Access on AssistantBench.
Authors: Ruisi Zhang, Farinaz Koushanfar
Abstract: The widely adopted and powerful generative large language models (LLMs) have raised concerns about intellectual property rights violations and the spread of machine-generated misinformation. Watermarking serves as a promising approch to establish ownership, prevent unauthorized use, and trace the origins of LLM-generated content. This paper summarizes and shares the challenges and opportunities we found when watermarking LLMs. We begin by introducing techniques for watermarking LLMs themselves under different threat models and scenarios. Next, we investigate watermarking methods designed for the content generated by LLMs, assessing their effectiveness and resilience against various attacks. We also highlight the importance of watermarking domain-specific models and data, such as those used in code generation, chip design, and medical applications. Furthermore, we explore methods like hardware acceleration to improve the efficiency of the watermarking process. Finally, we discuss the limitations of current approaches and outline future research directions for the responsible use and protection of these generative AI tools.
Authors: Yifan Wang, Vera Demberg
Abstract: Despite significant advancements in natural language generation, controlling language models to produce texts with desired attributes remains a formidable challenge. In this work, we introduce RSA-Control, a training-free controllable text generation framework grounded in pragmatics. RSA-Control directs the generation process by recursively reasoning between imaginary speakers and listeners, enhancing the likelihood that target attributes are correctly interpreted by listeners amidst distractors. Additionally, we introduce a self-adjustable rationality parameter, which allows for automatic adjustment of control strength based on context. Our experiments, conducted with two task types and two types of language models, demonstrate that RSA-Control achieves strong attribute control while maintaining language fluency and content consistency. Our code is available at https://github.com/Ewanwong/RSA-Control.
Authors: Abhirama Subramanyam Penamakuri, Anand Mishra
Abstract: We revisit knowledge-aware text-based visual question answering, also known as Text-KVQA, in the light of modern advancements in large multimodal models (LMMs), and make the following contributions: (i) We propose VisTEL - a principled approach to perform visual text entity linking. The proposed VisTEL module harnesses a state-of-the-art visual text recognition engine and the power of a large multimodal model to jointly reason using textual and visual context obtained using surrounding cues in the image to link the visual text entity to the correct knowledge base entity. (ii) We present KaLMA - a knowledge-aware large multimodal assistant that augments an LMM with knowledge associated with visual text entity in the image to arrive at an accurate answer. Further, we provide a comprehensive experimental analysis and comparison of our approach with traditional visual question answering, pre-large multimodal models, and large multimodal models, as well as prior top-performing approaches. Averaging over three splits of Text-KVQA, our proposed approach surpasses the previous best approach by a substantial 23.3% on an absolute scale and establishes a new state of the art. We make our implementation publicly available.
Authors: Abraham Israeli, David Jurgens, Daniel Romero
Abstract: The Internet has significantly expanded the potential for global collaboration, allowing millions of users to contribute to collective projects like Wikipedia. While prior work has assessed the success of online collaborations, most approaches are time-agnostic, evaluating success without considering its longevity. Research on the factors that ensure the long-term preservation of high-quality standards in online collaboration is scarce. In this study, we address this gap. We propose a novel metric, `Sustainable Success,' which measures the ability of collaborative efforts to maintain their quality over time. Using Wikipedia as a case study, we introduce the SustainPedia dataset, which compiles data from over 40K Wikipedia articles, including each article's sustainable success label and more than 300 explanatory features such as edit history, user experience, and team composition. Using this dataset, we develop machine learning models to predict the sustainable success of Wikipedia articles. Our best-performing model achieves a high AU-ROC score of 0.88 on average. Our analysis reveals important insights. For example, we find that the longer an article takes to be recognized as high-quality, the more likely it is to maintain that status over time (i.e., be sustainable). Additionally, user experience emerged as the most critical predictor of sustainability. Our analysis provides insights into broader collective actions beyond Wikipedia (e.g., online activism, crowdsourced open-source software), where the same social dynamics that drive success on Wikipedia might play a role. We make all data and code used for this study publicly available for further research.
Authors: Samuel Jacob Chacko, Sajib Biswas, Chashi Mahiul Islam, Fatema Tabassum Liza, Xiuwen Liu
Abstract: As powerful Large Language Models (LLMs) are now widely used for numerous practical applications, their safety is of critical importance. While alignment techniques have significantly improved overall safety, LLMs remain vulnerable to carefully crafted adversarial inputs. Consequently, adversarial attack methods are extensively used to study and understand these vulnerabilities. However, current attack methods face significant limitations. Those relying on optimizing discrete tokens suffer from limited efficiency, while continuous optimization techniques fail to generate valid tokens from the model's vocabulary, rendering them impractical for real-world applications. In this paper, we propose a novel technique for adversarial attacks that overcomes these limitations by leveraging regularized gradients with continuous optimization methods. Our approach is two orders of magnitude faster than the state-of-the-art greedy coordinate gradient-based method, significantly improving the attack success rate on aligned language models. Moreover, it generates valid tokens, addressing a fundamental limitation of existing continuous optimization methods. We demonstrate the effectiveness of our attack on five state-of-the-art LLMs using four datasets.
Authors: S Sakshi, Utkarsh Tyagi, Sonal Kumar, Ashish Seth, Ramaneswaran Selvakumar, Oriol Nieto, Ramani Duraiswami, Sreyan Ghosh, Dinesh Manocha
Abstract: The ability to comprehend audio--which includes speech, non-speech sounds, and music--is crucial for AI agents to interact effectively with the world. We present MMAU, a novel benchmark designed to evaluate multimodal audio understanding models on tasks requiring expert-level knowledge and complex reasoning. MMAU comprises 10k carefully curated audio clips paired with human-annotated natural language questions and answers spanning speech, environmental sounds, and music. It includes information extraction and reasoning questions, requiring models to demonstrate 27 distinct skills across unique and challenging tasks. Unlike existing benchmarks, MMAU emphasizes advanced perception and reasoning with domain-specific knowledge, challenging models to tackle tasks akin to those faced by experts. We assess 18 open-source and proprietary (Large) Audio-Language Models, demonstrating the significant challenges posed by MMAU. Notably, even the most advanced Gemini Pro v1.5 achieves only 52.97% accuracy, and the state-of-the-art open-source Qwen2-Audio achieves only 52.50%, highlighting considerable room for improvement. We believe MMAU will drive the audio and multimodal research community to develop more advanced audio understanding models capable of solving complex audio tasks.
Authors: Maren Eckhoff, Valmir Selimi, Alexander Aranovitch, Ian Lyons, Emily Briggs, Jennifer Hou, Alex Devereson, Matej Macak, David Champagne, Chris Anagnostopoulos
Abstract: Many therapies are effective in treating multiple diseases. We present an approach that leverages methods developed in natural language processing and real-world data to prioritize potential, new indications for a mechanism of action (MoA). We specifically use representation learning to generate embeddings of indications and prioritize them based on their proximity to the indications with the strongest available evidence for the MoA. We demonstrate the successful deployment of our approach for anti-IL-17A using embeddings generated with SPPMI and present an evaluation framework to determine the quality of indication finding results and the derived embeddings.
Authors: Suparna De, Ionut Bostan, Nishanth Sastry
Abstract: Recent studies have outlined the accessibility challenges faced by blind or visually impaired, and less-literate people, in interacting with social networks, in-spite of facilitating technologies such as monotone text-to-speech (TTS) screen readers and audio narration of visual elements such as emojis. Emotional speech generation traditionally relies on human input of the expected emotion together with the text to synthesise, with additional challenges around data simplification (causing information loss) and duration inaccuracy, leading to lack of expressive emotional rendering. In real-life communications, the duration of phonemes can vary since the same sentence might be spoken in a variety of ways depending on the speakers' emotional states or accents (referred to as the one-to-many problem of text to speech generation). As a result, an advanced voice synthesis system is required to account for this unpredictability. We propose an end-to-end context-aware Text-to-Speech (TTS) synthesis system that derives the conveyed emotion from text input and synthesises audio that focuses on emotions and speaker features for natural and expressive speech, integrating advanced natural language processing (NLP) and speech synthesis techniques for real-time applications. Our system also showcases competitive inference time performance when benchmarked against the state-of-the-art TTS models, making it suitable for real-time accessibility applications.
Authors: Sadat Shahriar, Zheng Qi, Nikolaos Pappas, Srikanth Doss, Monica Sunkara, Kishaloy Halder, Manuel Mager, Yassine Benajiba
Abstract: Aligning Large Language Models (LLM) to address subjectivity and nuanced preference levels requires adequate flexibility and control, which can be a resource-intensive and time-consuming procedure. Existing training-time alignment methods require full re-training when a change is needed and inference-time ones typically require access to the reward model at each inference step. To address these limitations, we introduce inference-time model alignment method that learns encoded representations of preference dimensions, called \textit{Alignment Vectors} (AV). These representations are computed by subtraction of the base model from the aligned model as in model editing enabling dynamically adjusting the model behavior during inference through simple linear operations. Even though the preference dimensions can span various granularity levels, here we focus on three gradual response levels across three specialized domains: medical, legal, and financial, exemplifying its practical potential. This new alignment paradigm introduces adjustable preference knobs during inference, allowing users to tailor their LLM outputs while reducing the inference cost by half compared to the prompt engineering approach. Additionally, we find that AVs are transferable across different fine-tuning stages of the same model, demonstrating their flexibility. AVs also facilitate multidomain, diverse preference alignment, making the process 12x faster than the retraining approach.
Authors: Tianchun Wang, Yuanzhou Chen, Zichuan Liu, Zhanwen Chen, Haifeng Chen, Xiang Zhang, Wei Cheng
Abstract: The advent of large language models (LLMs) has revolutionized the field of text generation, producing outputs that closely mimic human-like writing. Although academic and industrial institutions have developed detectors to prevent the malicious usage of LLM-generated texts, other research has doubt about the robustness of these systems. To stress test these detectors, we introduce a proxy-attack strategy that effortlessly compromises LLMs, causing them to produce outputs that align with human-written text and mislead detection systems. Our method attacks the source model by leveraging a reinforcement learning (RL) fine-tuned humanized small language model (SLM) in the decoding phase. Through an in-depth analysis, we demonstrate that our attack strategy is capable of generating responses that are indistinguishable to detectors, preventing them from differentiating between machine-generated and human-written text. We conduct systematic evaluations on extensive datasets using proxy-attacked open-source models, including Llama2-13B, Llama3-70B, and Mixtral-8*7B in both white- and black-box settings. Our findings show that the proxy-attack strategy effectively deceives the leading detectors, resulting in an average AUROC drop of 70.4% across multiple datasets, with a maximum drop of 90.3% on a single dataset. Furthermore, in cross-discipline scenarios, our strategy also bypasses these detectors, leading to a significant relative decrease of up to 90.9%, while in cross-language scenario, the drop reaches 91.3%. Despite our proxy-attack strategy successfully bypassing the detectors with such significant relative drops, we find that the generation quality of the attacked models remains preserved, even within a modest utility budget, when compared to the text produced by the original, unattacked source model.
Authors: Leander Girrbach, Yiran Huang, Stephan Alaniz, Trevor Darrell, Zeynep Akata
Abstract: Pre-trained large language models (LLMs) have been reliably integrated with visual input for multimodal tasks. The widespread adoption of instruction-tuned image-to-text vision-language assistants (VLAs) like LLaVA and InternVL necessitates evaluating gender biases. We study gender bias in 22 popular open-source VLAs with respect to personality traits, skills, and occupations. Our results show that VLAs replicate human biases likely present in the data, such as real-world occupational imbalances. Similarly, they tend to attribute more skills and positive personality traits to women than to men, and we see a consistent tendency to associate negative personality traits with men. To eliminate the gender bias in these models, we find that finetuning-based debiasing methods achieve the best tradeoff between debiasing and retaining performance on downstream tasks. We argue for pre-deploying gender bias assessment in VLAs and motivate further development of debiasing strategies to ensure equitable societal outcomes.
Authors: Leonardo Ara\'ujo
Abstract: The vector notation adopted by GNU Octave plays a significant role as a tool for introspection, aligning itself with the vision of Kenneth E. Iverson. He believed that, just like mathematics, a programming language should be an effective thinking tool for representing and reasoning about problems we wish to address. This work aims to explore the use of vector notation in GNU Octave through the analysis of operators and functions, providing a closer alignment with mathematical notation and enhancing code efficiency. We will delve into fundamental concepts such as indexing, broadcasting, and function handles, and present case studies for a deeper understanding of these concepts. By adopting vector notation, GNU Octave becomes a powerful tool for mathematicians, scientists and engineers, enabling them to express and solve complex problems more effectively and intuitively.
Authors: Kaixian Qu, Jie Tan, Tingnan Zhang, Fei Xia, Cesar Cadena, Marco Hutter
Abstract: Navigating efficiently to an object in an unexplored environment is a critical skill for general-purpose intelligent robots. Recent approaches to this object goal navigation problem have embraced a modular strategy, integrating classical exploration algorithms-notably frontier exploration-with a learned semantic mapping/exploration module. This paper introduces a novel informative path planning and 3D object probability mapping approach. The mapping module computes the probability of the object of interest through semantic segmentation and a Bayes filter. Additionally, it stores probabilities for common objects, which semantically guides the exploration based on common sense priors from a large language model. The planner terminates when the current viewpoint captures enough voxels identified with high confidence as the object of interest. Although our planner follows a zero-shot approach, it achieves state-of-the-art performance as measured by the Success weighted by Path Length (SPL) and Soft SPL in the Habitat ObjectNav Challenge 2023, outperforming other works by more than 20%. Furthermore, we validate its effectiveness on real robots. Project webpage: https://ippon-paper.github.io/
Authors: Nicole Cho, Nishan Srishankar, Lucas Cecchi, William Watson
Abstract: Financial intelligence generation from vast data sources has typically relied on traditional methods of knowledge-graph construction or database engineering. Recently, fine-tuned financial domain-specific Large Language Models (LLMs), have emerged. While these advancements are promising, limitations such as high inference costs, hallucinations, and the complexity of concurrently analyzing high-dimensional financial data, emerge. This motivates our invention FISHNET (Financial Intelligence from Sub-querying, Harmonizing, Neural-Conditioning, Expert swarming, and Task planning), an agentic architecture that accomplishes highly complex analytical tasks for more than 98,000 regulatory filings that vary immensely in terms of semantics, data hierarchy, or format. FISHNET shows remarkable performance for financial insight generation (61.8% success rate over 5.0% Routing, 45.6% RAG R-Precision). We conduct rigorous ablations to empirically prove the success of FISHNET, each agent's importance, and the optimized performance of assembling all agents. Our modular architecture can be leveraged for a myriad of use-cases, enabling scalability, flexibility, and data integrity that are critical for financial tasks.
Authors: Sizhe Zhou, Siru Ouyang, Zhuosheng Zhang, Hai Zhao
Abstract: In open-retrieval conversational machine reading (OR-CMR) task, machines are required to do multi-turn question answering given dialogue history and a textual knowledge base. Existing works generally utilize two independent modules to approach this problem's two successive sub-tasks: first with a hard-label decision making and second with a question generation aided by various entailment reasoning methods. Such usual cascaded modeling is vulnerable to error propagation and prevents the two sub-tasks from being consistently optimized. In this work, we instead model OR-CMR as a unified text-to-text task in a fully end-to-end style. Experiments on the ShARC and OR-ShARC dataset show the effectiveness of our proposed end-to-end framework on both sub-tasks by a large margin, achieving new state-of-the-art results. Further ablation studies support that our framework can generalize to different backbone models.
Authors: Leilei Su, Jian Chen, Yifan Peng, Cong Sun
Abstract: Although deep learning techniques have shown significant achievements, they frequently depend on extensive amounts of hand-labeled data and tend to perform inadequately in few-shot scenarios. The objective of this study is to devise a strategy that can improve the model's capability to recognize biomedical entities in scenarios of few-shot learning. By redefining biomedical named entity recognition (BioNER) as a machine reading comprehension (MRC) problem, we propose a demonstration-based learning method to address few-shot BioNER, which involves constructing appropriate task demonstrations. In assessing our proposed method, we compared the proposed method with existing advanced methods using six benchmark datasets, including BC4CHEMD, BC5CDR-Chemical, BC5CDR-Disease, NCBI-Disease, BC2GM, and JNLPBA. We examined the models' efficacy by reporting F1 scores from both the 25-shot and 50-shot learning experiments. In 25-shot learning, we observed 1.1% improvements in the average F1 scores compared to the baseline method, reaching 61.7%, 84.1%, 69.1%, 70.1%, 50.6%, and 59.9% on six datasets, respectively. In 50-shot learning, we further improved the average F1 scores by 1.0% compared to the baseline method, reaching 73.1%, 86.8%, 76.1%, 75.6%, 61.7%, and 65.4%, respectively. We reported that in the realm of few-shot learning BioNER, MRC-based language models are much more proficient in recognizing biomedical entities compared to the sequence labeling approach. Furthermore, our MRC-language models can compete successfully with fully-supervised learning methodologies that rely heavily on the availability of abundant annotated data. These results highlight possible pathways for future advancements in few-shot BioNER methodologies.
Authors: Dihia Lanasri, Juan Olano, Sifal Klioui, Sin Liang Lee, Lamia Sekkai
Abstract: With the proliferation of hate speech on social networks under different formats, such as abusive language, cyberbullying, and violence, etc., people have experienced a significant increase in violence, putting them in uncomfortable situations and threats. Plenty of efforts have been dedicated in the last few years to overcome this phenomenon to detect hate speech in different structured languages like English, French, Arabic, and others. However, a reduced number of works deal with Arabic dialects like Tunisian, Egyptian, and Gulf, mainly the Algerian ones. To fill in the gap, we propose in this work a complete approach for detecting hate speech on online Algerian messages. Many deep learning architectures have been evaluated on the corpus we created from some Algerian social networks (Facebook, YouTube, and Twitter). This corpus contains more than 13.5K documents in Algerian dialect written in Arabic, labeled as hateful or non-hateful. Promising results are obtained, which show the efficiency of our approach.
Authors: Yijia Xiao, Yiqiao Jin, Yushi Bai, Yue Wu, Xianjun Yang, Xiao Luo, Wenchao Yu, Xujiang Zhao, Yanchi Liu, Quanquan Gu, Haifeng Chen, Wei Wang, Wei Cheng
Abstract: The proliferation of Large Language Models (LLMs) has driven considerable interest in fine-tuning them with domain-specific data to create specialized language models. Nevertheless, such domain-specific fine-tuning data often contains contextually sensitive personally identifiable information (PII). Direct fine-tuning of LLMs on this data without privacy protection poses a risk of data leakage of sensitive PII during inference time. To address this challenge, we introduce Contextual Privacy Protection Language Models (CPPLM), a novel paradigm for fine-tuning LLMs that effectively injects domain-specific knowledge while safeguarding inference-time data privacy. Our work offers a theoretical analysis for model design and benchmarks various techniques such as corpus curation, penalty-based unlikelihood in training loss, instruction-based tuning, etc. Extensive experiments across diverse datasets and scenarios demonstrate the effectiveness of our approaches. In particular, instruction tuning with both positive and negative examples stands out as a promising method, effectively protecting private data while enhancing the model's knowledge. Our work underscores the potential for Large Language Models as robust contextual privacy protection learners. The complete code and data for the work can be found at https://github.com/Yijia-Xiao/PPLM.
Authors: George Chrysostomou, Zhixue Zhao, Miles Williams, Nikolaos Aletras
Abstract: Despite the remarkable performance of generative large language models (LLMs) on abstractive summarization, they face two significant challenges: their considerable size and tendency to hallucinate. Hallucinations are concerning because they erode reliability and raise safety issues. Pruning is a technique that reduces model size by removing redundant weights, enabling more efficient sparse inference. Pruned models yield downstream task performance comparable to the original, making them ideal alternatives when operating on a limited budget. However, the effect that pruning has upon hallucinations in abstractive summarization with LLMs has yet to be explored. In this paper, we provide an extensive empirical study across five summarization datasets, two state-of-the-art pruning methods, and five instruction-tuned LLMs. Surprisingly, we find that hallucinations are less prevalent from pruned LLMs than the original models. Our analysis suggests that pruned models tend to depend more on the source document for summary generation. This leads to a higher lexical overlap between the generated summary and the source document, which could be a reason for the reduction in hallucination risk.
Authors: Wenda Xu, Daniel Deutsch, Mara Finkelstein, Juraj Juraska, Biao Zhang, Zhongtao Liu, William Yang Wang, Lei Li, Markus Freitag
Abstract: Recent large language models (LLM) are leveraging human feedback to improve their generation quality. However, human feedback is costly to obtain, especially during inference. In this work, we propose LLMRefine, an inference time optimization method to refine LLM's output. The core idea is to use a learned fine-grained feedback model to pinpoint defects and guide LLM to refine them iteratively. Using original LLM as a proposal of edits, LLMRefine searches for defect-less text via simulated annealing, trading off the exploration and exploitation. We conduct experiments on three text generation tasks, including machine translation, long-form question answering (QA), and topical summarization. LLMRefine consistently outperforms all baseline approaches, achieving improvements up to 1.7 MetricX points on translation tasks, 8.1 ROUGE-L on ASQA, 2.2 ROUGE-L on topical summarization.
Authors: Zonghai Yao, Nandyala Siddharth Kantu, Guanghao Wei, Hieu Tran, Zhangqi Duan, Sunjae Kwon, Zhichao Yang, README annotation team, Hong Yu
Abstract: The advancement in healthcare has shifted focus toward patient-centric approaches, particularly in self-care and patient education, facilitated by access to Electronic Health Records (EHR). However, medical jargon in EHRs poses significant challenges in patient comprehension. To address this, we introduce a new task of automatically generating lay definitions, aiming to simplify complex medical terms into patient-friendly lay language. We first created the README dataset, an extensive collection of over 50,000 unique (medical term, lay definition) pairs and 300,000 mentions, each offering context-aware lay definitions manually annotated by domain experts. We have also engineered a data-centric Human-AI pipeline that synergizes data filtering, augmentation, and selection to improve data quality. We then used README as the training data for models and leveraged a Retrieval-Augmented Generation method to reduce hallucinations and improve the quality of model outputs. Our extensive automatic and human evaluations demonstrate that open-source mobile-friendly models, when fine-tuned with high-quality data, are capable of matching or even surpassing the performance of state-of-the-art closed-source large language models like ChatGPT. This research represents a significant stride in closing the knowledge gap in patient education and advancing patient-centric healthcare solutions.
Authors: Aliz\'ee Pace, Jonathan Mallinson, Eric Malmi, Sebastian Krause, Aliaksei Severyn
Abstract: The success of reinforcement learning from human feedback (RLHF) in language model alignment is strongly dependent on the quality of the underlying reward model. In this paper, we present a novel approach to improve reward model quality by generating synthetic preference data, thereby augmenting the training dataset with on-policy, high-quality preference pairs. Motivated by the promising results of Best-of-N sampling strategies in language model training, we extend their application to reward model training. This results in a self-training strategy to generate preference pairs by selecting the best and worst candidates in a pool of responses to a given query. Empirically, we find that this approach improves the performance of any reward model, with an effect comparable to the addition of a similar quantity of human preference data. This work opens up new avenues of research for improving RLHF for language model alignment, by offering synthetic preference generation as a solution to reward modeling challenges.
Authors: Jian Guan, Wei Wu, Zujie Wen, Peng Xu, Hongning Wang, Minlie Huang
Abstract: The notable success of large language models (LLMs) has sparked an upsurge in building language agents to complete various complex tasks. We present AMOR, an agent framework based on open-source LLMs, which reasons with external knowledge bases and adapts to specific domains through human supervision to the reasoning process. AMOR builds reasoning logic over a finite state machine (FSM) that solves problems through autonomous executions and transitions over disentangled modules. This allows humans to provide direct feedback to the individual modules, and thus naturally forms process supervision. Based on this reasoning and feedback framework, we develop AMOR through two-stage fine-tuning: warm-up and adaptation. The former fine-tunes the LLM with examples automatically constructed from various public datasets, enabling AMOR to generalize across different knowledge environments, while the latter tailors AMOR to specific domains using process feedback. Extensive experiments across multiple domains demonstrate the advantage of AMOR to strong baselines, thanks to its FSM-based reasoning and process feedback mechanism. The code and data are publicly available at \url{https://github.com/JianGuanTHU/AMOR}.
Authors: Zhaoxuan Tan, Qingkai Zeng, Yijun Tian, Zheyuan Liu, Bing Yin, Meng Jiang
Abstract: Personalization in large language models (LLMs) is increasingly important, aiming to align the LLMs' interactions, content, and recommendations with individual user preferences. Recent advances have highlighted effective prompt design by enriching user queries with non-parametric knowledge through behavior history retrieval and textual profiles. However, these methods faced limitations due to a lack of model ownership, resulting in constrained customization and privacy issues, and often failed to capture complex, dynamic user behavior patterns. To address these shortcomings, we introduce One PEFT Per User (OPPU), employing personalized parameter-efficient fine-tuning (PEFT) modules to store user-specific behavior patterns and preferences. By plugging in personal PEFT parameters, users can own and use their LLMs individually. OPPU integrates parametric user knowledge in the personal PEFT parameters with non-parametric knowledge from retrieval and profiles, adapting LLMs to user behavior shifts. Experimental results demonstrate that OPPU significantly outperforms existing prompt-based methods across seven diverse tasks in the LaMP benchmark. Further studies reveal OPPU's enhanced capabilities in handling user behavior shifts, modeling users at different activity levels, maintaining robustness across various user history formats, and displaying versatility with different PEFT methods.
Authors: Xinbei Ma, Tianjie Ju, Jiyang Qiu, Zhuosheng Zhang, Hai Zhao, Lifeng Liu, Yulong Wang
Abstract: Large language models (LLMs) have played a pivotal role in building communicative AI, yet they encounter the challenge of efficient updates. Model editing enables the manipulation of specific knowledge memories and the behavior of language generation without retraining. However, the robustness of model editing remains an open question. This work seeks to understand the strengths and limitations of editing methods, facilitating practical applications of communicative AI. We focus on three key research questions. RQ1: Can edited LLMs behave consistently resembling communicative AI in realistic situations? RQ2: To what extent does the rephrasing of prompts lead LLMs to deviate from the edited knowledge memory? RQ3: Which knowledge features are correlated with the performance and robustness of editing? Our empirical studies uncover a substantial disparity between existing editing methods and the practical application of LLMs. On rephrased prompts that are flexible but common in realistic applications, the performance of editing experiences a significant decline. Further analysis shows that more popular knowledge is memorized better, easier to recall, and more challenging to edit effectively. Code is publicly available at https://github.com/xbmxb/edit_analysis .
Authors: Sizhe Zhou, Yu Meng, Bowen Jin, Jiawei Han
Abstract: Relation extraction (RE) aims to identify semantic relationships between entities within text. Despite considerable advancements, existing models predominantly require extensive annotated training data, which is both costly and labor-intensive to collect. Moreover, these models often struggle to adapt to new or unseen relations. Few-shot learning, aiming to lessen annotation demands, typically provides incomplete and biased supervision for target relations, leading to degraded and unstable performance. To accurately and explicitly describe relation semantics while minimizing annotation demands, we explore the definition only zero-shot RE setting where only relation definitions expressed in natural language are used to train a RE model. We introduce REPaL, comprising three stages: (1) We leverage large language models (LLMs) to generate initial seed instances from relation definitions and an unlabeled corpus. (2) We fine-tune a bidirectional Small Language Model (SLM) with initial seeds to learn relations for the target domain. (3) We expand pattern coverage and mitigate bias from initial seeds by integrating feedback from the SLM's predictions on the unlabeled corpus and the synthesis history. To accomplish this, we leverage the multi-turn conversation ability of LLMs to generate new instances in follow-up dialogues, informed by both the feedback and synthesis history. Studies reveal that definition-oriented seed synthesis enhances pattern coverage whereas indiscriminately increasing seed quantity leads to performance saturation. Experiments on two datasets show REPaL significantly improved cost-effective zero-shot performance by large margins.
Authors: Shubham Vatsal, Ayush Singh, Shabnam Tafreshi
Abstract: Health insurance companies have a defined process called prior authorization (PA) which is a health plan cost-control process that requires doctors and other healthcare professionals to get clearance in advance from a health plan before performing a particular procedure on a patient in order to be eligible for payment coverage. For health insurance companies, approving PA requests for patients in the medical domain is a time-consuming and challenging task. One of those key challenges is validating if a request matches up to certain criteria such as age, gender, etc. In this work, we evaluate whether GPT can validate numerous key factors, in turn helping health plans reach a decision drastically faster. We frame it as a question answering task, prompting GPT to answer a question from patient electronic health record. We experiment with different conventional prompting techniques as well as introduce our own novel prompting technique. Moreover, we report qualitative assessment by humans on the natural language generation outputs from our approach. Results show that our method achieves superior performance with the mean weighted F1 score of 0.61 as compared to its standard counterparts.
Authors: Lang Cao, Jimeng Sun, Adam Cross
Abstract: Rare diseases affect millions worldwide but often face limited research focus due to their low prevalence. This results in prolonged diagnoses and a lack of approved therapies. Recent advancements in Large Language Models (LLMs) have shown promise in automating the extraction of medical information, offering potential to improve medical diagnosis and management. However, most LLMs lack professional medical knowledge, especially concerning rare diseases, and struggle to handle the latest rare disease information. They also cannot effectively manage rare disease data and are not directly suitable for diagnosis and management tasks. Our objective is to create an end-to-end system called AutoRD, which automates the extraction of information from medical texts about rare diseases, focusing on entities and their relations. AutoRD integrates up-to-date structured knowledge and demonstrates superior performance in rare disease extraction tasks. We conduct various experiments to evaluate AutoRD's performance, aiming to surpass common LLMs and traditional methods.
Authors: Yingshan Chang, Yasi Zhang, Zhiyuan Fang, Yingnian Wu, Yonatan Bisk, Feng Gao
Abstract: The literature on text-to-image generation is plagued by issues of faithfully composing entities with relations. But there lacks a formal understanding of how entity-relation compositions can be effectively learned. Moreover, the underlying phenomenon space that meaningfully reflects the problem structure is not well-defined, leading to an arms race for larger quantities of data in the hope that generalization emerges out of large-scale pretraining. We hypothesize that the underlying phenomenological coverage has not been proportionally scaled up, leading to a skew of the presented phenomenon which harms generalization. We introduce statistical metrics that quantify both the linguistic and visual skew of a dataset for relational learning, and show that generalization failures of text-to-image generation are a direct result of incomplete or unbalanced phenomenological coverage. We first perform experiments in a synthetic domain and demonstrate that systematically controlled metrics are strongly predictive of generalization performance. Then we move to natural images and show that simple distribution perturbations in light of our theories boost generalization without enlarging the absolute data size. This work informs an important direction towards quality-enhancing the data diversity or balance orthogonal to scaling up the absolute size. Our discussions point out important open questions on 1) Evaluation of generated entity-relation compositions, and 2) Better models for reasoning with abstract relations.
Authors: Xiaocheng Luo, Yanping Chen, Ruixue Tang, Caiwei Yang, Ruizhang Huang, Yongbin Qin
Abstract: Current methods to extract relational triples directly make a prediction based on a possible entity pair in a raw sentence without depending on entity recognition. The task suffers from a serious semantic overlapping problem, in which several relation triples may share one or two entities in a sentence. In this paper, based on a two-dimensional sentence representation, a bi-consolidating model is proposed to address this problem by simultaneously reinforcing the local and global semantic features relevant to a relation triple. This model consists of a local consolidation component and a global consolidation component. The first component uses a pixel difference convolution to enhance semantic information of a possible triple representation from adjacent regions and mitigate noise in neighbouring neighbours. The second component strengthens the triple representation based a channel attention and a spatial attention, which has the advantage to learn remote semantic dependencies in a sentence. They are helpful to improve the performance of both entity identification and relation type classification in relation triple extraction. After evaluated on several publish datasets, the bi-consolidating model achieves competitive performance. Analytical experiments demonstrate the effectiveness of our model for relational triple extraction and give motivation for other natural language processing tasks.
Authors: Jonghyun Song, Cheyon Jin, Wenlong Zhao, Andrew McCallum, Jay-Yoon Lee
Abstract: A common retrieve-and-rerank paradigm involves retrieving relevant candidates from a broad set using a fast bi-encoder (BE), followed by applying expensive but accurate cross-encoders (CE) to a limited candidate set. However, relying on this small subset is often susceptible to error propagation from the bi-encoders, which limits the overall performance. To address these issues, we propose the Comparing Multiple Candidates (CMC) framework. CMC compares a query and multiple embeddings of similar candidates (i.e., neighbors) through shallow self-attention layers, delivering rich representations contextualized to each other. Furthermore, CMC is scalable enough to handle multiple comparisons simultaneously. For example, comparing ~10K candidates with CMC takes a similar amount of time as comparing 16 candidates with CE. Experimental results on the ZeSHEL dataset demonstrate that CMC, when plugged in between bi-encoders and cross-encoders as a seamless intermediate reranker (BE-CMC-CE), can effectively improve recall@k (+4.8%-p, +3.5%-p for R@16, R@64) compared to using only bi-encoders (BE-CE), with negligible slowdown (<7%). Additionally, to verify CMC's effectiveness as the final-stage reranker in improving top-1 accuracy, we conduct experiments on downstream tasks such as entity, passage, and dialogue ranking. The results indicate that CMC is not only faster (11x) but also often more effective than CE, with improved prediction accuracy in Wikipedia entity linking (+0.7%-p) and DSTC7 dialogue ranking (+3.3%-p).
Authors: Shubham Vatsal, Ayush Singh
Abstract: Large language models (LLMs) have shown remarkable performance on many tasks in different domains. However, their performance in closed-book biomedical machine reading comprehension (MRC) has not been evaluated in depth. In this work, we evaluate GPT on four closed-book biomedical MRC benchmarks. We experiment with different conventional prompting techniques as well as introduce our own novel prompting method. To solve some of the retrieval problems inherent to LLMs, we propose a prompting strategy named Implicit Retrieval Augmented Generation (RAG) that alleviates the need for using vector databases to retrieve important chunks in traditional RAG setups. Moreover, we report qualitative assessments on the natural language generation outputs from our approach. The results show that our new prompting technique is able to get the best performance in two out of four datasets and ranks second in rest of them. Experiments show that modern-day LLMs like GPT even in a zero-shot setting can outperform supervised models, leading to new state-of-the-art (SoTA) results on two of the benchmarks.
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: Cheng-Hsun Hsueh, Paul Kuo-Ming Huang, Tzu-Han Lin, Che-Wei Liao, Hung-Chieh Fang, Chao-Wei Huang, Yun-Nung Chen
Abstract: Knowledge editing is a rising technique for efficiently updating factual knowledge in large language models (LLMs) with minimal alteration of parameters. However, recent studies have identified side effects, such as knowledge distortion and the deterioration of general abilities, that have emerged after editing. Despite these findings, evaluating the pitfalls of knowledge editing often relies on inconsistent metrics and benchmarks, lacking a uniform standard. In response, this survey presents a comprehensive study of these side effects, providing a unified perspective on the challenges of knowledge editing in LLMs by conducting experiments with consistent metrics and benchmarks. Additionally, we review related works and outline potential research directions to address these limitations. Our survey highlights the limitations of current knowledge editing methods, emphasizing the need for a deeper understanding of the inner knowledge structures of LLMs and improved knowledge editing methods. To foster future research, we have released the complementary materials publicly in https://github.com/MiuLab/EditLLM-Survey.
Authors: Federico Barbero, Andrea Banino, Steven Kapturowski, Dharshan Kumaran, Jo\~ao G. M. Ara\'ujo, Alex Vitvitskyi, Razvan Pascanu, Petar Veli\v{c}kovi\'c
Abstract: We study how information propagates in decoder-only Transformers, which are the architectural backbone of most existing frontier large language models (LLMs). We rely on a theoretical signal propagation analysis -- specifically, we analyse the representations of the last token in the final layer of the Transformer, as this is the representation used for next-token prediction. Our analysis reveals a representational collapse phenomenon: we prove that certain distinct sequences of inputs to the Transformer can yield arbitrarily close representations in the final token. This effect is exacerbated by the low-precision floating-point formats frequently used in modern LLMs. As a result, the model is provably unable to respond to these sequences in different ways -- leading to errors in, e.g., tasks involving counting or copying. Further, we show that decoder-only Transformer language models can lose sensitivity to specific tokens in the input, which relates to the well-known phenomenon of over-squashing in graph neural networks. We provide empirical evidence supporting our claims on contemporary LLMs. Our theory also points to simple solutions towards ameliorating these issues.
Authors: Zhaoxuan Tan, Zheyuan Liu, Meng Jiang
Abstract: Personalized large language models (LLMs) aim to tailor interactions, content, and recommendations to individual user preferences. While parameter-efficient fine-tuning (PEFT) methods excel in performance and generalization, they are costly and limit communal benefits when used individually. To this end, we introduce Personalized Pieces (Per-Pcs), a framework that allows users to safely share and assemble personalized PEFT efficiently with collaborative efforts. Per-Pcs involves selecting sharers, breaking their PEFT into pieces, and training gates for each piece. These pieces are added to a pool, from which target users can select and assemble personalized PEFT using their history data. This approach preserves privacy and enables fine-grained user modeling without excessive storage and computation demands. Experimental results show Per-Pcs outperforms non-personalized and PEFT retrieval baselines, offering performance comparable to OPPU with significantly lower resource use across six tasks. Further analysis highlights Per-Pcs's robustness concerning sharer count and selection strategy, pieces sharing ratio, and scalability in computation time and storage space. Per-Pcs's modularity promotes safe sharing, making LLM personalization more efficient, effective, and widely accessible through collaborative efforts.
Authors: Sara Court, Micha Elsner
Abstract: This work investigates the in-context learning abilities of pretrained large language models (LLMs) when instructed to translate text from a low-resource language into a high-resource language as part of an automated machine translation pipeline. We conduct a set of experiments translating Southern Quechua to Spanish and examine the informativity of various types of context retrieved from a constrained database of digitized pedagogical materials (dictionaries and grammar lessons) and parallel corpora. Using both automatic and human evaluation of model output, we conduct ablation studies that manipulate (1) context type (morpheme translations, grammar descriptions, and corpus examples), (2) retrieval methods (automated vs. manual), and (3) model type. Our results suggest that even relatively small LLMs are capable of utilizing prompt context for zero-shot low-resource translation when provided a minimally sufficient amount of relevant linguistic information. However, the variable effects of context type, retrieval method, model type, and language-specific factors highlight the limitations of using even the best LLMs as translation systems for the majority of the world's 7,000+ languages and their speakers.
Authors: Sourish Dasgupta, Ankush Chander, Parth Borad, Isha Motiyani, Tanmoy Chakraborty
Abstract: Personalized summarization models cater to individuals' subjective understanding of saliency, as represented by their reading history and current topics of attention. Existing personalized text summarizers are primarily evaluated based on accuracy measures such as BLEU, ROUGE, and METEOR. However, a recent study argued that accuracy measures are inadequate for evaluating the degree of personalization of these models and proposed EGISES, the first metric to evaluate personalized text summaries. It was suggested that accuracy is a separate aspect and should be evaluated standalone. In this paper, we challenge the necessity of an accuracy leaderboard, suggesting that relying on accuracy-based aggregated results might lead to misleading conclusions. To support this, we delve deeper into EGISES, demonstrating both theoretically and empirically that it measures the degree of responsiveness, a necessary but not sufficient condition for degree-of-personalization. We subsequently propose PerSEval, a novel measure that satisfies the required sufficiency condition. Based on the benchmarking of ten SOTA summarization models on the PENS dataset, we empirically establish that -- (i) PerSEval is reliable w.r.t human-judgment correlation (Pearson's r = 0.73; Spearman's $\rho$ = 0.62; Kendall's $\tau$ = 0.42), (ii) PerSEval has high rank-stability, (iii) PerSEval as a rank-measure is not entailed by EGISES-based ranking, and (iv) PerSEval can be a standalone rank-measure without the need of any aggregated ranking.
Authors: Wataru Hashimoto, Hidetaka Kamigaito, Taro Watanabe
Abstract: This work investigates the impact of data augmentation on confidence calibration and uncertainty estimation in Named Entity Recognition (NER) tasks. For the future advance of NER in safety-critical fields like healthcare and finance, it is essential to achieve accurate predictions with calibrated confidence when applying Deep Neural Networks (DNNs), including Pre-trained Language Models (PLMs), as a real-world application. However, DNNs are prone to miscalibration, which limits their applicability. Moreover, existing methods for calibration and uncertainty estimation are computational expensive. Our investigation in NER found that data augmentation improves calibration and uncertainty in cross-genre and cross-lingual setting, especially in-domain setting. Furthermore, we showed that the calibration for NER tends to be more effective when the perplexity of the sentences generated by data augmentation is lower, and that increasing the size of the augmentation further improves calibration and uncertainty.
Authors: Aleksander Ficek, Jiaqi Zeng, Oleksii Kuchaiev
Abstract: Parameter-Efficient Fine-Tuning (PEFT) and Retrieval-Augmented Generation (RAG) have become popular methods for adapting large language models while minimizing compute requirements. In this paper, we apply PEFT methods (P-tuning, Adapters, and LoRA) to a modified Retrieval-Enhanced Transformer (RETRO) and a baseline GPT model across several sizes, ranging from 823 million to 48 billion parameters. We show that RETRO models outperform GPT models in zero-shot settings due to their unique pre-training process but GPT models have higher performance potential with PEFT. Additionally, our study indicates that 8B parameter models strike an optimal balance between cost and performance and P-tuning lags behind other PEFT techniques. We further provide a comparative analysis between applying PEFT to an Instruction-tuned RETRO model and base RETRO model. This work presents the first comprehensive comparison of various PEFT methods integrated with RAG, applied to both GPT and RETRO models, highlighting their relative performance.
Authors: Aditya K Surikuchi, Raquel Fern\'andez, Sandro Pezzelle
Abstract: Visual storytelling consists in generating a natural language story given a temporally ordered sequence of images. This task is not only challenging for models, but also very difficult to evaluate with automatic metrics since there is no consensus about what makes a story 'good'. In this paper, we introduce a novel method that measures story quality in terms of human likeness regarding three key aspects highlighted in previous work: visual grounding, coherence, and repetitiveness. We then use this method to evaluate the stories generated by several models, showing that the foundation model LLaVA obtains the best result, but only slightly so compared to TAPM, a 50-times smaller visual storytelling model. Upgrading the visual and language components of TAPM results in a model that yields competitive performance with a relatively low number of parameters. Finally, we carry out a human evaluation study, whose results suggest that a 'good' story may require more than a human-like level of visual grounding, coherence, and repetition.
Authors: Bo Wang, Tsunenori Mine
Abstract: This paper presents a novel and comprehensive solution to enhance both the robustness and efficiency of question answering (QA) systems through supervised contrastive learning (SCL). Training a high-performance QA system has become straightforward with pre-trained language models, requiring only a small amount of data and simple fine-tuning. However, despite recent advances, existing QA systems still exhibit significant deficiencies in functionality and training efficiency. We address the functionality issue by defining four key tasks: user input intent classification, out-of-domain input detection, new intent discovery, and continual learning. We then leverage a unified SCL-based representation learning method to efficiently build an intra-class compact and inter-class scattered feature space, facilitating both known intent classification and unknown intent detection and discovery. Consequently, with minimal additional tuning on downstream tasks, our approach significantly improves model efficiency and achieves new state-of-the-art performance across all tasks.
Authors: Lutfi Eren Erdogan, Nicholas Lee, Siddharth Jha, Sehoon Kim, Ryan Tabrizi, Suhong Moon, Coleman Hooper, Gopala Anumanchipalli, Kurt Keutzer, Amir Gholami
Abstract: Recent large language models (LLMs) have enabled the development of advanced agentic systems that can integrate various tools and APIs to fulfill user queries through function calling. However, the deployment of these LLMs on the edge has not been explored since they typically require cloud-based infrastructure due to their substantial model size and computational demands. To this end, we present TinyAgent, an end-to-end framework for training and deploying task-specific small language model agents capable of function calling for driving agentic systems at the edge. We first show how to enable accurate function calling for open-source models via the LLMCompiler framework. We then systematically curate a high-quality dataset for function calling, which we use to fine-tune two small language models, TinyAgent-1.1B and 7B. For efficient inference, we introduce a novel tool retrieval method to reduce the input prompt length and utilize quantization to further accelerate the inference speed. As a driving application, we demonstrate a local Siri-like system for Apple's MacBook that can execute user commands through text or voice input. Our results show that our models can achieve, and even surpass, the function-calling capabilities of larger models like GPT-4-Turbo, while being fully deployed at the edge. We open-source our dataset, models, and installable package and provide a demo video for our MacBook assistant agent.
Authors: Qin Dai, Benjamin Heinzerling, Kentaro Inui
Abstract: Entity tracking is essential for complex reasoning. To perform in-context entity tracking, language models (LMs) must bind an entity to its attribute (e.g., bind a container to its content) to recall attribute for a given entity. For example, given a context mentioning ``The coffee is in Box Z, the stone is in Box M, the map is in Box H'', to infer ``Box Z contains the coffee'' later, LMs must bind ``Box Z'' to ``coffee''. To explain the binding behaviour of LMs, existing research introduces a Binding ID mechanism and states that LMs use a abstract concept called Binding ID (BI) to internally mark entity-attribute pairs. However, they have not captured the Ordering ID (OI) from entity activations that directly determines the binding behaviour. In this work, we provide a novel view of the BI mechanism by localizing OI and proving the causality between OI and binding behaviour. Specifically, by leveraging dimension reduction methods (e.g., PCA), we discover that there exists a low-rank subspace in the activations of LMs, that primarily encodes the order (i.e., OI) of entity and attribute. Moreover, we also discover the causal effect of OI on binding that when editing representations along the OI encoding direction, LMs tend to bind a given entity to other attributes accordingly. For example, by patching activations along the OI encoding direction we can make the LM to infer ``Box Z contains the stone'' and ``Box Z contains the map''.
Authors: Chaithanya Bandi, Abir Harrasse
Abstract: This paper explores optimal architectures for evaluating the outputs of large language models (LLMs) using LLMs themselves. We propose a novel framework that interprets LLMs as advocates within an ensemble of interacting agents, allowing them to defend their answers and reach conclusions through a judge and jury system. This approach offers a more dynamic and comprehensive evaluation process compared to traditional human-based assessments or automated metrics. We discuss the motivation behind this framework, its key components, and comparative advantages. We also present a probabilistic model to evaluate the error reduction achieved by iterative advocate systems. Finally, we outline experiments to validate the effectiveness of multi-advocate architectures and discuss future research directions.
Authors: Xin Mao, Feng-Lin Li, Huimin Xu, Wei Zhang, Wang Chen, Anh Tuan Luu
Abstract: Direct Preference Optimization (DPO) has emerged as a more computationally efficient alternative to Reinforcement Learning from Human Feedback (RLHF) with Proximal Policy Optimization (PPO), eliminating the need for reward models and online sampling. Despite these benefits, DPO and its variants remain sensitive to hyper-parameters and prone to instability, particularly on mathematical datasets. We argue that these issues arise from the unidirectional likelihood-derivative negative feedback inherent in the log-likelihood loss function. To address this, we propose a novel LLM alignment loss that establishes a stable Bidirectional Negative Feedback (BNF) during optimization. Our proposed BNF loss eliminates the need for pairwise contrastive losses and does not require any extra tunable hyper-parameters or pairwise preference data, streamlining the alignment pipeline to be as simple as supervised fine-tuning. We conduct extensive experiments across two challenging QA benchmarks and four reasoning benchmarks. The experimental results show that BNF achieves comparable performance to the best methods on QA benchmarks, while its performance decrease on the four reasoning benchmarks is significantly lower compared to the best methods, thus striking a better balance between value alignment and reasoning ability. In addition, we further validate the performance of BNF on non-pairwise datasets, and conduct in-depth analysis of log-likelihood and logit shifts across different preference optimization methods.
Authors: Akira Kawabata, Saku Sugawara
Abstract: Answer verification identifies correct solutions among candidates generated by large language models (LLMs). Current approaches typically train verifier models by labeling solutions as correct or incorrect based solely on whether the final answer matches the gold answer. However, this approach neglects any flawed rationale in the solution yielding the correct answer, undermining the verifier's ability to distinguish between sound and flawed rationales. We empirically show that in StrategyQA, only 19% of LLM-generated solutions with correct answers have valid rationales, thus leading to an unreliable verifier. Furthermore, we demonstrate that training a verifier on valid rationales significantly improves its ability to distinguish valid and flawed rationale. To make a better verifier without extra human supervision, we introduce REPS (Rationale Enhancement through Pairwise Selection), a method for selecting valid rationales from candidates by iteratively applying pairwise self-evaluation using the same LLM that generates the solutions. Verifiers trained on solutions selected by REPS outperform those trained using conventional training methods on three reasoning benchmarks (ARC-Challenge, DROP, and StrategyQA). Our results suggest that training reliable verifiers requires ensuring the validity of rationales in addition to the correctness of the final answers, which would be critical for models assisting humans in solving complex reasoning tasks.
Authors: Zhuoqun Li, Xuanang Chen, Haiyang Yu, Hongyu Lin, Yaojie Lu, Qiaoyu Tang, Fei Huang, Xianpei Han, Le Sun, Yongbin Li
Abstract: Retrieval-augmented generation (RAG) is a key means to effectively enhance large language models (LLMs) in many knowledge-based tasks. However, existing RAG methods struggle with knowledge-intensive reasoning tasks, because useful information required to these tasks are badly scattered. This characteristic makes it difficult for existing RAG methods to accurately identify key information and perform global reasoning with such noisy augmentation. In this paper, motivated by the cognitive theories that humans convert raw information into various structured knowledge when tackling knowledge-intensive reasoning, we proposes a new framework, StructRAG, which can identify the optimal structure type for the task at hand, reconstruct original documents into this structured format, and infer answers based on the resulting structure. Extensive experiments across various knowledge-intensive tasks show that StructRAG achieves state-of-the-art performance, particularly excelling in challenging scenarios, demonstrating its potential as an effective solution for enhancing LLMs in complex real-world applications.
Authors: Lisa Dunlap, Krishna Mandal, Trevor Darrell, Jacob Steinhardt, Joseph E Gonzalez
Abstract: Large language models (LLMs) often exhibit subtle yet distinctive characteristics in their outputs that users intuitively recognize, but struggle to quantify. These "vibes" - such as tone, formatting, or writing style - influence user preferences, yet traditional evaluations focus primarily on the single axis of correctness. We introduce VibeCheck, a system for automatically comparing a pair of LLMs by discovering identifying traits of a model ("vibes") that are well-defined, differentiating, and user-aligned. VibeCheck iteratively discover vibes from model outputs, then utilizes a panel of LLM judges to quantitatively measure the utility of each vibe. We validate that the vibes generated by VibeCheck align with those found in human discovery and run VibeCheck on pairwise preference data from real-world user conversations with llama-3-70b VS GPT-4. VibeCheck reveals that Llama has a friendly, funny, and somewhat controversial vibe. These vibes predict model identity with 80% accuracy and human preference with 61% accuracy. Lastly, we run VibeCheck on a variety of models and tasks including summarization, math, and captioning to provide insight into differences in model behavior. VibeCheck discovers vibes like Command X prefers to add concrete intros and conclusions when summarizing in comparison to TNGL, Llama-405b often overexplains its thought process on math problems compared to GPT-4o, and GPT-4 prefers to focus on the mood and emotions of the scene when captioning compared to Gemini-1.5-Flash. Code can be found at https://github.com/lisadunlap/VibeCheck
Authors: Fabiha Haider, Fariha Tanjim Shifat, Md Farhan Ishmam, Deeparghya Dutta Barua, Md Sakib Ul Rahman Sourove, Md Fahim, Md Farhad Alam
Abstract: The proliferation of transliterated texts in digital spaces has emphasized the need for detecting and classifying hate speech in languages beyond English, particularly in low-resource languages. As online discourse can perpetuate discrimination based on target groups, e.g. gender, religion, and origin, multi-label classification of hateful content can help in comprehending hate motivation and enhance content moderation. While previous efforts have focused on monolingual or binary hate classification tasks, no work has yet addressed the challenge of multi-label hate speech classification in transliterated Bangla. We introduce BanTH, the first multi-label transliterated Bangla hate speech dataset comprising 37.3k samples. The samples are sourced from YouTube comments, where each instance is labeled with one or more target groups, reflecting the regional demographic. We establish novel transformer encoder-based baselines by further pre-training on transliterated Bangla corpus. We also propose a novel translation-based LLM prompting strategy for transliterated text. Experiments reveal that our further pre-trained encoders are achieving state-of-the-art performance on the BanTH dataset, while our translation-based prompting outperforms other strategies in the zero-shot setting. The introduction of BanTH not only fills a critical gap in hate speech research for Bangla but also sets the stage for future exploration into code-mixed and multi-label classification challenges in underrepresented languages.
Authors: Ruihan Yang, Caiqi Zhang, Zhisong Zhang, Xinting Huang, Sen Yang, Nigel Collier, Dong Yu, Deqing Yang
Abstract: While Large Language Models (LLMs) demonstrate impressive capabilities, they still struggle with generating factually incorrect content (i.e., hallucinations). A promising approach to mitigate this issue is enabling models to express uncertainty when unsure. Previous research on uncertainty modeling has primarily focused on short-form QA, but realworld applications often require much longer responses. In this work, we introduce the task of Long-form Generation with Uncertainty(LoGU). We identify two key challenges: Uncertainty Suppression, where models hesitate to express uncertainty, and Uncertainty Misalignment, where models convey uncertainty inaccurately. To tackle these challenges, we propose a refinement-based data collection framework and a two-stage training pipeline. Our framework adopts a divide-and-conquer strategy, refining uncertainty based on atomic claims. The collected data are then used in training through supervised fine-tuning (SFT) and direct preference optimization (DPO) to enhance uncertainty expression. Extensive experiments on three long-form instruction following datasets show that our method significantly improves accuracy, reduces hallucinations, and maintains the comprehensiveness of responses.
Authors: Lechen Zhang, Tolga Ergen, Lajanugen Logeswaran, Moontae Lee, David Jurgens
Abstract: Large Language Models (LLMs) have shown impressive capabilities in many scenarios, but their performance depends, in part, on the choice of prompt. Past research has focused on optimizing prompts specific to a task. However, much less attention has been given to optimizing the general instructions included in a prompt, known as a system prompt. To address this gap, we propose SPRIG, an edit-based genetic algorithm that iteratively constructs prompts from prespecified components to maximize the model's performance in general scenarios. We evaluate the performance of system prompts on a collection of 47 different types of tasks to ensure generalizability. Our study finds that a single optimized system prompt performs on par with task prompts optimized for each individual task. Moreover, combining system and task-level optimizations leads to further improvement, which showcases their complementary nature. Experiments also reveal that the optimized system prompts generalize effectively across model families, parameter sizes, and languages. This study provides insights into the role of system-level instructions in maximizing LLM potential.
Authors: Yu Zhao, Alessio Devoto, Giwon Hong, Xiaotang Du, Aryo Pradipta Gema, Hongru Wang, Xuanli He, Kam-Fai Wong, Pasquale Minervini
Abstract: Large language models (LLMs) can store a significant amount of factual knowledge in their parameters. However, their parametric knowledge may conflict with the information provided in the context -- this phenomenon, known as \emph{context-memory knowledge conflicts}, can lead to undesirable model behaviour, such as reliance on outdated or incorrect information. Analysing the internal activations of LLMs, we find that they can internally register the signals of knowledge conflict at mid-layers. Such signals allow us to detect whether a knowledge conflict occurs and use \emph{inference-time} intervention strategies to resolve it. In this work, we propose \textsc{SpARE}, a \emph{training-free} representation engineering method that uses pre-trained sparse auto-encoders (SAEs) to control the knowledge selection behaviour of LLMs. \textsc{SpARE} identifies the functional features that control the knowledge selection behaviours and applies them to edit the internal activations of LLMs at inference time. Our experimental results show that \textsc{SpARE} can effectively control the usage of either knowledge source to resolve knowledge conflict in open-domain question-answering tasks, surpassing existing representation engineering methods ($+10\%$) as well as contrastive decoding methods ($+15\%$).
Authors: Aryaman Arora, Dan Jurafsky, Christopher Potts, Noah D. Goodman
Abstract: In-context learning (ICL) is a powerful technique for getting language models to perform complex tasks with no training updates. Prior work has established strong correlations between the number of in-context examples provided and the accuracy of the model's predictions. In this paper, we seek to explain this correlation by showing that ICL approximates a Bayesian learner. This perspective gives rise to a family of novel Bayesian scaling laws for ICL. In experiments with \mbox{GPT-2} models of different sizes, our scaling laws exceed or match existing scaling laws in accuracy while also offering interpretable terms for task priors, learning efficiency, and per-example probabilities. To illustrate the analytic power that such interpretable scaling laws provide, we report on controlled synthetic dataset experiments designed to inform real-world studies of safety alignment. In our experimental protocol, we use SFT to suppress an unwanted existing model capability and then use ICL to try to bring that capability back (many-shot jailbreaking). We then experiment on real-world instruction-tuned LLMs using capabilities benchmarks as well as a new many-shot jailbreaking dataset. In all cases, Bayesian scaling laws accurately predict the conditions under which ICL will cause the suppressed behavior to reemerge, which sheds light on the ineffectiveness of post-training at increasing LLM safety.
Authors: Antoine Gorceix, Bastien Le Chenadec, Ahmad Rammal, Nelson Vadori, Manuela Veloso
Abstract: In this paper, we study the ability of large language models to learn specific mathematical rules such as distributivity or simplifying equations. We present an empirical analysis of their ability to generalize these rules, as well as to reuse them in the context of word problems. For this purpose, we provide a rigorous methodology to build synthetic data incorporating such rules, and perform fine-tuning of large language models on such data. Our experiments show that our model can learn and generalize these rules to some extent, as well as suitably reuse them in the context of word problems.
Authors: Wonje Jeung, Dongjae Jeon, Ashkan Yousefpour, Jonghyun Choi
Abstract: Existing fairness benchmarks for large language models (LLMs) primarily focus on simple tasks, such as multiple-choice questions, overlooking biases that may arise in more complex scenarios like long-text generation. To address this gap, we introduce the Long Text Fairness Test (LTF-TEST), a framework that evaluates biases in LLMs through essay-style prompts. LTF-TEST covers 14 topics and 10 demographic axes, including gender and race, resulting in 11,948 samples. By assessing both model responses and the reasoning behind them, LTF-TEST uncovers subtle biases that are difficult to detect in simple responses. In our evaluation of five recent LLMs, including GPT-4o and LLaMa3, we identify two key patterns of bias. First, these models frequently favor certain demographic groups in their responses. Second, they show excessive sensitivity toward traditionally disadvantaged groups, often providing overly protective responses while neglecting others. To mitigate these biases, we propose FT-REGARD, a finetuning approach that pairs biased prompts with neutral responses. FT-REGARD reduces gender bias by 34.6% and improves performance by 1.4 percentage points on the BBQ benchmark, offering a promising approach to addressing biases in long-text generation tasks.
Authors: Liangdong Wang, Bo-Wen Zhang, Chengwei Wu, Hanyu Zhao, Xiaofeng Shi, Shuhao Gu, Jijie Li, Quanyue Ma, TengFei Pan, Guang Liu
Abstract: We present CCI3.0-HQ (https://huggingface.co/datasets/BAAI/CCI3-HQ), a high-quality 500GB subset of the Chinese Corpora Internet 3.0 (CCI3.0)(https://huggingface.co/datasets/BAAI/CCI3-Data), developed using a novel two-stage hybrid filtering pipeline that significantly enhances data quality. To evaluate its effectiveness, we trained a 0.5B parameter model from scratch on 100B tokens across various datasets, achieving superior performance on 10 benchmarks in a zero-shot setting compared to CCI3.0, SkyPile, and WanjuanV1. The high-quality filtering process effectively distills the capabilities of the Qwen2-72B-instruct model into a compact 0.5B model, attaining optimal F1 scores for Chinese web data classification. We believe this open-access dataset will facilitate broader access to high-quality language models.
URLs: https://huggingface.co/datasets/BAAI/CCI3-HQ),, https://huggingface.co/datasets/BAAI/CCI3-Data),
Authors: Boyi Wei, Kaixuan Huang, Yangsibo Huang, Tinghao Xie, Xiangyu Qi, Mengzhou Xia, Prateek Mittal, Mengdi Wang, Peter Henderson
Abstract: Large language models (LLMs) show inherent brittleness in their safety mechanisms, as evidenced by their susceptibility to jailbreaking and even non-malicious fine-tuning. This study explores this brittleness of safety alignment by leveraging pruning and low-rank modifications. We develop methods to identify critical regions that are vital for safety guardrails, and that are disentangled from utility-relevant regions at both the neuron and rank levels. Surprisingly, the isolated regions we find are sparse, comprising about $3\%$ at the parameter level and $2.5\%$ at the rank level. Removing these regions compromises safety without significantly impacting utility, corroborating the inherent brittleness of the model's safety mechanisms. Moreover, we show that LLMs remain vulnerable to low-cost fine-tuning attacks even when modifications to the safety-critical regions are restricted. These findings underscore the urgent need for more robust safety strategies in LLMs.
Authors: Hengyuan Zhang, Zitao Liu, Shuyan Huang, Chenming Shang, Bojun Zhan, Yong Jiang
Abstract: Knowledge tracing (KT) aims to estimate student's knowledge mastery based on their historical interactions. Recently, the deep learning based KT (DLKT) approaches have achieved impressive performance in the KT task. These DLKT models heavily rely on the large number of available student interactions. However, due to various reasons such as budget constraints and privacy concerns, observed interactions are very limited in many real-world scenarios, a.k.a, low-resource KT datasets. Directly training a DLKT model on a low-resource KT dataset may lead to overfitting and it is difficult to choose the appropriate deep neural architecture. Therefore, in this paper, we propose a low-resource KT framework called LoReKT to address above challenges. Inspired by the prevalent "pre-training and fine-tuning" paradigm, we aim to learn transferable parameters and representations from rich-resource KT datasets during the pre-training stage and subsequently facilitate effective adaptation to low-resource KT datasets. Specifically, we simplify existing sophisticated DLKT model architectures with purely a stack of transformer decoders. We design an encoding mechanism to incorporate student interactions from multiple KT data sources and develop an importance mechanism to prioritize updating parameters with high importance while constraining less important ones during the fine-tuning stage. We evaluate LoReKT on six public KT datasets and experimental results demonstrate the superiority of our approach in terms of AUC and Accuracy. To encourage reproducible research, we make our data and code publicly available at https://github.com/rattlesnakey/LoReKT.
Authors: Aneesh Pappu, Billy Porter, Ilia Shumailov, Jamie Hayes
Abstract: Reinforcement learning with human feedback (RLHF) has become the dominant method to align large models to user preferences. Unlike fine-tuning, for which there are many studies regarding training data memorization, it is not clear how memorization is affected by or introduced in the RLHF alignment process. Understanding this relationship is important as real user data may be collected and used to align large models; if user data is memorized during RLHF and later regurgitated, this could raise privacy concerns. In addition to RLHF, other methods such as Direct Preference Optimization (DPO) and $\Psi$PO have gained popularity for learning directly from human preferences, removing the need for optimizing intermediary reward models with reinforcement learning. In this work, we analyze how training data memorization can surface and propagate through each phase of RLHF and direct preference learning. We focus our study on code completion models, as code completion is one of the most popular use cases for large language models. We find that RLHF significantly decreases the chance that data used for reward modeling and reinforcement learning is memorized in comparison to directly fine-tuning on this data, but that examples already memorized during the fine-tuning stage of RLHF, will, in the majority of cases, remain memorized after RLHF. In contrast, we find that aligning by learning directly from human preference data via a special case of $\Psi$PO, Identity Preference Optimization (IPO), increases the likelihood that training data is regurgitated compared to RLHF. Our work suggests that RLHF, as opposed to direct preference learning, is a safer way to mitigate the risk of regurgitating sensitive preference data when aligning large language models. We find our conclusions are robust across multiple code completion datasets, tasks, and model scales.
Authors: Rickard Br\"uel-Gabrielsson, Jiacheng Zhu, Onkar Bhardwaj, Leshem Choshen, Kristjan Greenewald, Mikhail Yurochkin, Justin Solomon
Abstract: Fine-tuning large language models (LLMs) with low-rank adaptations (LoRAs) has become common practice, often yielding numerous copies of the same LLM differing only in their LoRA updates. This paradigm presents challenges for systems that serve real-time responses to queries that each involve a different LoRA. Prior works optimize the design of such systems but still require continuous loading and offloading of LoRAs, as it is infeasible to store thousands of LoRAs in GPU memory. To mitigate this issue, we investigate the efficacy of model compression when serving LoRAs. We propose a method for joint compression of LoRAs into a shared basis paired with LoRA-specific scaling matrices. We extend our algorithm to learn clusters of LoRAs that are more amenable to joint compression, allowing it to scale gracefully to large LoRA collections. Our experiments with up to 500 LoRAs demonstrate that compressed LoRAs preserve performance while offering major throughput gains in realistic serving scenarios with over a thousand LoRAs, maintaining 80% of the throughput of serving a single LoRA.
Authors: Zafeirios Fountas, Martin A Benfeghoul, Adnan Oomerjee, Fenia Christopoulou, Gerasimos Lampouras, Haitham Bou-Ammar, Jun Wang
Abstract: Large language models (LLMs) have shown remarkable capabilities, but still struggle with processing extensive contexts, limiting their ability to maintain coherence and accuracy over long sequences. In contrast, the human brain excels at organising and retrieving episodic experiences across vast temporal scales, spanning a lifetime. In this work, we introduce EM-LLM, a novel approach that integrates key aspects of human episodic memory and event cognition into LLMs with no fine-tuning, enabling them to handle practically infinite context lengths while maintaining computational efficiency. EM-LLM organises sequences of tokens into coherent episodic events using a combination of Bayesian surprise and graph-theoretic boundary refinement in an online fashion. When needed, these events are retrieved through a two-stage memory process, combining similarity-based and temporally contiguous retrieval for efficient and human-like access to relevant information. Experiments on the LongBench and InfiniteBench benchmarks demonstrate EM-LLM's superior performance, consistently outperforming the state-of-the-art retrieval model InfLLM across various baseline LLMs. In addition, EM-LLM outperforms its popular counterpart, RAG, in a wide range of tasks, while requiring similar resources. Notably, EM-LLM's performance even surpasses full-context models in most tasks, while successfully performing retrieval across 10 million tokens - a scale computationally infeasible for such models. Finally, our analysis reveals strong correlations between EM-LLM's event segmentation and human-perceived events, suggesting a bridge between this artificial system and its biological counterpart, thereby offering a novel computational framework for exploring human memory mechanisms.
Authors: Kaiwen Zheng, Yongxin Chen, Hanzi Mao, Ming-Yu Liu, Jun Zhu, Qinsheng Zhang
Abstract: Masked diffusion models (MDMs) have emerged as a popular research topic for generative modeling of discrete data, thanks to their superior performance over other discrete diffusion models, and are rivaling the auto-regressive models (ARMs) for language modeling tasks. The recent effort in simplifying the masked diffusion framework further leads to alignment with continuous-space diffusion models and more principled training and sampling recipes. In this paper, however, we reveal that both training and sampling of MDMs are theoretically free from the time variable, arguably the key signature of diffusion models, and are instead equivalent to masked models. The connection on the sampling aspect is drawn by our proposed first-hitting sampler (FHS). Specifically, we show that the FHS is theoretically equivalent to MDMs' original generation process while significantly alleviating the time-consuming categorical sampling and achieving a 20$\times$ speedup. In addition, our investigation raises doubts about whether MDMs can truly beat ARMs in text generation. We identify, for the first time, an underlying numerical issue, even with the commonly used 32-bit floating-point precision, which results in inaccurate categorical sampling. We show that it lowers the effective temperature both theoretically and empirically, and the resulting decrease in token diversity makes previous evaluations, which assess the generation quality solely through the incomplete generative perplexity metric, somewhat unfair.
Authors: Arkadeep Acharya, Rudra Murthy, Vishwajeet Kumar, Jaydeep Sen
Abstract: Given the large number of Hindi speakers worldwide, there is a pressing need for robust and efficient information retrieval systems for Hindi. Despite ongoing research, comprehensive benchmarks for evaluating retrieval models in Hindi are lacking. To address this gap, we introduce the Hindi-BEIR benchmark, comprising 15 datasets across seven distinct tasks. We evaluate state-of-the-art multilingual retrieval models on the Hindi-BEIR benchmark, identifying task and domain-specific challenges that impact Hindi retrieval performance. Building on the insights from these results, we introduce NLLB-E5, a multilingual retrieval model that leverages a zero-shot approach to support Hindi without the need for Hindi training data. We believe our contributions, which include the release of the Hindi-BEIR benchmark and the NLLB-E5 model, will prove to be a valuable resource for researchers and promote advancements in multilingual retrieval models.
Authors: Chuang Niu, Parisa Kaviani, Qing Lyu, Mannudeep K. Kalra, Christopher T. Whitlow, Ge Wang
Abstract: Current LLMs for creating fully-structured reports face the challenges of formatting errors, content hallucinations, and privacy leakage issues when uploading data to external servers.We aim to develop an open-source, accurate LLM for creating fully-structured and standardized LCS reports from varying free-text reports across institutions and demonstrate its utility in automatic statistical analysis and individual lung nodule retrieval. With IRB approvals, our retrospective study included 5,442 de-identified LDCT LCS radiology reports from two institutions. We constructed two evaluation datasets by labeling 500 pairs of free-text and fully-structured radiology reports and one large-scale consecutive dataset from January 2021 to December 2023. Two radiologists created a standardized template for recording 27 lung nodule features on LCS. We designed a dynamic-template-constrained decoding method to enhance existing LLMs for creating fully-structured reports from free-text radiology reports. Using consecutive structured reports, we automated descriptive statistical analyses and a nodule retrieval prototype. Our best LLM for creating fully-structured reports achieved high performance on cross-institutional datasets with an F1 score of about 97%, with neither formatting errors nor content hallucinations. Our method consistently improved the best open-source LLMs by up to 10.42%, and outperformed GPT-4o by 17.19%. The automatically derived statistical distributions were consistent with prior findings regarding attenuation, location, size, stability, and Lung-RADS. The retrieval system with structured reports allowed flexible nodule-level search and complex statistical analysis. Our developed software is publicly available for local deployment and further research.
Authors: Joost de Winter, Dimitra Dodou, Yke Bauke Eisma
Abstract: The processes underlying human cognition are often divided into System 1, which involves fast, intuitive thinking, and System 2, which involves slow, deliberate reasoning. Previously, large language models were criticized for lacking the deeper, more analytical capabilities of System 2. In September 2024, OpenAI introduced the o1 model series, designed to handle System 2-like reasoning. While OpenAI's benchmarks are promising, independent validation is still needed. In this study, we tested the o1-preview model twice on the Dutch 'Mathematics B' final exam. It scored a near-perfect 76 and 74 out of 76 points. For context, only 24 out of 16,414 students in the Netherlands achieved a perfect score. By comparison, the GPT-4o model scored 66 and 62 out of 76, well above the Dutch students' average of 40.63 points. Neither model had access to the exam figures. Since there was a risk of model contami-nation (i.e., the knowledge cutoff for o1-preview and GPT-4o was after the exam was published online), we repeated the procedure with a new Mathematics B exam that was published after the cutoff date. The results again indicated that o1-preview performed strongly (97.8th percentile), which suggests that contamination was not a factor. We also show that there is some variability in the output of o1-preview, which means that sometimes there is 'luck' (the answer is correct) or 'bad luck' (the output has diverged into something that is incorrect). We demonstrate that the self-consistency approach, where repeated prompts are given and the most common answer is selected, is a useful strategy for identifying the correct answer. It is concluded that while OpenAI's new model series holds great potential, certain risks must be considered.
Authors: Vishnu Sarukkai, Brennan Shacklett, Zander Majercik, Kush Bhatia, Christopher R\'e, Kayvon Fatahalian
Abstract: Large Language Models (LLMs) have the potential to automate reward engineering by leveraging their broad domain knowledge across various tasks. However, they often need many iterations of trial-and-error to generate effective reward functions. This process is costly because evaluating every sampled reward function requires completing the full policy optimization process for each function. In this paper, we introduce an LLM-driven reward generation framework that is able to produce state-of-the-art policies on the challenging Bi-DexHands benchmark with 20x fewer reward function samples than the prior state-of-the-art work. Our key insight is that we reduce the problem of generating task-specific rewards to the problem of coarsely estimating task progress. Our two-step solution leverages the task domain knowledge and the code synthesis abilities of LLMs to author progress functions that estimate task progress from a given state. Then, we use this notion of progress to discretize states, and generate count-based intrinsic rewards using the low-dimensional state space. We show that the combination of LLM-generated progress functions and count-based intrinsic rewards is essential for our performance gains, while alternatives such as generic hash-based counts or using progress directly as a reward function fall short.
Authors: Michael Zhang, Simran Arora, Rahul Chalamala, Alan Wu, Benjamin Spector, Aaryan Singhal, Krithik Ramesh, Christopher R\'e
Abstract: Recent works show we can linearize large language models (LLMs) -- swapping the quadratic attentions of popular Transformer-based LLMs with subquadratic analogs, such as linear attention -- avoiding the expensive pretraining costs. However, linearizing LLMs often significantly degrades model quality, still requires training over billions of tokens, and remains limited to smaller 1.3B to 7B LLMs. We thus propose Low-rank Linear Conversion via Attention Transfer (LoLCATs), a simple two-step method that improves LLM linearizing quality with orders of magnitudes less memory and compute. We base these steps on two findings. First, we can replace an LLM's softmax attentions with closely-approximating linear attentions, simply by training the linear attentions to match their softmax counterparts with an output MSE loss ("attention transfer"). Then, this enables adjusting for approximation errors and recovering LLM quality simply with low-rank adaptation (LoRA). LoLCATs significantly improves linearizing quality, training efficiency, and scalability. We significantly reduce the linearizing quality gap and produce state-of-the-art subquadratic LLMs from Llama 3 8B and Mistral 7B v0.1, leading to 20+ points of improvement on 5-shot MMLU. Furthermore, LoLCATs does so with only 0.2% of past methods' model parameters and 0.4% of their training tokens. Finally, we apply LoLCATs to create the first linearized 70B and 405B LLMs (50x larger than prior work). When compared with prior approaches under the same compute budgets, LoLCATs significantly improves linearizing quality, closing the gap between linearized and original Llama 3.1 70B and 405B LLMs by 77.8% and 78.1% on 5-shot MMLU.
Authors: Long Li, Weiwen Xu, Jiayan Guo, Ruochen Zhao, Xinxuan Li, Yuqian Yuan, Boqiang Zhang, Yuming Jiang, Yifei Xin, Ronghao Dang, Deli Zhao, Yu Rong, Tian Feng, Lidong Bing
Abstract: Effective research ideation is a critical step for scientific research. However, the exponential increase in scientific literature makes it challenging for researchers to stay current with recent advances and identify meaningful research directions. Recent developments in large language models~(LLMs) suggest a promising avenue for automating the generation of novel research ideas. However, existing methods for idea generation either trivially prompt LLMs or directly expose LLMs to extensive literature without indicating useful information. Inspired by the research process of human researchers, we propose a Chain-of-Ideas~(CoI) agent, an LLM-based agent that organizes relevant literature in a chain structure to effectively mirror the progressive development in a research domain. This organization facilitates LLMs to capture the current advancements in research, thereby enhancing their ideation capabilities. Furthermore, we propose Idea Arena, an evaluation protocol that can comprehensively evaluate idea generation methods from different perspectives, aligning closely with the preferences of human researchers. Experimental results indicate that the CoI agent consistently outperforms other methods and shows comparable quality as humans in research idea generation. Moreover, our CoI agent is budget-friendly, with a minimum cost of \$0.50 to generate a candidate idea and its corresponding experimental design.
Authors: Ted Kwartler, Matthew Berman, Alan Aqrawi
Abstract: This study explores the ability of Large Language Model (LLM) agents to detect and correct hallucinations in AI-generated content. A primary agent was tasked with creating a blog about a fictional Danish artist named Flipfloppidy, which was then reviewed by another agent for factual inaccuracies. Most LLMs hallucinated the existence of this artist. Across 4,900 test runs involving various combinations of primary and reviewing agents, advanced AI models such as Llama3-70b and GPT-4 variants demonstrated near-perfect accuracy in identifying hallucinations and successfully revised outputs in 85% to 100% of cases following feedback. These findings underscore the potential of advanced AI models to significantly enhance the accuracy and reliability of generated content, providing a promising approach to improving AI workflow orchestration.
Authors: Jiaxuan Gao, Shusheng Xu, Wenjie Ye, Weilin Liu, Chuyi He, Wei Fu, Zhiyu Mei, Guangju Wang, Yi Wu
Abstract: Reward models have been increasingly critical for improving the reasoning capability of LLMs. Existing research has shown that a well-trained reward model can substantially improve model performances at inference time via search. However, the potential of reward models during RL training time still remains largely under-explored. It is currently unclear whether these reward models can provide additional training signals to enhance the reasoning capabilities of LLMs in RL training that uses sparse success rewards, which verify the correctness of solutions. In this work, we evaluate popular reward models for RL training, including the Outcome-supervised Reward Model (ORM) and the Process-supervised Reward Model (PRM), and train a collection of LLMs for math problems using RL by combining these learned rewards with success rewards. Surprisingly, even though these learned reward models have strong inference-time performances, they may NOT help or even hurt RL training, producing worse performances than LLMs trained with the success reward only. Our analysis reveals that an LLM can receive high rewards from some of these reward models by repeating correct but unnecessary reasoning steps, leading to a severe reward hacking issue. Therefore, we introduce two novel reward refinement techniques, including Clipping and Delta. The key idea is to ensure the accumulative reward of any reasoning trajectory is upper-bounded to keep a learned reward model effective without being exploited. We evaluate our techniques with multiple reward models over a set of 1.5B and 7B LLMs on MATH and GSM8K benchmarks and demonstrate that with a carefully designed reward function, RL training without any additional supervised tuning can improve all the evaluated LLMs, including the state-of-the-art 7B LLM Qwen2.5-Math-7B-Instruct on MATH and GSM8K benchmarks.
Authors: Mete Ismayilzada, Debjit Paul, Antoine Bosselut, Lonneke van der Plas
Abstract: Creativity is the ability to produce novel, useful, and surprising ideas, and has been widely studied as a crucial aspect of human cognition. Machine creativity on the other hand has been a long-standing challenge. With the rise of advanced generative AI, there has been renewed interest and debate regarding AI's creative capabilities. Therefore, it is imperative to revisit the state of creativity in AI and identify key progresses and remaining challenges. In this work, we survey leading works studying the creative capabilities of AI systems, focusing on creative problem-solving, linguistic, artistic, and scientific creativity. Our review suggests that while the latest AI models are largely capable of producing linguistically and artistically creative outputs such as poems, images, and musical pieces, they struggle with tasks that require creative problem-solving, abstract thinking and compositionality and their generations suffer from a lack of diversity, originality, long-range incoherence and hallucinations. We also discuss key questions concerning copyright and authorship issues with generative models. Furthermore, we highlight the need for a comprehensive evaluation of creativity that is process-driven and considers several dimensions of creativity. Finally, we propose future research directions to improve the creativity of AI outputs, drawing inspiration from cognitive science and psychology.
Authors: Woosung Koh, Jang Han Yoon, MinHyung Lee, Youngjin Song, Jaegwan Cho, Jaehyun Kang, Taehyeon Kim, Se-young Yun, Youngjae Yu, Bongshin Lee
Abstract: Generating high-quality charts with Large Language Models presents significant challenges due to limited data and the high cost of scaling through human curation. Instruction, data, and code triplets are scarce and expensive to manually curate as their creation demands technical expertise. To address this scalability issue, we introduce a reference-free automatic feedback generator, which eliminates the need for costly human intervention. Our novel framework, $C^2$, consists of (1) an automatic feedback provider (ChartAF) and (2) a diverse, reference-free dataset (ChartUIE-8K). Quantitative results are compelling: in our first experiment, 74% of respondents strongly preferred, and 10% preferred, the results after feedback. The second post-feedback experiment demonstrates that ChartAF outperforms nine baselines. Moreover, ChartUIE-8K significantly improves data diversity by increasing queries, datasets, and chart types by 5982%, 1936%, and 91%, respectively, over benchmarks. Finally, an LLM user study revealed that 94% of participants preferred ChartUIE-8K's queries, with 93% deeming them aligned with real-world use cases. Core contributions are available as open-source at an anonymized project site, with ample qualitative examples.
Authors: Yuxing Chen, Weijie Wang, Sylvain Lobry, Camille Kurtz
Abstract: Large language models (LLMs) are being used in data science code generation tasks, but they often struggle with complex sequential tasks, leading to logical errors. Their application to geospatial data processing is particularly challenging due to difficulties in incorporating complex data structures and spatial constraints, effectively utilizing diverse function calls, and the tendency to hallucinate less-used geospatial libraries. To tackle these problems, we introduce GeoAgent, a new interactive framework designed to help LLMs handle geospatial data processing more effectively. GeoAgent pioneers the integration of a code interpreter, static analysis, and Retrieval-Augmented Generation (RAG) techniques within a Monte Carlo Tree Search (MCTS) algorithm, offering a novel approach to geospatial data processing. In addition, we contribute a new benchmark specifically designed to evaluate the LLM-based approach in geospatial tasks. This benchmark leverages a variety of Python libraries and includes both single-turn and multi-turn tasks such as data acquisition, data analysis, and visualization. By offering a comprehensive evaluation among diverse geospatial contexts, this benchmark sets a new standard for developing LLM-based approaches in geospatial data analysis tasks. Our findings suggest that relying solely on knowledge of LLM is insufficient for accurate geospatial task programming, which requires coherent multi-step processes and multiple function calls. Compared to the baseline LLMs, the proposed GeoAgent has demonstrated superior performance, yielding notable improvements in function calls and task completion. In addition, these results offer valuable insights for the future development of LLM agents in automatic geospatial data analysis task programming.