Authors: Anahita Samadi, Allison Sullivan
Abstract: As machine learning systems become more widely used, especially for safety critical applications, there is a growing need to ensure that these systems behave as intended, even in the face of adversarial examples. Adversarial examples are inputs that are designed to trick the decision making process, and are intended to be imperceptible to humans. However, for text-based classification systems, changes to the input, a string of text, are always perceptible. Therefore, text-based adversarial examples instead focus on trying to preserve semantics. Unfortunately, recent work has shown this goal is often not met. To improve the quality of text-based adversarial examples, we need to know what elements of the input text are worth focusing on. To address this, in this paper, we explore what parts of speech have the highest impact of text-based classifiers. Our experiments highlight a distinct bias in CNN algorithms against certain parts of speech tokens within review datasets. This finding underscores a critical vulnerability in the linguistic processing capabilities of CNNs.
Authors: Krisztian Balog, John Palowitch, Barbara Ikica, Filip Radlinski, Hamidreza Alvari, Mehdi Manshadi
Abstract: The emergence of synthetic data represents a pivotal shift in modern machine learning, offering a solution to satisfy the need for large volumes of data in domains where real data is scarce, highly private, or difficult to obtain. We investigate the feasibility of creating realistic, large-scale synthetic datasets of user-generated content, noting that such content is increasingly prevalent and a source of frequently sought information. Large language models (LLMs) offer a starting point for generating synthetic social media discussion threads, due to their ability to produce diverse responses that typify online interactions. However, as we demonstrate, straightforward application of LLMs yields limited success in capturing the complex structure of online discussions, and standard prompting mechanisms lack sufficient control. We therefore propose a multi-step generation process, predicated on the idea of creating compact representations of discussion threads, referred to as scaffolds. Our framework is generic yet adaptable to the unique characteristics of specific social media platforms. We demonstrate its feasibility using data from two distinct online discussion platforms. To address the fundamental challenge of ensuring the representativeness and realism of synthetic data, we propose a portfolio of evaluation measures to compare various instantiations of our framework.
Authors: Mohammad Ghiasvand Mohammadkhani, Ali Ghiasvand Mohammadkhani, Hamid Beigy
Abstract: Automated fact-checking is an important task because determining the accurate status of a proposed claim within the vast amount of information available online is a critical challenge. This challenge requires robust evaluation to prevent the spread of false information. Modern large language models (LLMs) have demonstrated high capability in performing a diverse range of Natural Language Processing (NLP) tasks. By utilizing proper prompting strategies, their versatility due to their understanding of large context sizes and zero-shot learning ability enables them to simulate human problem-solving intuition and move towards being an alternative to humans for solving problems. In this work, we introduce a straightforward framework based on Zero-Shot Learning and Key Points (ZSL-KeP) for automated fact-checking, which despite its simplicity, performed well on the AVeriTeC shared task dataset by robustly improving the baseline and achieving 10th place.
Authors: Deepak Pandita, Tharindu Cyril Weerasooriya, Sujan Dutta, Sarah K. Luger, Tharindu Ranasinghe, Ashiqur R. KhudaBukhsh, Marcos Zampieri, Christopher M. Homan
Abstract: Human feedback is essential for building human-centered AI systems across domains where disagreement is prevalent, such as AI safety, content moderation, or sentiment analysis. Many disagreements, particularly in politically charged settings, arise because raters have opposing values or beliefs. Vicarious annotation is a method for breaking down disagreement by asking raters how they think others would annotate the data. In this paper, we explore the use of vicarious annotation with analytical methods for moderating rater disagreement. We employ rater cohesion metrics to study the potential influence of political affiliations and demographic backgrounds on raters' perceptions of offense. Additionally, we utilize CrowdTruth's rater quality metrics, which consider the demographics of the raters, to score the raters and their annotations. We study how the rater quality metrics influence the in-group and cross-group rater cohesion across the personal and vicarious levels.
Authors: Jinming Nian, Zhiyuan Peng, Qifan Wang, Yi Fang
Abstract: In knowledge-intensive tasks such as open-domain question answering (OpenQA), Large Language Models (LLMs) often struggle to generate factual answers relying solely on their internal (parametric) knowledge. To address this limitation, Retrieval-Augmented Generation (RAG) systems enhance LLMs by retrieving relevant information from external sources, thereby positioning the retriever as a pivotal component. Although dense retrieval demonstrates state-of-the-art performance, its training poses challenges due to the scarcity of ground-truth evidence, largely attributed to the high costs of human annotation. In this paper, we propose W-RAG by utilizing the ranking capabilities of LLMs to create weakly labeled data for training dense retrievers. Specifically, we rerank the top-$K$ passages retrieved via BM25 by assessing the probability that LLMs will generate the correct answer based on the question and each passage. The highest-ranking passages are then used as positive training examples for dense retrieval. Our comprehensive experiments across four publicly available OpenQA datasets demonstrate that our approach enhances both retrieval and OpenQA performance compared to baseline models.
Authors: Xiaochuang Han, Marjan Ghazvininejad, Pang Wei Koh, Yulia Tsvetkov
Abstract: Recent work in image and video generation has been adopting the autoregressive LLM architecture due to its generality and potentially easy integration into multi-modal systems. The crux of applying autoregressive training in language generation to visual generation is discretization -- representing continuous data like images and videos as discrete tokens. Common methods of discretizing images and videos include modeling raw pixel values, which are prohibitively lengthy, or vector quantization, which requires convoluted pre-hoc training. In this work, we propose to directly model images and videos as compressed files saved on computers via canonical codecs (e.g., JPEG, AVC/H.264). Using the default Llama architecture without any vision-specific modifications, we pretrain JPEG-LM from scratch to generate images (and AVC-LM to generate videos as a proof of concept), by directly outputting compressed file bytes in JPEG and AVC formats. Evaluation of image generation shows that this simple and straightforward approach is more effective than pixel-based modeling and sophisticated vector quantization baselines (on which our method yields a 31% reduction in FID). Our analysis shows that JPEG-LM has an especial advantage over vector quantization models in generating long-tail visual elements. Overall, we show that using canonical codec representations can help lower the barriers between language generation and visual generation, facilitating future research on multi-modal language/image/video LLMs.
Authors: Huang Lei, Jiaming Guo, Guanhua He, Xishan Zhang, Rui Zhang, Shaohui Peng, Shaoli Liu, Tianshi Chen
Abstract: Generating long-term texts such as novels using artificial intelligence has always been a challenge. A common approach is to use large language models (LLMs) to construct a hierarchical framework that first plans and then writes. Despite the fact that the generated novels reach a sufficient length, they exhibit poor logical coherence and appeal in their plots and deficiencies in character and event depiction, ultimately compromising the overall narrative quality. In this paper, we propose a method named Extracting Excelsior and Expanding. Ex3 initially extracts structure information from raw novel data. By combining this structure information with the novel data, an instruction-following dataset is meticulously crafted. This dataset is then utilized to fine-tune the LLM, aiming for excelsior generation performance. In the final stage, a tree-like expansion method is deployed to facilitate the generation of arbitrarily long novels. Evaluation against previous methods showcases Ex3's ability to produce higher-quality long-form novels.
Authors: Rong-Ching Chang, Jiawei Zhang
Abstract: Despite advancements in Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG) systems, their effectiveness is often hindered by a lack of integration with entity relationships and community structures, limiting their ability to provide contextually rich and accurate information retrieval for fact-checking. We introduce CommunityKG-RAG (Community Knowledge Graph-Retrieval Augmented Generation), a novel zero-shot framework that integrates community structures within Knowledge Graphs (KGs) with RAG systems to enhance the fact-checking process. Capable of adapting to new domains and queries without additional training, CommunityKG-RAG utilizes the multi-hop nature of community structures within KGs to significantly improve the accuracy and relevance of information retrieval. Our experimental results demonstrate that CommunityKG-RAG outperforms traditional methods, representing a significant advancement in fact-checking by offering a robust, scalable, and efficient solution.
Authors: Renato Lui Geh, Honghua Zhang, Kareem Ahmed, Benjie Wang, Guy Van den Broeck
Abstract: Large Language Models (LLMs) are typically shipped with tokenizers that deterministically encode text into so-called canonical token sequences, to which the LLMs assign probability values. One common assumption is that the probability of a piece of text is the probability of its canonical token sequence. However, the tokenization of a string is not unique: e.g., the Llama2 tokenizer encodes Tokens as [Tok,ens], but [Tok,en,s] also represents the same text. In this paper, we study non-canonical tokenizations. We prove that, given a string, it is computationally hard to find the most likely tokenization for an autoregressive LLM, as well as to compute the marginal probability over all possible tokenizations. We then show how the marginal is, in most cases, indistinguishable from the canonical probability. Surprisingly, we then empirically demonstrate the existence of a significant amount of signal hidden within tokenization space. Notably, by simply aggregating the probabilities of non-canonical tokenizations, we achieve improvements across a range of LLM evaluation benchmarks for a variety of architectures, including transformers and state space models.
Authors: Kaushal Kumar Maurya, KV Aditya Srivatsa, Ekaterina Kochmar
Abstract: Large language models (LLMs) have gained increased popularity due to their remarkable success across various tasks, which has led to the active development of a large set of diverse LLMs. However, individual LLMs have limitations when applied to complex tasks because of such factors as training biases, model sizes, and the datasets used. A promising approach is to efficiently harness the diverse capabilities of LLMs to overcome these individual limitations. Towards this goal, we introduce a novel LLM selection algorithm called SelectLLM. This algorithm directs input queries to the most suitable subset of LLMs from a large pool, ensuring they collectively provide the correct response efficiently. SelectLLM uses a multi-label classifier, utilizing the classifier's predictions and confidence scores to design optimal policies for selecting an optimal, query-aware, and lightweight subset of LLMs. Our findings show that the proposed model outperforms individual LLMs and achieves competitive performance compared to similarly sized, computationally expensive top-performing LLM subsets. Specifically, with a similarly sized top-performing LLM subset, we achieve a significant reduction in latency on two standard reasoning benchmarks: 13% lower latency for GSM8K and 70% lower latency for MMLU. Additionally, we conduct comprehensive analyses and ablation studies, which validate the robustness of the proposed model.
Authors: Haohao Zhu, Xiaokun Zhang, Junyu Lu, Liang Yang, Hongfei Lin
Abstract: Textual personality detection aims to identify personality traits by analyzing user-generated content. To achieve this effectively, it is essential to thoroughly examine user-generated content from various perspectives. However, previous studies have struggled with automatically extracting and effectively integrating information from multiple perspectives, thereby limiting their performance on personality detection. To address these challenges, we propose the Multi-view Mixture-of-Experts Model for Textual Personality Detection (MvP). MvP introduces a Multi-view Mixture-of-Experts (MoE) network to automatically analyze user posts from various perspectives. Additionally, it employs User Consistency Regularization to mitigate conflicts among different perspectives and learn a multi-view generic user representation. The model's training is optimized via a multi-task joint learning strategy that balances supervised personality detection with self-supervised user consistency constraints. Experimental results on two widely-used personality detection datasets demonstrate the effectiveness of the MvP model and the benefits of automatically analyzing user posts from diverse perspectives for textual personality detection.
Authors: Tomas Goldsack, Carolina Scarton, Matthew Shardlow, Chenghua Lin
Abstract: This paper presents the setup and results of the second edition of the BioLaySumm shared task on the Lay Summarisation of Biomedical Research Articles, hosted at the BioNLP Workshop at ACL 2024. In this task edition, we aim to build on the first edition's success by further increasing research interest in this important task and encouraging participants to explore novel approaches that will help advance the state-of-the-art. Encouragingly, we found research interest in the task to be high, with this edition of the task attracting a total of 53 participating teams, a significant increase in engagement from the previous edition. Overall, our results show that a broad range of innovative approaches were adopted by task participants, with a predictable shift towards the use of Large Language Models (LLMs).
Authors: Geonhee Kim, Marco Valentino, Andr\'e Freitas
Abstract: Recent studies on logical reasoning in auto-regressive Language Models (LMs) have sparked a debate on whether such models can learn systematic reasoning principles during pre-training or merely exploit superficial patterns in the training data. This paper presents a mechanistic interpretation of syllogistic reasoning in LMs to further enhance our understanding of internal dynamics. Specifically, we present a methodology for circuit discovery aimed at disentangling content-independent reasoning mechanisms from world knowledge acquired during pre-training. Through two distinct intervention methods, we uncover a sufficient and necessary circuit involving middle-term suppression that elucidates how LMs transfer information to derive valid conclusions from premises. Furthermore, we investigate how belief biases manifest in syllogistic reasoning, finding evidence of partial contamination from additional attention heads responsible for encoding commonsense and contextualized knowledge. Finally, we explore the generalization of the discovered mechanisms across various syllogistic schemes and model sizes, finding that the identified circuit is sufficient and necessary for all the schemes on which the model achieves high downstream accuracy ($\geq$ 60\%). Overall, our findings suggest that LMs indeed learn transferable content-independent reasoning mechanisms, but that, at the same time, such mechanisms do not involve generalisable and abstract logical primitives, being susceptible to contamination by the same world knowledge acquired during pre-training.
Authors: Gregory Kell, Angus Roberts, Serge Umansky, Yuti Khare, Najma Ahmed, Nikhil Patel, Chloe Simela, Jack Coumbe, Julian Rozario, Ryan-Rhys Griffiths, Iain J. Marshall
Abstract: Clinical question answering systems have the potential to provide clinicians with relevant and timely answers to their questions. Nonetheless, despite the advances that have been made, adoption of these systems in clinical settings has been slow. One issue is a lack of question-answering datasets which reflect the real-world needs of health professionals. In this work, we present RealMedQA, a dataset of realistic clinical questions generated by humans and an LLM. We describe the process for generating and verifying the QA pairs and assess several QA models on BioASQ and RealMedQA to assess the relative difficulty of matching answers to questions. We show that the LLM is more cost-efficient for generating "ideal" QA pairs. Additionally, we achieve a lower lexical similarity between questions and answers than BioASQ which provides an additional challenge to the top two QA models, as per the results. We release our code and our dataset publicly to encourage further research.
Authors: Junseok Kim, Nakyeong Yang, Kyomin Jung
Abstract: Recent studies demonstrate that prompting an appropriate role-playing persona to an LLM improves its reasoning capability. However, assigning a proper persona is difficult since an LLM's performance is extremely sensitive to assigned prompts; therefore, personas sometimes hinder LLMs and degrade their reasoning capabilities. In this paper, we propose a novel framework, Jekyll \& Hyde, which ensembles the results of role-playing and neutral prompts to eradicate performance degradation via unilateral use of role-playing prompted LLM and enhance the robustness of an LLM's reasoning ability. Specifically, Jekyll \& Hyde collects two potential solutions from both role-playing and neutral prompts and selects a better solution after cross-checking via an LLM evaluator. However, LLM-based evaluators tend to be affected by the order of those potential solutions within the prompt when selecting the proper solution; thus, we also propose a robust LLM evaluator to mitigate the position bias. The experimental analysis demonstrates that role-playing prompts distract LLMs and degrade their reasoning abilities in 4 out of 12 datasets, even when using GPT-4. In addition, we reveal that Jekyll \& Hyde improves reasoning capabilities by selecting better choices among the potential solutions on twelve widely-used reasoning datasets. We further show that our proposed LLM evaluator outperforms other baselines, proving the LLMs' position bias is successfully mitigated.
Authors: Jian Li, Weiheng Lu
Abstract: Multimodal Large Language Models (MLLMs) are gaining increasing popularity in both academia and industry due to their remarkable performance in various applications such as visual question answering, visual perception, understanding, and reasoning. Over the past few years, significant efforts have been made to examine MLLMs from multiple perspectives. This paper presents a comprehensive review of \textbf{180 benchmarks} and evaluation for MLLMs, focusing on (1)perception and understanding, (2)cognition and reasoning, (3)specific domains, (4)key capabilities, and (5)other modalities. Finally, we discuss the limitations of the current evaluation methods for MLLMs and explore promising future directions. Our key argument is that evaluation should be regarded as a crucial discipline to better support the development of MLLMs. For more details, please visit our GitHub repository: https://github.com/swordlidev/Evaluation-Multimodal-LLMs-Survey.
URLs: https://github.com/swordlidev/Evaluation-Multimodal-LLMs-Survey.
Authors: Wenwen Zhuang, Xin Huang, Xiantao Zhang, Jin Zeng
Abstract: Multimodal Large Language Models (MLLMs) excel in solving text-based mathematical problems, but they struggle with mathematical diagrams since they are primarily trained on natural scene images. For humans, visual aids generally enhance problem-solving, but MLLMs perform worse as information shifts from textual to visual modality. This decline is mainly due to their shortcomings in aligning images and text. To tackle aforementioned challenges, we propose Math-PUMA, a methodology focused on Progressive Upward Multimodal Alignment. This approach is designed to improve the mathematical reasoning skills of MLLMs through a three-stage training process, with the second stage being the critical alignment stage. We first enhance the language model's mathematical reasoning capabilities with extensive set of textual mathematical problems. We then construct a multimodal dataset with varying degrees of textual and visual information, creating data pairs by presenting each problem in at least two forms. By leveraging the Kullback-Leibler (KL) divergence of next-token prediction distributions to align visual and textual modalities, consistent problem-solving abilities are ensured. Finally, we utilize multimodal instruction tuning for MLLMs with high-quality multimodal data. Experimental results on multiple mathematical reasoning benchmarks demonstrate that the MLLMs trained with Math-PUMA surpass most open-source MLLMs. Our approach effectively narrows the performance gap for problems presented in different modalities.
Authors: Peiming Guo, Sinuo Liu, Yanzhao Zhang, Dingkun Long, Pengjun Xie, Meishan Zhang, Min Zhang
Abstract: Photo-Sharing Multi-modal dialogue generation requires a dialogue agent not only to generate text responses but also to share photos at the proper moment. Using image text caption as the bridge, a pipeline model integrates an image caption model, a text generation model, and an image generation model to handle this complex multi-modal task. However, representing the images with text captions may loss important visual details and information and cause error propagation in the complex dialogue system. Besides, the pipeline model isolates the three models separately because discrete image text captions hinder end-to-end gradient propagation. We propose the first end-to-end model for photo-sharing multi-modal dialogue generation, which integrates an image perceptron and an image generator with a large language model. The large language model employs the Q-Former to perceive visual images in the input end. For image generation in the output end, we propose a dynamic vocabulary transformation matrix and use straight-through and gumbel-softmax techniques to align the large language model and stable diffusion model and achieve end-to-end gradient propagation. We perform experiments on PhotoChat and DialogCC datasets to evaluate our end-to-end model. Compared with pipeline models, the end-to-end model gains state-of-the-art performances on various metrics of text and image generation. More analysis experiments also verify the effectiveness of the end-to-end model for photo-sharing multi-modal dialogue generation.
Authors: Kyle Moore, Jesse Roberts, Thao Pham, Douglas Fisher
Abstract: Language models are known to absorb biases from their training data, leading to predictions driven by statistical regularities rather than semantic relevance. We investigate the impact of these biases on answer choice preferences in the Massive Multi-Task Language Understanding (MMLU) task. Our findings reveal that differences in learned regularities across answer options are predictive of model preferences and mirror human test-taking strategies. To address this issue, we introduce two novel methods: Counterfactual Prompting with Chain of Thought (CoT) and Counterfactual Prompting with Agnostically Primed CoT (APriCoT). We demonstrate that while Counterfactual Prompting with CoT alone is insufficient to mitigate bias, our novel Primed Counterfactual Prompting with CoT approach effectively reduces the influence of base-rate probabilities while improving overall accuracy. Our results suggest that mitigating bias requires a "System-2" like process and that CoT reasoning is susceptible to confirmation bias under some prompting methodologies. Our contributions offer practical solutions for developing more robust and fair language models.
Authors: Do Xuan Long, Hai Nguyen Ngoc, Tiviatis Sim, Hieu Dao, Shafiq Joty, Kenji Kawaguchi, Nancy F. Chen, Min-Yen Kan
Abstract: We present the first systematic evaluation examining format bias in performance of large language models (LLMs). Our approach distinguishes between two categories of an evaluation metric under format constraints to reliably and accurately assess performance: one measures performance when format constraints are adhered to, while the other evaluates performance regardless of constraint adherence. We then define a metric for measuring the format bias of LLMs and establish effective strategies to reduce it. Subsequently, we present our empirical format bias evaluation spanning four commonly used categories -- multiple-choice question-answer, wrapping, list, and mapping -- covering 15 widely-used formats. Our evaluation on eight generation tasks uncovers significant format bias across state-of-the-art LLMs. We further discover that improving the format-instruction following capabilities of LLMs across formats potentially reduces format bias. Based on our evaluation findings, we study prompting and fine-tuning with synthesized format data techniques to mitigate format bias. Our methods successfully reduce the variance in ChatGPT's performance among wrapping formats from 235.33 to 0.71 (%$^2$).
Authors: Wenjie Fu, Huandong Wang, Chen Gao, Guanghua Liu, Yong Li, Tao Jiang
Abstract: The increasing parameters and expansive dataset of large language models (LLMs) highlight the urgent demand for a technical solution to audit the underlying privacy risks and copyright issues associated with LLMs. Existing studies have partially addressed this need through an exploration of the pre-training data detection problem, which is an instance of a membership inference attack (MIA). This problem involves determining whether a given piece of text has been used during the pre-training phase of the target LLM. Although existing methods have designed various sophisticated MIA score functions to achieve considerable detection performance in pre-trained LLMs, how to achieve high-confidence detection and how to perform MIA on aligned LLMs remain challenging. In this paper, we propose MIA-Tuner, a novel instruction-based MIA method, which instructs LLMs themselves to serve as a more precise pre-training data detector internally, rather than design an external MIA score function. Furthermore, we design two instruction-based safeguards to respectively mitigate the privacy risks brought by the existing methods and MIA-Tuner. To comprehensively evaluate the most recent state-of-the-art LLMs, we collect a more up-to-date MIA benchmark dataset, named WIKIMIA-24, to replace the widely adopted benchmark WIKIMIA. We conduct extensive experiments across various aligned and unaligned LLMs over the two benchmark datasets. The results demonstrate that MIA-Tuner increases the AUC of MIAs from 0.7 to a significantly high level of 0.9.
Authors: Samee Arif, Sualeha Farid, Abdul Hameed Azeemi, Awais Athar, Agha Ali Raza
Abstract: This paper presents and evaluates multi-agent workflows for synthetic Preference Optimization (PO) dataset generation. PO dataset generation requires two modules: (1) response evaluation, and (2) response generation. In the response evaluation module, the responses from Large Language Models (LLMs) are evaluated and ranked - a task typically carried out by human annotators that we automate using LLMs. We assess the response evaluation module in a 2 step process. In step 1, we assess LLMs as evaluators using three distinct prompting strategies. In step 2, we apply the winning prompting strategy to compare the performance of LLM-as-a-Judge, LLMs-as-a-Jury, and LLM Debate. In each step, we use inter-rater agreement using Cohen's Kappa between human annotators and LLMs. For the response generation module, we compare different configurations for the LLM Feedback Loop using the identified LLM evaluator configuration. We use the win rate (the fraction of times a generation framework is selected as the best by an LLM evaluator) to determine the best multi-agent configuration for generation. After identifying the best configurations for both modules, we use models from the GPT, Gemma, and Llama families to generate our PO datasets using the above pipeline. We generate two types of PO datasets, one to improve the generation capabilities of individual LLM and the other to improve the multi-agent workflow. Our evaluation shows that GPT-4o-as-a-Judge is more consistent across datasets when the candidate responses do not include responses from the GPT family. Additionally, we find that the LLM Feedback Loop, with Llama as the generator and Gemma as the reviewer, achieves a notable 71.8% and 73.8% win rate over single-agent Llama and Gemma, respectively.
Authors: Hongcheng Liu, Yusheng Liao, Siqv Ou, Yuhao Wang, Heyang Liu, Yanfeng Wang, Yu Wang
Abstract: The application of the Multi-modal Large Language Models (MLLMs) in medical clinical scenarios remains underexplored. Previous benchmarks only focus on the capacity of the MLLMs in medical visual question-answering (VQA) or report generation and fail to assess the performance of the MLLMs on complex clinical multi-modal tasks. In this paper, we propose a novel Medical Personalized Multi-modal Consultation (Med-PMC) paradigm to evaluate the clinical capacity of the MLLMs. Med-PMC builds a simulated clinical environment where the MLLMs are required to interact with a patient simulator to complete the multi-modal information-gathering and decision-making task. Specifically, the patient simulator is decorated with personalized actors to simulate diverse patients in real scenarios. We conduct extensive experiments to access 12 types of MLLMs, providing a comprehensive view of the MLLMs' clinical performance. We found that current MLLMs fail to gather multimodal information and show potential bias in the decision-making task when consulted with the personalized patient simulators. Further analysis demonstrates the effectiveness of Med-PMC, showing the potential to guide the development of robust and reliable clinical MLLMs. Code and data are available at https://github.com/LiuHC0428/Med-PMC.
Authors: Lyberius Ennio F. Taruc, Arvin R. De La Cruz
Abstract: Student extracurricular activities play an important role in enriching the students' educational experiences. With the increasing popularity of Machine Learning and Natural Language Processing, it becomes a logical step that incorporating ML-NLP in improving extracurricular activities is a potential focus of study in Artificial Intelligence (AI). This research study aims to develop a machine learning workflow that will quantify the effectiveness of student-organized activities based on student emotional responses using sentiment analysis. The study uses the Bidirectional Encoder Representations from Transformers (BERT) Large Language Model (LLM) called via the pysentimiento toolkit, as a Transformer pipeline in Hugging Face. A sample data set from Organization C, a Recognized Student Organization (RSO) of a higher educational institute in the Philippines, College X, was used to develop the workflow. The workflow consisted of data preprocessing, key feature selection, LLM feature processing, and score aggregation, resulting in an Event Score for each data set. The results show that the BERT LLM can also be used effectively in analyzing sentiment beyond product reviews and post comments. For the student affairs offices of educational institutions, this study can provide a practical example of how NLP can be applied to real-world scenarios, showcasing the potential impact of data-driven decision making.
Authors: Xianzhen Luo, Yixuan Wang, Qingfu Zhu, Zhiming Zhang, Xuanyu Zhang, Qing Yang, Dongliang Xu, Wanxiang Che
Abstract: The rapid growth in the parameters of large language models (LLMs) has made inference latency a fundamental bottleneck, limiting broader application of LLMs. Speculative decoding represents a lossless approach to accelerate inference through a guess-and-verify paradigm, leveraging the parallel capabilities of modern hardware. Some speculative decoding methods rely on additional structures to guess draft tokens, such as small models or parameter-efficient architectures, which need extra training before use. Alternatively, retrieval-based train-free techniques build libraries from pre-existing corpora or by n-gram generation. However, they face challenges like large storage requirements, time-consuming retrieval, and limited adaptability. Observing that candidate tokens generated during the decoding process are likely to reoccur in future sequences, we propose Token Recycling. This approach stores candidate tokens in an adjacency matrix and employs a breadth-first search (BFS)-like algorithm on the matrix to construct a draft tree. The tree is then validated through tree attention. New candidate tokens from the decoding process are then used to update the matrix. Token Recycling requires \textless2MB of additional storage and achieves approximately 2x speedup across all sizes of LLMs. It significantly outperforms existing train-free methods by 30\% and even a training method by 25\%. It can be directly applied to any existing LLMs and tasks without the need for adaptation.
Authors: Yongkang Liu, Feng Shi, Daling Wang, Yifei Zhang, Hinrich Sch\"utze
Abstract: Although large language models(LLMs) show amazing capabilities, among various exciting applications discovered for LLMs fall short in other low-resource languages. Besides, most existing methods depend on large-scale dialogue corpora and thus building systems for dialogue generation in a zero-shot scenario remains a considerable challenge. To address this challenge, we propose a novel end-to-end zero-shot dialogue generation model ChatZero based on cross-lingual code-switching method. First, we construct code-switching language and pseudo-target language with placeholders. Then for cross-lingual semantic transfer, we employ unsupervised contrastive learning to minimize the semantics gap of the source language, code-switching language, and pseudo-target language that are mutually positive examples in the high dimensional semantic space. Experiments on the multilingual DailyDialog and DSTC7-AVSD datasets demonstrate that ChatZero can achieve more than 90\% of the original performance under the zero-shot case compared to supervised learning, and achieve state-of-the-art performance compared with other baselines.
Authors: Dingwei Chen, Feiteng Fang, Shiwen Ni, Feng Liang, Ruifeng Xu, Min Yang, Chengming Li
Abstract: Large Language Models (LLMs) have demonstrated exceptional performance across various natural language processing tasks, yet they occasionally tend to yield content that factually inaccurate or discordant with the expected output, a phenomenon empirically referred to as "hallucination". To tackle this issue, recent works have investigated contrastive decoding between the original model and an amateur model with induced hallucination, which has shown promising results. Nonetheless, this method may undermine the output distribution of the original LLM caused by its coarse contrast and simplistic subtraction operation, potentially leading to errors in certain cases. In this paper, we introduce a novel contrastive decoding framework termed LOL (LOwer Layer Matters). Our approach involves concatenating the contrastive decoding of both the final and lower layers between the original model and the amateur model, thereby achieving multi-layer fusion to aid in the mitigation of hallucination. Additionally, we incorporate a truthfulness refocused module that leverages contextual guidance to enhance factual encoding, further capturing truthfulness during contrastive decoding. Extensive experiments conducted on two publicly available datasets illustrate that our proposed LOL framework can substantially alleviate hallucination while surpassing existing baselines in most cases. Compared with the best baseline, we improve by average 4.5 points on all metrics of TruthfulQA. The source code is coming soon.
Authors: Dingzirui Wang, Longxu Dou, Xuanliang Zhang, Qingfu Zhu, Wanxiang Che
Abstract: Text-to-SQL is an important task that helps people obtain information from databases by automatically generating SQL queries. Considering the brilliant performance, approaches based on Large Language Models (LLMs) become the mainstream for text-to-SQL. Among these approaches, automated correction is an effective approach that further enhances performance by correcting the mistakes in the generated results. The existing correction methods require LLMs to directly correct with generated SQL, while previous research shows that LLMs do not know how to detect mistakes, leading to poor performance. Therefore, in this paper, we propose to employ the decomposed correction to enhance text-to-SQL performance. We first demonstrate that decomposed correction outperforms direct correction since detecting and fixing mistakes with the results of the decomposed sub-tasks is easier than with SQL. Based on this analysis, we introduce Decomposed Automation Correction (DAC), which corrects SQL by decomposing text-to-SQL into entity linking and skeleton parsing. DAC first generates the entity and skeleton corresponding to the question and then compares the differences between the initial SQL and the generated entities and skeleton as feedback for correction. Experimental results show that our method improves performance by $3.7\%$ on average of Spider, Bird, and KaggleDBQA compared with the baseline method, demonstrating the effectiveness of DAC.
Authors: Chenming Tang, Zhixiang Wang, Yunfang Wu
Abstract: With the help of in-context learning (ICL), large language models (LLMs) have achieved impressive performance across various tasks. However, the function of descriptive instructions during ICL remains under-explored. In this work, we propose an ensemble prompt framework to describe the selection criteria of multiple in-context examples, and preliminary experiments on machine translation (MT) across six translation directions confirm that this framework boosts ICL perfromance. But to our surprise, LLMs might not necessarily care what the descriptions actually say, and the performance gain is primarily caused by the ensemble format, since the framework could lead to improvement even with random descriptive nouns. We further apply this new ensemble prompt on a range of commonsense, math, logical reasoning and hallucination tasks with three LLMs and achieve promising results, suggesting again that designing a proper prompt format would be much more effective and efficient than paying effort into specific descriptions. Our code will be publicly available once this paper is published.
Authors: Chenwei Wan, Matthieu Labeau, Chlo\'e Clavel
Abstract: Designing emotionally intelligent conversational systems to provide comfort and advice to people experiencing distress is a compelling area of research. Previous efforts have focused on developing modular dialogue systems that treat socio-emotional strategy prediction as an auxiliary task and generate strategy-conditioned responses with customized decoders. Recently, with advancements in large language models (LLMs), end-to-end dialogue agents without explicit socio-emotional strategy prediction steps have become prevalent. However, despite their excellence in language generation, recent studies show that LLMs' inherent preference bias towards certain socio-emotional strategies hinders the delivery of high-quality emotional support. To address this challenge, we propose decoupling strategy prediction from language generation, and introduce a novel dialogue strategy predictor, EmoDynamiX, which models the discourse dynamics between user emotions and system strategies using a heterogeneous graph. Additionally, we make use of the Emotion Recognition in Conversations (ERC) task and design a flexible mixed-emotion module to capture fine-grained emotional states of the user. Experimental results on two ESC datasets show EmoDynamiX outperforms previous state-of-the-art methods with a significant margin.
Authors: Abdullah Al Imran, Md Farhan Ishmam
Abstract: For many years, transformer-based pre-trained models with Multi-layer Perceptron (MLP) heads have been the standard for text classification tasks. However, the fixed non-linear functions employed by MLPs often fall short of capturing the intricacies of the contextualized embeddings produced by pre-trained encoders. Furthermore, MLPs usually require a significant number of training parameters, which can be computationally expensive. In this work, we introduce FourierKAN (FR-KAN), a variant of the promising MLP alternative called Kolmogorov-Arnold Networks (KANs), as classification heads for transformer-based encoders. Our studies reveal an average increase of 10% in accuracy and 11% in F1-score when incorporating FR-KAN heads instead of traditional MLP heads for several transformer-based pre-trained models across multiple text classification tasks. Beyond improving model accuracy, FR-KAN heads train faster and require fewer parameters. Our research opens new grounds for broader applications of KAN across several Natural Language Processing (NLP) tasks.
Authors: Joanito Agili Lopo, Marina Indah Prasasti, Alma Permatasari
Abstract: In this study, we introduce CIKMar, an efficient approach to educational dialogue systems powered by the Gemma Language model. By leveraging a Dual-Encoder ranking system that incorporates both BERT and SBERT model, we have designed CIKMar to deliver highly relevant and accurate responses, even with the constraints of a smaller language model size. Our evaluation reveals that CIKMar achieves a robust recall and F1-score of 0.70 using BERTScore metrics. However, we have identified a significant challenge: the Dual-Encoder tends to prioritize theoretical responses over practical ones. These findings underscore the potential of compact and efficient models like Gemma in democratizing access to advanced educational AI systems, ensuring effective and contextually appropriate responses.
Authors: Xuanliang Zhang, Dingzirui Wang, Longxu Dou, Baoxin Wang, Dayong Wu, Qingfu Zhu, Wanxiang Che
Abstract: The table reasoning task aims to answer the question according to the given table. Currently, using Large Language Models (LLMs) is the predominant method for table reasoning. Most existing methods employ a fixed tabular format to represent the table, which could limit the performance. Given that each instance requires different capabilities and models possess varying abilities, we assert that different instances and models suit different tabular formats. We prove the aforementioned claim through quantitative analysis of experimental results, where different instances and models achieve different performances using various tabular formats. Building on this discussion, we propose FLEXTAF-Single and FLEXTAF-Vote to enhance table reasoning performance by employing flexible tabular formats. Specifically, (i) FLEXTAF-Single trains a classifier to predict the most suitable tabular format based on the instance and the LLM. (ii) FLEXTAF-Vote integrates the results across different formats. Our experiments on WikiTableQuestions and TabFact reveal significant improvements, with average gains of 2.3% and 4.8% compared to the best performance achieved using a fixed tabular format with greedy decoding and self-consistency decoding, thereby validating the effectiveness of our methods.
Authors: Mohammad Amin Abbasi, Farnaz Sadat Mirnezami, Hassan Naderi
Abstract: This paper explores the intersection of psychology and artificial intelligence through the development and evaluation of specialized Large Language Models (LLMs). We introduce PsychoLex, a suite of resources designed to enhance LLMs' proficiency in psychological tasks in both Persian and English. Key contributions include the PsychoLexQA dataset for instructional content and the PsychoLexEval dataset for rigorous evaluation of LLMs in complex psychological scenarios. Additionally, we present the PsychoLexLLaMA model, optimized specifically for psychological applications, demonstrating superior performance compared to general-purpose models. The findings underscore the potential of tailored LLMs for advancing psychological research and applications, while also highlighting areas for further refinement. This research offers a foundational step towards integrating LLMs into specialized psychological domains, with implications for future advancements in AI-driven psychological practice.
Authors: Sumanth Prabhu
Abstract: Self-ensembling techniques with diverse reasoning paths such as Self-Consistency have demonstrated remarkable gains in accuracy for Large Language Models (LLMs). However, such techniques depend on the availability of an accurate answer extraction process to aggregate across multiple outputs. Moreover, they acquire higher inference cost, in comparison to Greedy Decoding, due to generation of relatively higher number of output tokens. Research has shown that the free form text outputs from Self-Consistency can be aggregated reliably using LLMs to produce the final output. Additionally, recent advancements in LLM inference have demonstrated that usage of diverse exemplars in prompts have the ability to induce diversity in the LLM outputs. Such proven techniques can be easily extended to self-ensembling based approaches to achieve enhanced results in text generation. In this paper, we introduce PEDAL (Prompts based on Exemplar Diversity Aggregated using LLMs), a hybrid self-ensembling approach, that combines the strengths of diverse exemplar based prompts and LLM based aggregation to achieve improvement in overall performance. On the publicly available SVAMP and ARC datasets, our experiments reveal that PEDAL can achieve better accuracy than Greedy Decoding based strategies with lower inference cost compared to Self Consistency based approaches.
Authors: Vibhor Agarwal, Yulong Pei, Salwa Alamir, Xiaomo Liu
Abstract: Large Language Models (LLMs) have shown promising potentials in program generation and no-code automation. However, LLMs are prone to generate hallucinations, i.e., they generate text which sounds plausible but is incorrect. Although there has been a recent surge in research on LLM hallucinations for text generation, similar hallucination phenomenon can happen in code generation. Sometimes the generated code can have syntactical or logical errors as well as more advanced issues like security vulnerabilities, memory leaks, etc. Given the wide adaptation of LLMs to enhance efficiency in code generation and development in general, it becomes imperative to investigate hallucinations in code generation. To the best of our knowledge, this is the first attempt at studying hallucinations in the code generated by LLMs. We start by introducing the code hallucination definition and a comprehensive taxonomy of code hallucination types. We propose the first benchmark CodeMirage dataset for code hallucinations. The benchmark contains 1,137 GPT-3.5 generated hallucinated code snippets for Python programming problems from two base datasets - HumanEval and MBPP. We then propose the methodology for code hallucination detection and experiment with open source LLMs such as CodeLLaMA as well as OpenAI's GPT-3.5 and GPT-4 models using one-shot prompt. We find that GPT-4 performs the best on HumanEval dataset and gives comparable results to the fine-tuned CodeBERT baseline on MBPP dataset. Towards the end, we discuss various mitigation strategies for code hallucinations and conclude our work.
Authors: Nastaran Bassamzadeh, Chhaya Methani
Abstract: Planning in code is considered a more reliable approach for many orchestration tasks. This is because code is more tractable than steps generated via Natural Language and make it easy to support more complex sequences by abstracting deterministic logic into functions. It also allows spotting issues with incorrect function names with the help of parsing checks that can be run on code. Progress in Code Generation methodologies, however, remains limited to general-purpose languages like C, C++, and Python. LLMs continue to face challenges with custom function names in Domain Specific Languages or DSLs, leading to higher hallucination rates and syntax errors. This is more common for custom function names, that are typically part of the plan. Moreover, keeping LLMs up-to-date with newer function names is an issue. This poses a challenge for scenarios like task planning over a large number of APIs, since the plan is represented as a DSL having custom API names. In this paper, we focus on workflow automation in RPA (Robotic Process Automation) domain as a special case of task planning. We present optimizations for using Retrieval Augmented Generation (or RAG) with LLMs for DSL generation along with an ablation study comparing these strategies with a fine-tuned model. Our results showed that the fine-tuned model scored the best on code similarity metric. However, with our optimizations, RAG approach is able to match the quality for in-domain API names in the test set. Additionally, it offers significant advantage for out-of-domain or unseen API names, outperforming Fine-Tuned model on similarity metric by 7 pts.
Authors: Daniele Rege Cambrin, Gabriele Scaffidi Militone, Luca Colomba, Giovanni Malnati, Daniele Apiletti, Paolo Garza
Abstract: Designing effective game tutorials is crucial for a smooth learning curve for new players, especially in games with many rules and complex core mechanics. Evaluating the effectiveness of these tutorials usually requires multiple iterations with testers who have no prior knowledge of the game. Recent Vision-Language Models (VLMs) have demonstrated significant capabilities in understanding and interpreting visual content. VLMs can analyze images, provide detailed insights, and answer questions about their content. They can recognize objects, actions, and contexts in visual data, making them valuable tools for various applications, including automated game testing. In this work, we propose an automated game-testing solution to evaluate the quality of game tutorials. Our approach leverages VLMs to analyze frames from video game tutorials, answer relevant questions to simulate human perception, and provide feedback. This feedback is compared with expected results to identify confusing or problematic scenes and highlight potential errors for developers. In addition, we publish complete tutorial videos and annotated frames from different game versions used in our tests. This solution reduces the need for extensive manual testing, especially by speeding up and simplifying the initial development stages of the tutorial to improve the final game experience.
Authors: Zhengyuan Zhu, Daniel Lee, Hong Zhang, Sai Sree Harsha, Loic Feujio, Akash Maharaj, Yunyao Li
Abstract: Recent advancements in retrieval-augmented generation (RAG) have demonstrated impressive performance in the question-answering (QA) task. However, most previous works predominantly focus on text-based answers. While some studies address multimodal data, they still fall short in generating comprehensive multimodal answers, particularly for explaining concepts or providing step-by-step tutorials on how to accomplish specific goals. This capability is especially valuable for applications such as enterprise chatbots and settings such as customer service and educational systems, where the answers are sourced from multimodal data. In this paper, we introduce a simple and effective framework named MuRAR (Multimodal Retrieval and Answer Refinement). MuRAR enhances text-based answers by retrieving relevant multimodal data and refining the responses to create coherent multimodal answers. This framework can be easily extended to support multimodal answers in enterprise chatbots with minimal modifications. Human evaluation results indicate that multimodal answers generated by MuRAR are more useful and readable compared to plain text answers.
Authors: Zhongzhou Liu, Hao Zhang, Kuicai Dong, Yuan Fang
Abstract: Despite the success of conventional collaborative filtering (CF) approaches for recommendation systems, they exhibit limitations in leveraging semantic knowledge within the textual attributes of users and items. Recent focus on the application of large language models for recommendation (LLM4Rec) has highlighted their capability for effective semantic knowledge capture. However, these methods often overlook the collaborative signals in user behaviors. Some simply instruct-tune a language model, while others directly inject the embeddings of a CF-based model, lacking a synergistic fusion of different modalities. To address these issues, we propose a framework of Collaborative Cross-modal Fusion with Large Language Models, termed CCF-LLM, for recommendation. In this framework, we translate the user-item interactions into a hybrid prompt to encode both semantic knowledge and collaborative signals, and then employ an attentive cross-modal fusion strategy to effectively fuse latent embeddings of both modalities. Extensive experiments demonstrate that CCF-LLM outperforms existing methods by effectively utilizing semantic and collaborative signals in the LLM4Rec context.
Authors: Jonathan Ben-Naim, Victor David, Anthony Hunter
Abstract: Argument mining is natural language processing technology aimed at identifying arguments in text. Furthermore, the approach is being developed to identify the premises and claims of those arguments, and to identify the relationships between arguments including support and attack relationships. In this paper, we assume that an argument map contains the premises and claims of arguments, and support and attack relationships between them, that have been identified by argument mining. So from a piece of text, we assume an argument map is obtained automatically by natural language processing. However, to understand and to automatically analyse that argument map, it would be desirable to instantiate that argument map with logical arguments. Once we have the logical representation of the arguments in an argument map, we can use automated reasoning to analyze the argumentation (e.g. check consistency of premises, check validity of claims, and check the labelling on each arc corresponds with thw logical arguments). We address this need by using classical logic for representing the explicit information in the text, and using default logic for representing the implicit information in the text. In order to investigate our proposal, we consider some specific options for instantiation.
Authors: Yuqi Ye, Wei Gao
Abstract: The key to effective point cloud compression is to obtain a robust context model consistent with complex 3D data structures. Recently, the advancement of large language models (LLMs) has highlighted their capabilities not only as powerful generators for in-context learning and generation but also as effective compressors. These dual attributes of LLMs make them particularly well-suited to meet the demands of data compression. Therefore, this paper explores the potential of using LLM for compression tasks, focusing on lossless point cloud geometry compression (PCGC) experiments. However, applying LLM directly to PCGC tasks presents some significant challenges, i.e., LLM does not understand the structure of the point cloud well, and it is a difficult task to fill the gap between text and point cloud through text description, especially for large complicated and small shapeless point clouds. To address these problems, we introduce a novel architecture, namely the Large Language Model-based Point Cloud Geometry Compression (LLM-PCGC) method, using LLM to compress point cloud geometry information without any text description or aligning operation. By utilizing different adaptation techniques for cross-modality representation alignment and semantic consistency, including clustering, K-tree, token mapping invariance, and Low Rank Adaptation (LoRA), the proposed method can translate LLM to a compressor/generator for point cloud. To the best of our knowledge, this is the first structure to employ LLM as a compressor for point cloud data. Experiments demonstrate that the LLM-PCGC outperforms the other existing methods significantly, by achieving -40.213% bit rate reduction compared to the reference software of MPEG Geometry-based Point Cloud Compression (G-PCC) standard, and by achieving -2.267% bit rate reduction compared to the state-of-the-art learning-based method.
Authors: Mhd Modar Halimeh, Matteo Torcoli, Emanu\"el Habets
Abstract: Dialogue separation involves isolating a dialogue signal from a mixture, such as a movie or a TV program. This can be a necessary step to enable dialogue enhancement for broadcast-related applications. In this paper, ConcateNet for dialogue separation is proposed, which is based on a novel approach for processing local and global features aimed at better generalization for out-of-domain signals. ConcateNet is trained using a noise reduction-focused, publicly available dataset and evaluated using three datasets: two noise reduction-focused datasets (in-domain), which show competitive performance for ConcateNet, and a broadcast-focused dataset (out-of-domain), which verifies the better generalization performance for the proposed architecture compared to considered state-of-the-art noise-reduction methods.
Authors: Bhuvanashree Murugadoss, Christian Poelitz, Ian Drosos, Vu Le, Nick McKenna, Carina Suzana Negreanu, Chris Parnin, Advait Sarkar
Abstract: LLMs-as-a-judge is a recently popularized method which replaces human judgements in task evaluation (Zheng et al. 2024) with automatic evaluation using LLMs. Due to widespread use of RLHF (Reinforcement Learning from Human Feedback), state-of-the-art LLMs like GPT4 and Llama3 are expected to have strong alignment with human preferences when prompted for a quality judgement, such as the coherence of a text. While this seems beneficial, it is not clear whether the assessments by an LLM-as-a-judge constitute only an evaluation based on the instructions in the prompts, or reflect its preference for high-quality data similar to its fine-tune data. To investigate how much influence prompting the LLMs-as-a-judge has on the alignment of AI judgements to human judgements, we analyze prompts with increasing levels of instructions about the target quality of an evaluation, for several LLMs-as-a-judge. Further, we compare to a prompt-free method using model perplexity as a quality measure instead. We aggregate a taxonomy of quality criteria commonly used across state-of-the-art evaluations with LLMs and provide this as a rigorous benchmark of models as judges. Overall, we show that the LLMs-as-a-judge benefit only little from highly detailed instructions in prompts and that perplexity can sometimes align better with human judgements than prompting, especially on textual quality.
Authors: Le Xue, Manli Shu, Anas Awadalla, Jun Wang, An Yan, Senthil Purushwalkam, Honglu Zhou, Viraj Prabhu, Yutong Dai, Michael S Ryoo, Shrikant Kendre, Jieyu Zhang, Can Qin, Shu Zhang, Chia-Chih Chen, Ning Yu, Juntao Tan, Tulika Manoj Awalgaonkar, Shelby Heinecke, Huan Wang, Yejin Choi, Ludwig Schmidt, Zeyuan Chen, Silvio Savarese, Juan Carlos Niebles, Caiming Xiong, Ran Xu
Abstract: This report introduces xGen-MM (also known as BLIP-3), a framework for developing Large Multimodal Models (LMMs). The framework comprises meticulously curated datasets, a training recipe, model architectures, and a resulting suite of LMMs. xGen-MM, short for xGen-MultiModal, expands the Salesforce xGen initiative on foundation AI models. Our models undergo rigorous evaluation across a range of tasks, including both single and multi-image benchmarks. Our pre-trained base model exhibits strong in-context learning capabilities and the instruction-tuned model demonstrates competitive performance among open-source LMMs with similar model sizes. In addition, we introduce a safety-tuned model with DPO, aiming to mitigate harmful behaviors such as hallucinations and improve safety. We open-source our models, curated large-scale datasets, and our fine-tuning codebase to facilitate further advancements in LMM research. Associated resources will be available on our project page above.
Authors: Yucheng Shi, Shaochen Xu, Tianze Yang, Zhengliang Liu, Tianming Liu, Quanzheng Li, Xiang Li, Ninghao Liu
Abstract: Large Language Models (LLMs), although powerful in general domains, often perform poorly on domain-specific tasks such as medical question answering (QA). In addition, LLMs tend to function as "black-boxes", making it challenging to modify their behavior. To address the problem, our work employs a transparent process of retrieval augmented generation (RAG), aiming to improve LLM responses without the need for fine-tuning or retraining. Specifically, we propose a comprehensive retrieval strategy to extract medical facts from an external knowledge base, and then inject them into the LLM's query prompt. Focusing on medical QA, we evaluate the impact of different retrieval models and the number of facts on LLM performance using the MedQA-SMILE dataset. Notably, our retrieval-augmented Vicuna-7B model exhibited an accuracy improvement from 44.46% to 48.54%. This work underscores the potential of RAG to enhance LLM performance, offering a practical approach to mitigate the challenges posed by black-box LLMs.
Authors: Eric Zelikman, Eliana Lorch, Lester Mackey, Adam Tauman Kalai
Abstract: Several recent advances in AI systems solve problems by providing a "scaffolding" program that structures multiple calls to language models (LMs) to generate better outputs. A scaffolding program is written in a programming language such as Python. In this work, we use a language-model-infused scaffolding program to improve itself. We start with a seed "improver" that improves an input program according to a given utility function by querying an LM several times and returning the best solution. We then run this seed improver to improve itself. Across a small set of downstream tasks, the resulting improved improver generates programs with significantly better performance than its seed improver. A variety of self-improvement strategies are proposed by the language model, including beam search, genetic algorithms, and simulated annealing. Since the language models themselves are not altered, this is not full recursive self-improvement. Nonetheless, it demonstrates that a modern language model, GPT-4 in our experiments, is capable of writing code that can call itself to improve itself. We consider concerns around the development of self-improving technologies and evaluate the frequency with which the generated code bypasses a sandbox.
Authors: Lyle Muller, Patricia S. Churchland, Terrence J. Sejnowski
Abstract: The capabilities of transformer networks such as ChatGPT and other Large Language Models (LLMs) have captured the world's attention. The crucial computational mechanism underlying their performance relies on transforming a complete input sequence - for example, all the words in a sentence - into a long "encoding vector" that allows transformers to learn long-range temporal dependencies in naturalistic sequences. Specifically, "self-attention" applied to this encoding vector enhances temporal context in transformers by computing associations between pairs of words in the input sequence. We suggest that waves of neural activity traveling across single cortical areas or multiple regions at the whole-brain scale could implement a similar encoding principle. By encapsulating recent input history into a single spatial pattern at each moment in time, cortical waves may enable temporal context to be extracted from sequences of sensory inputs, the same computational principle used in transformers.
Authors: Haitao Mao, Guangliang Liu, Yao Ma, Rongrong Wang, Kristen Johnson, Jiliang Tang
Abstract: In-Context Learning (ICL) empowers Large Language Models (LLMs) with the capacity to learn in context, achieving downstream generalization without gradient updates but with a few in-context examples. Despite the encouraging empirical success, the underlying mechanism of ICL remains unclear, and existing research offers various viewpoints of understanding. These studies propose intuition-driven and ad-hoc technical solutions for interpreting ICL, illustrating an ambiguous road map. In this paper, we leverage a data generation perspective to reinterpret recent efforts and demonstrate the potential broader usage of popular technical solutions, approaching a systematic angle. For a conceptual definition, we rigorously adopt the terms of skill learning and skill recognition. The difference between them is skill learning can learn new data generation functions from in-context data. We also provide a comprehensive study on the merits and weaknesses of different solutions, and highlight the uniformity among them given the perspective of data generation, establishing a technical foundation for future research to incorporate the strengths of different lines of research.
Authors: Xuanliang Zhang, Dingzirui Wang, Longxu Dou, Qingfu Zhu, Wanxiang Che
Abstract: Open-domain text-to-SQL is an important task that retrieves question-relevant tables from massive databases and then generates SQL. However, existing retrieval methods that retrieve in a single hop do not pay attention to the text-to-SQL challenge of schema linking, which is aligning the entities in the question with table entities, reflected in two aspects: similar irrelevant entity and domain mismatch entity. Therefore, we propose our method, the multi-hop table retrieval with rewrite and beam search (Murre). To reduce the effect of the similar irrelevant entity, our method focuses on unretrieved entities at each hop and considers the low-ranked tables by beam search. To alleviate the limitation of domain mismatch entity, Murre rewrites the question based on retrieved tables in multiple hops, decreasing the domain gap with relevant tables. We conduct experiments on SpiderUnion and BirdUnion+, reaching new state-of-the-art results with an average improvement of 6.38%.
Authors: Junjie Ye, Sixian Li, Guanyu Li, Caishuang Huang, Songyang Gao, Yilong Wu, Qi Zhang, Tao Gui, Xuanjing Huang
Abstract: Tool learning is widely acknowledged as a foundational approach or deploying large language models (LLMs) in real-world scenarios. While current research primarily emphasizes leveraging tools to augment LLMs, it frequently neglects emerging safety considerations tied to their application. To fill this gap, we present *ToolSword*, a comprehensive framework dedicated to meticulously investigating safety issues linked to LLMs in tool learning. Specifically, ToolSword delineates six safety scenarios for LLMs in tool learning, encompassing **malicious queries** and **jailbreak attacks** in the input stage, **noisy misdirection** and **risky cues** in the execution stage, and **harmful feedback** and **error conflicts** in the output stage. Experiments conducted on 11 open-source and closed-source LLMs reveal enduring safety challenges in tool learning, such as handling harmful queries, employing risky tools, and delivering detrimental feedback, which even GPT-4 is susceptible to. Moreover, we conduct further studies with the aim of fostering research on tool learning safety. The data is released in https://github.com/Junjie-Ye/ToolSword.
Authors: Jerry Ngo, Yoon Kim
Abstract: This work explores whether language models encode meaningfully grounded representations of sounds of objects. We learn a linear probe that retrieves the correct text representation of an object given a snippet of audio related to that object, where the sound representation is given by a pretrained audio model. This probe is trained via a contrastive loss that pushes the language representations and sound representations of an object to be close to one another. After training, the probe is tested on its ability to generalize to objects that were not seen during training. Across different language models and audio models, we find that the probe generalization is above chance in many cases, indicating that despite being trained only on raw text, language models encode grounded knowledge of sounds for some objects.
Authors: Xidong Wang, Nuo Chen, Junyin Chen, Yidong Wang, Guorui Zhen, Yan Hu, Xiangbo Wu, Anningzhe Gao, Xiang Wan, Haizhou Li, Benyou Wang
Abstract: Despite the vast repository of global medical knowledge predominantly being in English, local languages are crucial for delivering tailored healthcare services, particularly in areas with limited medical resources. To extend the reach of medical AI advancements to a broader population, we aim to develop medical LLMs across the six most widely spoken languages, encompassing a global population of 6.1 billion. This effort culminates in the creation of the ApolloCorpora multilingual medical dataset and the XMedBench benchmark. In the multilingual medical benchmark, the released Apollo models, at various relatively-small sizes (i.e., 0.5B, 1.8B, 2B, 6B, and 7B), achieve the best performance among models of equivalent size. Especially, Apollo-7B is the state-of-the-art multilingual medical LLMs up to 70B. Additionally, these lite models could be used to improve the multi-lingual medical capabilities of larger models without fine-tuning in a proxy-tuning fashion. We will open-source training corpora, code, model weights and evaluation benchmark.
Authors: Cristiano Mesquita Garcia, Alessandro Lameiras Koerich, Alceu de Souza Britto Jr, Jean Paul Barddal
Abstract: The proliferation of textual data on the Internet presents a unique opportunity for institutions and companies to monitor public opinion about their services and products. Given the rapid generation of such data, the text stream mining setting, which handles sequentially arriving, potentially infinite text streams, is often more suitable than traditional batch learning. While pre-trained language models are commonly employed for their high-quality text vectorization capabilities in streaming contexts, they face challenges adapting to concept drift - the phenomenon where the data distribution changes over time, adversely affecting model performance. Addressing the issue of concept drift, this study explores the efficacy of seven text sampling methods designed to selectively fine-tune language models, thereby mitigating performance degradation. We precisely assess the impact of these methods on fine-tuning the SBERT model using four different loss functions. Our evaluation, focused on Macro F1-score and elapsed time, employs two text stream datasets and an incremental SVM classifier to benchmark performance. Our findings indicate that Softmax loss and Batch All Triplets loss are particularly effective for text stream classification, demonstrating that larger sample sizes generally correlate with improved macro F1-scores. Notably, our proposed WordPieceToken ratio sampling method significantly enhances performance with the identified loss functions, surpassing baseline results.
Authors: Bohao Yang, Kun Zhao, Chen Tang, Dong Liu, Liang Zhan, Chenghua Lin
Abstract: Automatic open-domain dialogue evaluation has attracted increasing attention. Trainable evaluation metrics, typically trained with true positive and randomly selected negative responses, tend to assign higher scores to responses that share greater content similarity with a given context. However, adversarial negative responses, despite possessing high content similarity with the contexts, are semantically different. Consequently, existing evaluation metrics are not robust enough to evaluate such responses, resulting in low correlations with human judgments. While recent studies have demonstrated the effectiveness of Large Language Models (LLMs) for open-domain dialogue evaluation, they still face challenges in effectively handling adversarial negative examples. In this paper, we propose an effective framework for open-domain dialogue evaluation, which combines domain-specific language models (SLMs) enhanced with Abstract Meaning Representation (AMR) knowledge with LLMs. The SLMs can explicitly incorporate AMR graph information of the dialogue through a gating mechanism for enhanced dialogue semantic representation learning. Both the evaluation result from the SLMs and the AMR graph information are incorporated into the LLM's prompt for enhanced evaluation performance. Experimental results on open-domain dialogue evaluation tasks demonstrate the superiority of our method compared to a wide range of state-of-the-art baselines, especially in discriminating adversarial negative responses. Our code and data are publicly available at https://github.com/Bernard-Yang/SIMAMR.
Authors: Xiaoshu Chen, Sihang Zhou, Ke Liang, Xinwang Liu
Abstract: Chain of thought finetuning (cot-finetuning) aims to endow small language models (SLM) with reasoning ability to improve their performance towards specific tasks by allowing them to imitate the reasoning procedure of large language models (LLM) beyond simply predicting the answers. Most existing cot-finetuning methods adopt a pre-thinking mechanism, allowing the SLM to generate a rationale before providing an answer. This mechanism enables SLM to analyze and think about complex questions, but it also makes answer correctness highly sensitive to minor errors in rationale. Therefore, we propose a robust post-thinking mechanism to generate answers before rationale. Thanks to this answer-first setting, 1) the answer can escape from the adverse effects caused by minor errors in the rationale; 2) the rationale serves as an error amplifier to the answer, which makes the SLM focus on learning hard samples; 3) the inferring efficiency can also benefit from the setting since users can stop the generation right after answers are outputted when inference is conducted. However, although the post-thinking mechanism brings many advantages and improves the overall performance of SLM on specific tasks, it may lose the ability to think about the questions and decompose complex questions into simple sub-questions compared to pre-thinking mechanism. Therefore, a plug-and-play adaptive-thinking mechanism is proposed with the aid of the soft prompt tuning to integrate the merits of the pre-thinking mechanism and post-thinking mechanism, in which a perception module is introduced to adaptively prompt SLM answer or think first based on perceiving the complexity of the questions. Extensive experiments are conducted across 12 reasoning tasks and 2 representative language models to demonstrate the effectiveness of the proposed mechanism.
Authors: Huanjun Kong
Abstract: This paper introduces a multi-stage manual annotation calibrated by the scaling law, offering a high-quality Supervised Fine-Tuning data acquisition method for environments with constrained resources like GPU poor, limited GPT access, and funding restrictions. We have preprocessed 58k authentic chat data and manually annotated 2.3k questions. After this, we conducted fine-tuning on Qwen models, ranging from 0.5B to 32B parameters. The optimal version improved 29.07 in F1 score. This confirms the viability of fine-tuning Large Language Model (LLM) for downstream Natural Language Processing (NLP) tasks. Our contributions are: 1) Created Supervised Fine-Tuning (SFT) training data in alpaca format, along with a set of Low-Rank Adaptation (LoRA) weights, and 2) Developed a method for acquiring high-quality data leveraging scaling law principle. The script, raw data with alpaca format and experiments track are open-sourced on Github (https://github.com/InternLM/HuixiangDou/tree/main/web/tools), HuggingFace (https://huggingface.co/tpoisonooo) and WandB (https://wandb.ai/tpoisonooo/huixiangdou-cr/table?nw=nwusertpoisonooo). The privacy of the data involved has been authorized by users. SFT data and license comes from ncnn contributors group.
URLs: https://github.com/InternLM/HuixiangDou/tree/main/web/tools),, https://huggingface.co/tpoisonooo), https://wandb.ai/tpoisonooo/huixiangdou-cr/table?nw=nwusertpoisonooo).
Authors: Carolina Fortuna, Vid Han\v{z}el, Bla\v{z} Bertalani\v{c}
Abstract: Domain specific digital twins, representing a digital replica of various segments of the smart grid, are foreseen as able to model, simulate, and control the respective segments. At the same time, knowledge-based digital twins, coupled with AI, may also empower humans to understand aspects of the system through natural language interaction in view of planning and policy making. This paper is the first to assess and report on the potential of Retrieval Augmented Generation (RAG) question answers related to household electrical energy measurement aspects leveraging a knowledge-based energy digital twin. Relying on the recently published electricity consumption knowledge graph that actually represents a knowledge-based digital twin, we study the capabilities of ChatGPT, Gemini and Llama in answering electricity related questions. Furthermore, we compare the answers with the ones generated through a RAG techniques that leverages an existing electricity knowledge-based digital twin. Our findings illustrate that the RAG approach not only reduces the incidence of incorrect information typically generated by LLMs but also significantly improves the quality of the output by grounding responses in verifiable data. This paper details our methodology, presents a comparative analysis of responses with and without RAG, and discusses the implications of our findings for future applications of AI in specialized sectors like energy data analysis.
Authors: Hao Yang, Yanyan Zhao, Yang Wu, Shilong Wang, Tian Zheng, Hongbo Zhang, Zongyang Ma, Wanxiang Che, Bing Qin
Abstract: Compared to traditional sentiment analysis, which only considers text, multimodal sentiment analysis needs to consider emotional signals from multimodal sources simultaneously and is therefore more consistent with the way how humans process sentiment in real-world scenarios. It involves processing emotional information from various sources such as natural language, images, videos, audio, physiological signals, etc. However, although other modalities also contain diverse emotional cues, natural language usually contains richer contextual information and therefore always occupies a crucial position in multimodal sentiment analysis. The emergence of ChatGPT has opened up immense potential for applying large language models (LLMs) to text-centric multimodal tasks. However, it is still unclear how existing LLMs can adapt better to text-centric multimodal sentiment analysis tasks. This survey aims to (1) present a comprehensive review of recent research in text-centric multimodal sentiment analysis tasks, (2) examine the potential of LLMs for text-centric multimodal sentiment analysis, outlining their approaches, advantages, and limitations, (3) summarize the application scenarios of LLM-based multimodal sentiment analysis technology, and (4) explore the challenges and potential research directions for multimodal sentiment analysis in the future.
Authors: Martin Juan Jos\'e Bucher, Marco Martini
Abstract: Generative AI offers a simple, prompt-based alternative to fine-tuning smaller BERT-style LLMs for text classification tasks. This promises to eliminate the need for manually labeled training data and task-specific model training. However, it remains an open question whether tools like ChatGPT can deliver on this promise. In this paper, we show that smaller, fine-tuned LLMs (still) consistently and significantly outperform larger, zero-shot prompted models in text classification. We compare three major generative AI models (ChatGPT with GPT-3.5/GPT-4 and Claude Opus) with several fine-tuned LLMs across a diverse set of classification tasks (sentiment, approval/disapproval, emotions, party positions) and text categories (news, tweets, speeches). We find that fine-tuning with application-specific training data achieves superior performance in all cases. To make this approach more accessible to a broader audience, we provide an easy-to-use toolkit alongside this paper. Our toolkit, accompanied by non-technical step-by-step guidance, enables users to select and fine-tune BERT-like LLMs for any classification task with minimal technical and computational effort.
Authors: Lynn Chua, Badih Ghazi, Yangsibo Huang, Pritish Kamath, Ravi Kumar, Daogao Liu, Pasin Manurangsi, Amer Sinha, Chiyuan Zhang
Abstract: Large language models (LLMs) have emerged as powerful tools for tackling complex tasks across diverse domains, but they also raise privacy concerns when fine-tuned on sensitive data due to potential memorization. While differential privacy (DP) offers a promising solution by ensuring models are 'almost indistinguishable' with or without any particular privacy unit, current evaluations on LLMs mostly treat each example (text record) as the privacy unit. This leads to uneven user privacy guarantees when contributions per user vary. We therefore study user-level DP motivated by applications where it necessary to ensure uniform privacy protection across users. We present a systematic evaluation of user-level DP for LLM fine-tuning on natural language generation tasks. Focusing on two mechanisms for achieving user-level DP guarantees, Group Privacy and User-wise DP-SGD, we investigate design choices like data selection strategies and parameter tuning for the best privacy-utility tradeoff.
Authors: Bohao Yang, Dong Liu, Chen Tang, Chenghao Xiao, Kun Zhao, Chao Li, Lin Yuan, Guang Yang, Lanxiao Huang, Chenghua Lin
Abstract: Large Language Models (LLMs) demonstrate a remarkable ability to comprehend human instructions and generate high-quality text. This capability allows LLMs to function as agents that can emulate human beings at a more sophisticated level, beyond the mere replication of basic human behaviours. However, there is a lack of exploring into leveraging LLMs to craft characters from diverse aspects. In this work, we introduce the Customisable Conversation Agent Framework, which leverages LLMs to simulate real-world characters that can be freely customised according to various user preferences. This adaptable framework is beneficial for the design of customisable characters and role-playing agents aligned with human preferences. We propose the SimsConv dataset, which encompasses 68 different customised characters, 1,360 multi-turn role-playing dialogues, and a total of 13,971 interaction dialogues. The characters are created from several real-world elements, such as career, aspiration, trait, and skill. Building upon these foundations, we present SimsChat, a freely customisable role-playing agent. It incorporates diverse real-world scenes and topic-specific character interaction dialogues, thereby simulating characters' life experiences in various scenarios and topic-specific interactions with specific emotions. Experimental results indicate that our proposed framework achieves desirable performance and provides a valuable guideline for the construction of more accurate human simulacra in the future. Our data and code are publicly available at https://github.com/Bernard-Yang/SimsChat.
Authors: Davide Venditti, Elena Sofia Ruzzetti, Giancarlo A. Xompero, Cristina Giannone, Andrea Favalli, Raniero Romagnoli, Fabio Massimo Zanzotto
Abstract: Large Language Models (LLMs) are powerful tools with extensive applications, but their tendency to memorize private information raises significant concerns as private data leakage can easily happen. In this paper, we introduce Private Association Editing (PAE), a novel defense approach for private data leakage. PAE is designed to effectively remove Personally Identifiable Information (PII) without retraining the model. Our approach consists of a four-step procedure: detecting memorized PII, applying PAE cards to mitigate memorization of private data, verifying resilience to targeted data extraction (TDE) attacks, and ensuring consistency in the post-edit LLMs. The versatility and efficiency of PAE, which allows for batch modifications, significantly enhance data privacy in LLMs. Experimental results demonstrate the effectiveness of PAE in mitigating private data leakage. We believe PAE will serve as a critical tool in the ongoing effort to protect data privacy in LLMs, encouraging the development of safer models for real-world applications.
Authors: Yuan Feng, Junlin Lv, Yukun Cao, Xike Xie, S. Kevin Zhou
Abstract: Large Language Models have excelled in various fields but encounter challenges in memory and time efficiency due to the expanding Key-Value (KV) cache required for long-sequence inference. Recent efforts try to reduce KV cache size to a given memory budget by evicting vast non-critical cache elements during runtime, while preserving generation quality. Our revisiting of current eviction methods reveals that they fundamentally minimize an upper bound of the $L_1$ eviction loss between the pre- and post-eviction outputs of multi-head self-attention mechanisms. Moreover, our analysis indicates that the common practices of uniformly assigning budgets across attention heads harm their post-eviction generation quality. In light of these findings, we propose a simple yet effective adaptive budget allocation algorithm. This algorithm not only optimizes the theoretical loss upper bound but also reduces the $L_1$ eviction loss in practice by aligning with the varied characteristics across different heads. By integrating this algorithm into two state-of-the-art methods, we demonstrate the effectiveness of using adaptive budget allocation to optimize KV cache eviction. Extensive evaluations on 16 datasets and the Needle-in-a-Haystack test confirm significant performance improvements across various tasks.
Authors: Baixuan Li, Yunlong Fan, Zhiqiang Gao
Abstract: Multilingual large language models (MLLMs) struggle to answer questions posed in non-dominant languages, even though they have acquired the relevant knowledge from their dominant language corpus. In contrast, human multilinguals can overcome such non-native language context limitations through Positive Native Language Transfer (PNLT). Inspired by the process of PNLT, we analogize the dominant language of MLLMs to the native language of human multilinguals, and propose Native Language Prompting (NatLan) to simulate the PNLT observed in human multilinguals. It explicitly creates native language contexts for MLLMs to facilitate the elicitation of the rich native language knowledge during question-answering, unlocking the limitations imposed by non-native language contexts. By employing multi-MLLM collaboration, NatLan reduces the workload on each MLLM in simulating PNLT and refines semantic transfer. On the C-Eval benchmark, NatLan provides up to a 10.1% average accuracy improvement and up to a 5.0% increase in the hard-level subset across five MLLMs, surpassing all top-notch related methods. Our code is available at https://github.com/AnonyNLP/NatLan.
Authors: Sangjoon Park, Chan Woo Wee, Seo Hee Choi, Kyung Hwan Kim, Jee Suk Chang, Hong In Yoon, Ik Jae Lee, Yong Bae Kim, Jaeho Cho, Ki Chang Keum, Chang Geol Lee, Hwa Kyung Byun, Woong Sub Koom
Abstract: Accurate patient selection is critical in radiotherapy (RT) to prevent ineffective treatments. Traditional survival prediction models, relying on structured data, often lack precision. This study explores the potential of large language models (LLMs) to structure unstructured electronic health record (EHR) data, thereby improving survival prediction accuracy through comprehensive clinical information integration. Data from 34,276 patients treated with RT at Yonsei Cancer Center between 2013 and 2023 were analyzed, encompassing both structured and unstructured data. An open-source LLM was used to structure the unstructured EHR data via single-shot learning, with its performance compared against a domain-specific medical LLM and a smaller variant. Survival prediction models were developed using statistical, machine learning, and deep learning approaches, incorporating both structured and LLM-structured data. Clinical experts evaluated the accuracy of the LLM-structured data. The open-source LLM achieved 87.5% accuracy in structuring unstructured EHR data without additional training, significantly outperforming the domain-specific medical LLM, which reached only 35.8% accuracy. Larger LLMs were more effective, particularly in extracting clinically relevant features like general condition and disease extent, which closely correlated with patient survival. Incorporating LLM-structured clinical features into survival prediction models significantly improved accuracy, with the C-index of deep learning models increasing from 0.737 to 0.820. These models also became more interpretable by emphasizing clinically significant factors. This study shows that general-domain LLMs, even without specific medical training, can effectively structure large-scale unstructured EHR data, substantially enhancing the accuracy and interpretability of clinical predictive models.
Authors: Wenxuan Xie, Gaochen Wu, Bowen Zhou
Abstract: Recent In-Context Learning based methods have achieved remarkable success in Text-to-SQL task. However, there is still a large gap between the performance of these models and human performance on datasets with complex database schema and difficult questions, such as BIRD. Besides, existing work has neglected to supervise intermediate steps when solving questions iteratively with question decomposition methods, and the schema linking methods used in these works are very rudimentary. To address these issues, we propose MAG-SQL, a multi-agent generative approach with soft schema linking and iterative Sub-SQL refinement. In our framework, an entity-based method with tables' summary is used to select the columns in database, and a novel targets-conditions decomposition method is introduced to decompose those complex questions. Additionally, we build a iterative generating module which includes a Sub-SQL Generator and Sub-SQL Refiner, introducing external oversight for each step of generation. Through a series of ablation studies, the effectiveness of each agent in our framework has been demonstrated. When evaluated on the BIRD benchmark with GPT-4, MAG-SQL achieves an execution accuracy of 61.08%, compared to the baseline accuracy of 46.35% for vanilla GPT-4 and the baseline accuracy of 57.56% for MAC-SQL. Besides, our approach makes similar progress on Spider.
Authors: Abeer Aldayel, Areej Alokaili, Rehab Alahmadi
Abstract: While various approaches have recently been studied for bias identification, little is known about how implicit language that does not explicitly convey a viewpoint affects bias amplification in large language models. To examine the severity of bias toward a view, we evaluated the performance of two downstream tasks where the implicit and explicit knowledge of social groups were used. First, we present a stress test evaluation by using a biased model in edge cases of excessive bias scenarios. Then, we evaluate how LLMs calibrate linguistically in response to both implicit and explicit opinions when they are aligned with conflicting viewpoints. Our findings reveal a discrepancy in LLM performance in identifying implicit and explicit opinions, with a general tendency of bias toward explicit opinions of opposing stances. Moreover, the bias-aligned models generate more cautious responses using uncertainty phrases compared to the unaligned (zero-shot) base models. The direct, incautious responses of the unaligned models suggest a need for further refinement of decisiveness by incorporating uncertainty markers to enhance their reliability, especially on socially nuanced topics with high subjectivity.
Authors: Yongshuo Zong, Oisin Mac Aodha, Timothy Hospedales
Abstract: Multimodal learning, which aims to understand and analyze information from multiple modalities, has achieved substantial progress in the supervised regime in recent years. However, the heavy dependence on data paired with expensive human annotations impedes scaling up models. Meanwhile, given the availability of large-scale unannotated data in the wild, self-supervised learning has become an attractive strategy to alleviate the annotation bottleneck. Building on these two directions, self-supervised multimodal learning (SSML) provides ways to learn from raw multimodal data. In this survey, we provide a comprehensive review of the state-of-the-art in SSML, in which we elucidate three major challenges intrinsic to self-supervised learning with multimodal data: (1) learning representations from multimodal data without labels, (2) fusion of different modalities, and (3) learning with unaligned data. We then detail existing solutions to these challenges. Specifically, we consider (1) objectives for learning from multimodal unlabeled data via self-supervision, (2) model architectures from the perspective of different multimodal fusion strategies, and (3) pair-free learning strategies for coarse-grained and fine-grained alignment. We also review real-world applications of SSML algorithms in diverse fields such as healthcare, remote sensing, and machine translation. Finally, we discuss challenges and future directions for SSML. A collection of related resources can be found at: https://github.com/ys-zong/awesome-self-supervised-multimodal-learning.
URLs: https://github.com/ys-zong/awesome-self-supervised-multimodal-learning.
Authors: Haoran Wang, Kai Shu
Abstract: To ensure AI safety, instruction-tuned Large Language Models (LLMs) are specifically trained to ensure alignment, which refers to making models behave in accordance with human intentions. While these models have demonstrated commendable results on various safety benchmarks, the vulnerability of their safety alignment has not been extensively studied. This is particularly troubling given the potential harm that LLMs can inflict. Existing attack methods on LLMs often rely on poisoned training data or the injection of malicious prompts. These approaches compromise the stealthiness and generalizability of the attacks, making them susceptible to detection. Additionally, these models often demand substantial computational resources for implementation, making them less practical for real-world applications. In this work, we study a different attack scenario, called Trojan Activation Attack (TA^2), which injects trojan steering vectors into the activation layers of LLMs. These malicious steering vectors can be triggered at inference time to steer the models toward attacker-desired behaviors by manipulating their activations. Our experiment results on four primary alignment tasks show that TA^2 is highly effective and adds little or no overhead to attack efficiency. Additionally, we discuss potential countermeasures against such activation attacks.
Authors: Dimitri Coelho Mollo
Abstract: Artificial Intelligence is a field that lives many lives, and the term has come to encompass a motley collection of scientific and commercial endeavours. In this paper, I articulate the contours of a rather neglected but central scientific role that AI has to play, which I dub `AI-as-exploration'.The basic thrust of AI-as-exploration is that of creating and studying systems that can reveal candidate building blocks of intelligence that may differ from the forms of human and animal intelligence we are familiar with. In other words, I suggest that AI is one of the best tools we have for exploring intelligence space, namely the space of possible intelligent systems. I illustrate the value of AI-as-exploration by focusing on a specific case study, i.e., recent work on the capacity to combine novel and invented concepts in humans and Large Language Models. I show that the latter, despite showing human-level accuracy in such a task, probably solve it in ways radically different, but no less relevant to intelligence research, to those hypothesised for humans.
Authors: Gang Hu, Ke Qin, Chenhan Yuan, Min Peng, Alejandro Lopez-Lira, Benyou Wang, Sophia Ananiadou, Jimin Huang, Qianqian Xie
Abstract: While the progression of Large Language Models (LLMs) has notably propelled financial analysis, their application has largely been confined to singular language realms, leaving untapped the potential of bilingual Chinese-English capacity. To bridge this chasm, we introduce ICE-PIXIU, seamlessly amalgamating the ICE-INTENT model and ICE-FLARE benchmark for bilingual financial analysis. ICE-PIXIU uniquely integrates a spectrum of Chinese tasks, alongside translated and original English datasets, enriching the breadth and depth of bilingual financial modeling. It provides unrestricted access to diverse model variants, a substantial compilation of diverse cross-lingual and multi-modal instruction data, and an evaluation benchmark with expert annotations, comprising 10 NLP tasks, 20 bilingual specific tasks, totaling 95k datasets. Our thorough evaluation emphasizes the advantages of incorporating these bilingual datasets, especially in translation tasks and utilizing original English data, enhancing both linguistic flexibility and analytical acuity in financial contexts. Notably, ICE-INTENT distinguishes itself by showcasing significant enhancements over conventional LLMs and existing financial LLMs in bilingual milieus, underscoring the profound impact of robust bilingual data on the accuracy and efficacy of financial NLP.
Authors: Emile Contal, Garrin McGoldrick
Abstract: Large Language Models (LLM) hold immense promise for real-world applications, but their generic knowledge often falls short of domain-specific needs. Fine-tuning, a common approach, can suffer from catastrophic forgetting and hinder generalizability. In-Context Learning (ICL) offers an alternative, which can leverage Retrieval-Augmented Generation (RAG) to provide LLMs with relevant demonstrations for few-shot learning tasks. This paper explores the desired qualities of a demonstration retrieval system for ICL. We argue that ICL retrieval in this context resembles item-cold-start recommender systems, prioritizing discovery and maximizing information gain over strict relevance. We propose a novel evaluation method that measures the LLM's subsequent performance on NLP tasks, eliminating the need for subjective diversity scores. Our findings demonstrate the critical role of diversity and quality bias in retrieved demonstrations for effective ICL, and highlight the potential of recommender system techniques in this domain.
Authors: Yanming Liu, Xinyue Peng, Yuwei Zhang, Xiaolan Ke, Songhang Deng, Jiannan Cao, Chen Ma, Mengchen Fu, Xuhong Zhang, Sheng Cheng, Xun Wang, Jianwei Yin, Tianyu Du
Abstract: Large language models have repeatedly shown outstanding performance across diverse applications. However, deploying these models can inadvertently risk user privacy. The significant memory demands during training pose a major challenge in terms of resource consumption. This substantial size places a heavy load on memory resources, raising considerable practical concerns. In this paper, we introduce DP-MemArc, a novel training framework aimed at reducing the memory costs of large language models while emphasizing the protection of user data privacy. DP-MemArc incorporates side network or reversible network designs to support a variety of differential privacy memory-efficient fine-tuning schemes. Our approach not only achieves in memory optimization but also ensures robust privacy protection, keeping user data secure and confidential. Extensive experiments have demonstrated that DP-MemArc effectively provides differential privacy-efficient fine-tuning across different task scenarios.
Authors: Yu-An Liu, Ruqing Zhang, Jiafeng Guo, Maarten de Rijke, Yixing Fan, Xueqi Cheng
Abstract: Recent advances in neural information retrieval (IR) models have significantly enhanced their effectiveness over various IR tasks. The robustness of these models, essential for ensuring their reliability in practice, has also garnered significant attention. With a wide array of research on robust IR being proposed, we believe it is the opportune moment to consolidate the current status, glean insights from existing methodologies, and lay the groundwork for future development. We view the robustness of IR to be a multifaceted concept, emphasizing its necessity against adversarial attacks, out-of-distribution (OOD) scenarios and performance variance. With a focus on adversarial and OOD robustness, we dissect robustness solutions for dense retrieval models (DRMs) and neural ranking models (NRMs), respectively, recognizing them as pivotal components of the neural IR pipeline. We provide an in-depth discussion of existing methods, datasets, and evaluation metrics, shedding light on challenges and future directions in the era of large language models. To the best of our knowledge, this is the first comprehensive survey on the robustness of neural IR models, and we will also be giving our first tutorial presentation at SIGIR 2024 \url{https://sigir2024-robust-information-retrieval.github.io}. Along with the organization of existing work, we introduce a Benchmark for robust IR (BestIR), a heterogeneous evaluation benchmark for robust neural information retrieval, which is publicly available at \url{https://github.com/Davion-Liu/BestIR}. We hope that this study provides useful clues for future research on the robustness of IR models and helps to develop trustworthy search engines \url{https://github.com/Davion-Liu/Awesome-Robustness-in-Information-Retrieval}.
URLs: https://sigir2024-robust-information-retrieval.github.io, https://github.com/Davion-Liu/BestIR, https://github.com/Davion-Liu/Awesome-Robustness-in-Information-Retrieval
Authors: Haoran Zheng, Utku Pamuksuz
Abstract: Explainable Artificial Intelligence (XAI) plays a crucial role in enhancing the transparency and accountability of AI models, particularly in natural language processing (NLP) tasks. However, popular XAI methods such as LIME and SHAP have been found to be unstable and potentially misleading, underscoring the need for a standardized evaluation approach. This paper introduces SCENE (Soft Counterfactual Evaluation for Natural language Explainability), a novel evaluation method that leverages large language models (LLMs) to generate Soft Counterfactual explanations in a zero-shot manner. By focusing on token-based substitutions, SCENE creates contextually appropriate and semantically meaningful Soft Counterfactuals without extensive fine-tuning. SCENE adopts Validitysoft and Csoft metrics to assess the effectiveness of model-agnostic XAI methods in text classification tasks. Applied to CNN, RNN, and Transformer architectures, SCENE provides valuable insights into the strengths and limitations of various XAI techniques.
Authors: Kyudan Jung, Sieun Hyeon, Jeong Youn Kwon, Nam-Joon Kim, Hyun Gon Ryu, Hyuk-Jae Lee, Jaeyoung Do
Abstract: Improving the readability of mathematical expressions in text-based document such as subtitle of mathematical video, is an significant task. To achieve this, mathematical expressions should be convert to compiled formulas. For instance, the spoken expression ``x equals minus b plus or minus the square root of b squared minus four a c, all over two a'' from automatic speech recognition is more readily comprehensible when displayed as a compiled formula $x = \frac{-b \pm \sqrt{b^2 - 4ac}}{2a}$. To convert mathematical spoken sentences to compiled formulas, two processes are required: spoken sentences are converted into LaTeX formulas, and LaTeX formulas are converted into compiled formulas. The latter can be managed by using LaTeX engines. However, there is no way to do the former effectively. Even if we try to solve this using language models, there is no paired data between spoken sentences and LaTeX formulas to train it. In this paper, we introduce MathBridge, the first extensive dataset for translating mathematical spoken sentences into LaTeX formulas. MathBridge comprises approximately 23 million LaTeX formulas paired with the corresponding mathematical spoken sentences. Through comprehensive evaluations, including fine-tuning with proposed data, we discovered that MathBridge significantly enhances the capabilities of pretrained language models for converting to LaTeX formulas from mathematical spoken sentences. Specifically, for the T5-large model, the sacreBLEU score increased from 4.77 to 46.8, demonstrating substantial enhancement.