Abstract
The rapid development of large language models (LLMs) enables them to convey factual knowledge in a more human-like fashion. Extensive efforts have been made to reduce factual hallucinations by modifying LLMs with factuality decoding. However, they also pose risks of hindering knowledge updates, as they make models overly confident in known facts. In this work, we first revisite the current factuality decoding methods and verified their effectiveness in enhancing factual accuracy. Subsequently, we conduct further evaluation of several strong factuality decoding methods on the knowledge editing benchmark. All these decoding methods significantly diminish the performance of llama2 models compared to their original decoding, with the largest decrease being a staggering 81.3\%. This further indicates that the current existing decoding methods still cannot perfectly address the factual hallucinations, as they overlook the importance of preserving the flexibility for knowledge editing. Therefore, our work suggests that research into factual alignment should simultaneously focus on the effectiveness of knowledge editing.
DeFT: Flash Tree-attention with IO-Awareness for Efficient Tree-search-based LLM Inference
Authors: Jinwei Yao, Kaiqi Chen, Kexun Zhang, Jiaxuan You, Binhang Yuan, Zeke Wang, Tao Lin
Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
Abstract
Decoding using tree search can greatly enhance the inference quality for transformer-based Large Language Models (LLMs). Depending on the guidance signal, it searches for the best path from root to leaf in the tree by forming LLM outputs to improve controllability, reasoning ability, alignment, et cetera. However, current tree decoding strategies and their inference systems do not suit each other well due to redundancy in computation, memory footprints, and memory access, resulting in inefficient inference. To address this issue, we propose DeFT, an IO-aware tree attention algorithm that maintains memory-efficient attention calculation with low memory footprints in two stages: (1) QKV Preparation: we propose a KV-Guided Tree Split strategy to group QKV wisely for high utilization of GPUs and reduction of memory reads/writes for the KV cache between GPU global memory and on-chip shared memory as much as possible; (2) Attention Calculation: we calculate partial attention of each QKV groups in a fused kernel then apply a Tree-topology-aware Global Reduction strategy to get final attention. Thanks to a reduction in KV cache IO by 3.6-4.5$\times$, along with an additional reduction in IO for $\mathbf{Q} \mathbf{K}^\top$ and Softmax equivalent to 25% of the total KV cache IO, DeFT can achieve a speedup of 1.7-2.4$\times$ in end-to-end latency across two practical reasoning tasks over the SOTA attention algorithms.
Dialectical Alignment: Resolving the Tension of 3H and Security Threats of LLMs
Authors: Shu Yang, Jiayuan Su, Han Jiang, Mengdi Li, Keyuan Cheng, Muhammad Asif Ali, Lijie Hu, Di Wang
Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
Abstract
With the rise of large language models (LLMs), ensuring they embody the principles of being helpful, honest, and harmless (3H), known as Human Alignment, becomes crucial. While existing alignment methods like RLHF, DPO, etc., effectively fine-tune LLMs to match preferences in the preference dataset, they often lead LLMs to highly receptive human input and external evidence, even when this information is poisoned. This leads to a tendency for LLMs to be Adaptive Chameleons when external evidence conflicts with their parametric memory. This exacerbates the risk of LLM being attacked by external poisoned data, which poses a significant security risk to LLM system applications such as Retrieval-augmented generation (RAG). To address the challenge, we propose a novel framework: Dialectical Alignment (DA), which (1) utilizes AI feedback to identify optimal strategies for LLMs to navigate inter-context conflicts and context-memory conflicts with different external evidence in context window (i.e., different ratios of poisoned factual contexts); (2) constructs the SFT dataset as well as the preference dataset based on the AI feedback and strategies above; (3) uses the above datasets for LLM alignment to defense poisoned context attack while preserving the effectiveness of in-context knowledge editing. Our experiments show that the dialectical alignment model improves poisoned data attack defense by 20 and does not require any additional prompt engineering or prior declaration of you may be attacked to the LLMs' context window.
Comparing Bad Apples to Good Oranges: Aligning Large Language Models via Joint Preference Optimization
Abstract
A common technique for aligning large language models (LLMs) relies on acquiring human preferences by comparing multiple generations conditioned on a fixed context. This only leverages the pairwise comparisons when the generations are placed in an identical context. However, such conditional rankings often fail to capture the complex and multidimensional aspects of human preferences. In this work, we revisit the traditional paradigm of preference acquisition and propose a new axis that is based on eliciting preferences jointly over the instruction-response pairs. While prior preference optimizations are designed for conditional ranking protocols (e.g., DPO), our proposed preference acquisition protocol introduces DOVE, a new preference optimization objective that upweights the joint probability of the chosen instruction-response pair over the rejected instruction-response pair. Interestingly, we find that the LLM trained with joint instruction-response preference data using DOVE outperforms the LLM trained with DPO by 5.2% and 3.3% win-rate for the summarization and open-ended dialogue datasets, respectively. Our findings reveal that joint preferences over instruction and response pairs can significantly enhance the alignment of LLMs by tapping into a broader spectrum of human preference elicitation. The data and code is available at https://github.com/Hritikbansal/dove.
EvoCodeBench: An Evolving Code Generation Benchmark Aligned with Real-World Code Repositories
Authors: Jia Li, Ge Li, Xuanming Zhang, Yihong Dong, Zhi Jin
Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Software Engineering (cs.SE)
Abstract
How to evaluate Large Language Models (LLMs) in code generation is an open question. Existing benchmarks demonstrate poor alignment with real-world code repositories and are insufficient to evaluate the coding abilities of LLMs. This paper proposes a new benchmark - EvoCodeBench to address the preceding problems, which has three primary advances. (1) EvoCodeBench aligns with real-world repositories in multiple dimensions, e.g., code distributions and dependency distributions. (2) EvoCodeBench offers comprehensive annotations (e.g., requirements, reference code, and reference dependencies), and robust evaluation metrics (e.g., Pass@k and Recall@k). (3) EvoCodeBench is an evolving benchmark to avoid data leakage. We build an automatic pipeline to update EvoCodeBench from the latest repositories. We release the first version - EvoCodeBench-2403, containing 275 samples from 25 real-world repositories. Based on EvoCodeBench, we propose repository-level code generation and evaluate 10 popular LLMs (e.g., gpt-4, gpt-3.5, DeepSeek Coder, StarCoder 2, CodeLLaMa, Gemma, and Qwen 1.5). Our experiments reveal the coding abilities of these LLMs in real-world repositories. For example, the highest Pass@1 of gpt-4 only is 20.73% in our experiments. We also analyze failed cases and summarize the shortcomings of existing LLMs in EvoCodeBench. We release EvoCodeBench, all prompts, and LLMs' completions for further community analysis.
Extensive Self-Contrast Enables Feedback-Free Language Model Alignment
Authors: Xiao Liu, Xixuan Song, Yuxiao Dong, Jie Tang
Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Abstract
Reinforcement learning from human feedback (RLHF) has been a central technique for recent large language model (LLM) alignment. However, its heavy dependence on costly human or LLM-as-Judge preference feedback could stymie its wider applications. In this work, we introduce Self-Contrast, a feedback-free large language model alignment method via exploiting extensive self-generated negatives. With only supervised fine-tuning (SFT) targets, Self-Contrast leverages the LLM itself to generate massive diverse candidates, and harnesses a pre-trained embedding model to filter multiple negatives according to text similarity. Theoretically, we illustrate that in this setting, merely scaling negative responses can still effectively approximate situations with more balanced positive and negative preference annotations. Our experiments with direct preference optimization (DPO) on three datasets show that, Self-Contrast could consistently outperform SFT and standard DPO training by large margins. And as the number of self-generated negatives increases, the performance of Self-Contrast continues to grow. Code and data are available at https://github.com/THUDM/Self-Contrast.
Bailong: Bilingual Transfer Learning based on QLoRA and Zip-tie Embedding
Authors: Lung-Chuan Chen, Zong-Ru Li
Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
Abstract
Large language models (LLMs) have demonstrated exceptional performance in various NLP applications. However, the majority of existing open-source LLMs are pre-trained primarily on English data and little part of other languages. This deficiency in multilingual training data results in suboptimal performance when applied to languages with fewer available resources. Furthermore, enhancing the performance of LLMs on low-resource languages by full-parameter fine-tuning with additional data requires substantial computational resources, posing computational barriers for research organizations and individual researchers. Consequently, several techniques such as parameter-efficient tuning and advanced embedding initialization have been proposed to address these challenges. In this work, we combine them to facilitate cross-lingual transfer on English-dominated open-source LLM. To effectively enhance the model's proficiency in Traditional Chinese, we conduct secondary pre-training on Llama 2 7B with Traditional Chinese data by leveraging QLoRA and our proposed zip-tie embedding initialization. The resulting model called Bailong, which stands for Bilingual trAnsfer learnIng based on qLOra and zip-tie embeddiNG. We present Bailong-instruct 7B, a fine-tuned version of Bailong 7B optimized for multi-turn dialogue scenarios. Recognizing the inadequacy of benchmark datasets in Traditional Chinese, we further introduce Bailong-bench to assess the alignment of models with human preferences and the capability to follow instructions in both Traditional Chinese and English tasks. In our evaluation, Bailong-instruct 7B exhibits competitive performance on Bailong-bench and other benchmark datasets when compared to other open-source models of similar or even larger parameter sizes. Bailong-instruct 7B and Bailong-bench are publicly available with the aim of empowering the community to build upon our efforts.
A Survey on Multilingual Large Language Models: Corpora, Alignment, and Bias
Abstract
Based on the foundation of Large Language Models (LLMs), Multilingual Large Language Models (MLLMs) have been developed to address the challenges of multilingual natural language processing tasks, hoping to achieve knowledge transfer from high-resource to low-resource languages. However, significant limitations and challenges still exist, such as language imbalance, multilingual alignment, and inherent bias. In this paper, we aim to provide a comprehensive analysis of MLLMs, delving deeply into discussions surrounding these critical issues. First of all, we start by presenting an overview of MLLMs, covering their evolution, key techniques, and multilingual capacities. Secondly, we explore widely utilized multilingual corpora for MLLMs' training and multilingual datasets oriented for downstream tasks that are crucial for enhancing the cross-lingual capability of MLLMs. Thirdly, we survey the existing studies on multilingual representations and investigate whether the current MLLMs can learn a universal language representation. Fourthly, we discuss bias on MLLMs including its category and evaluation metrics, and summarize the existing debiasing techniques. Finally, we discuss existing challenges and point out promising research directions. By demonstrating these aspects, this paper aims to facilitate a deeper understanding of MLLMs and their potentiality in various domains.
ChatGLM-RLHF: Practices of Aligning Large Language Models with Human Feedback
Abstract
ChatGLM is a free-to-use AI service powered by the ChatGLM family of large language models (LLMs). In this paper, we present the ChatGLM-RLHF pipeline -- a reinforcement learning from human feedback (RLHF) system -- designed to enhance ChatGLM's alignment with human preferences. ChatGLM-RLHF encompasses three major components: the collection of human preference data, the training of the reward model, and the optimization of policies. Throughout the process of integrating ChatGLM-RLHF into production, we encountered and addressed several unprecedented challenges. We introduce the strategies to mitigate reward variance for stabilized large-scale training, implement model parallelism with fused gradient-descent, and design regularization constraints to avoid catastrophic forgetting in LLMs. Experiments show that ChatGLM-RLHF brings significant improvements in alignment tasks compared to the supervised fine-tuned (SFT) version of ChatGLM. For instance, it achieves on average 15\% more wins against ChatGLM-SFT in Chinese alignment tasks. The work presents our practices of aligning LLMs with human preferences, offering insights into the challenges and solutions in RLHF implementations.
Prior Constraints-based Reward Model Training for Aligning Large Language Models
Abstract
Reinforcement learning with human feedback for aligning large language models (LLMs) trains a reward model typically using ranking loss with comparison pairs.However, the training procedure suffers from an inherent problem: the uncontrolled scaling of reward scores during reinforcement learning due to the lack of constraints while training the reward model.This paper proposes a Prior Constraints-based Reward Model (namely PCRM) training method to mitigate this problem. PCRM incorporates prior constraints, specifically, length ratio and cosine similarity between outputs of each comparison pair, during reward model training to regulate optimization magnitude and control score margins. We comprehensively evaluate PCRM by examining its rank correlation with human preferences and its effectiveness in aligning LLMs via RL. Experimental results demonstrate that PCRM significantly improves alignment performance by effectively constraining reward score scaling. As another bonus, our method is easily integrated into arbitrary rank-based alignment methods, such as direct preference optimization, and can yield consistent improvement.
Generating Faithful and Complete Hospital-Course Summaries from the Electronic Health Record
Abstract
The rapid adoption of Electronic Health Records (EHRs) has been instrumental in streamlining administrative tasks, increasing transparency, and enabling continuity of care across providers. An unintended consequence of the increased documentation burden, however, has been reduced face-time with patients and, concomitantly, a dramatic rise in clinician burnout. In this thesis, we pinpoint a particularly time-intensive, yet critical, documentation task: generating a summary of a patient's hospital admissions, and propose and evaluate automated solutions. In Chapter 2, we construct a dataset based on 109,000 hospitalizations (2M source notes) and perform exploratory analyses to motivate future work on modeling and evaluation [NAACL 2021]. In Chapter 3, we address faithfulness from a modeling perspective by revising noisy references [EMNLP 2022] and, to reduce the reliance on references, directly calibrating model outputs to metrics [ACL 2023]. These works relied heavily on automatic metrics as human annotations were limited. To fill this gap, in Chapter 4, we conduct a fine-grained expert annotation of system errors in order to meta-evaluate existing metrics and better understand task-specific issues of domain adaptation and source-summary alignments. To learn a metric less correlated to extractiveness (copy-and-paste), we derive noisy faithfulness labels from an ensemble of existing metrics and train a faithfulness classifier on these pseudo labels [MLHC 2023]. Finally, in Chapter 5, we demonstrate that fine-tuned LLMs (Mistral and Zephyr) are highly prone to entity hallucinations and cover fewer salient entities. We improve both coverage and faithfulness by performing sentence-level entity planning based on a set of pre-computed salient entities from the source text, which extends our work on entity-guided news summarization [ACL, 2023], [EMNLP, 2023].
UniArk: Improving Generalisation and Consistency for Factual Knowledge Extraction through Debiasing
Authors: Yijun Yang, Jie He, Pinzhen Chen, Víctor Gutiérrez-Basulto, Jeff Z. Pan
Abstract
Several recent papers have investigated the potential of language models as knowledge bases as well as the existence of severe biases when extracting factual knowledge. In this work, we focus on the factual probing performance over unseen prompts from tuning, and using a probabilistic view we show the inherent misalignment between pre-training and downstream tuning objectives in language models for probing knowledge. We hypothesize that simultaneously debiasing these objectives can be the key to generalisation over unseen prompts. We propose an adapter-based framework, UniArk, for generalised and consistent factual knowledge extraction through simple methods without introducing extra parameters. Extensive experiments show that UniArk can significantly improve the model's out-of-domain generalisation as well as consistency under various prompts. Additionally, we construct ParaTrex, a large-scale and diverse dataset for measuring the inconsistency and out-of-domain generation of models. Further, ParaTrex offers a reference method for constructing paraphrased datasets using large language models.
Design as Desired: Utilizing Visual Question Answering for Multimodal Pre-training
Authors: Tongkun Su, Jun Li, Xi Zhang, Haibo Jin, Hao Chen, Qiong Wang, Faqin Lv, Baoliang Zhao, Yin Hu
Subjects: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL)
Abstract
Multimodal pre-training demonstrates its potential in the medical domain, which learns medical visual representations from paired medical reports. However, many pre-training tasks require extra annotations from clinicians, and most of them fail to explicitly guide the model to learn the desired features of different pathologies. To the best of our knowledge, we are the first to utilize Visual Question Answering (VQA) for multimodal pre-training to guide the framework focusing on targeted pathological features. In this work, we leverage descriptions in medical reports to design multi-granular question-answer pairs associated with different diseases, which assist the framework in pre-training without requiring extra annotations from experts. We also propose a novel pre-training framework with a quasi-textual feature transformer, a module designed to transform visual features into a quasi-textual space closer to the textual domain via a contrastive learning strategy. This narrows the vision-language gap and facilitates modality alignment. Our framework is applied to four downstream tasks: report generation, classification, segmentation, and detection across five datasets. Extensive experiments demonstrate the superiority of our framework compared to other state-of-the-art methods. Our code will be released upon acceptance.
LLaMA-Excitor: General Instruction Tuning via Indirect Feature Interaction
Abstract
Existing methods to fine-tune LLMs, like Adapter, Prefix-tuning, and LoRA, which introduce extra modules or additional input sequences to inject new skills or knowledge, may compromise the innate abilities of LLMs. In this paper, we propose LLaMA-Excitor, a lightweight method that stimulates the LLMs' potential to better follow instructions by gradually paying more attention to worthwhile information. Specifically, the LLaMA-Excitor does not directly change the intermediate hidden state during the self-attention calculation of the transformer structure. We designed the Excitor block as a bypass module for the similarity score computation in LLMs' self-attention to reconstruct keys and change the importance of values by learnable prompts. LLaMA-Excitor ensures a self-adaptive allocation of additional attention to input instructions, thus effectively preserving LLMs' pre-trained knowledge when fine-tuning LLMs on low-quality instruction-following datasets. Furthermore, we unify the modeling of multi-modal tuning and language-only tuning, extending LLaMA-Excitor to a powerful visual instruction follower without the need for complex multi-modal alignment. Our proposed approach is evaluated in language-only and multi-modal tuning experimental scenarios. Notably, LLaMA-Excitor is the only method that maintains basic capabilities while achieving a significant improvement (+6%) on the MMLU benchmark. In the visual instruction tuning, we achieve a new state-of-the-art image captioning performance of 157.5 CIDEr on MSCOCO, and a comparable performance (88.39%) on ScienceQA to cutting-edge models with more parameters and extensive vision-language pertaining.
LLMs are Good Sign Language Translators
Authors: Jia Gong, Lin Geng Foo, Yixuan He, Hossein Rahmani, Jun Liu
Subjects: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL)
Abstract
Sign Language Translation (SLT) is a challenging task that aims to translate sign videos into spoken language. Inspired by the strong translation capabilities of large language models (LLMs) that are trained on extensive multilingual text corpora, we aim to harness off-the-shelf LLMs to handle SLT. In this paper, we regularize the sign videos to embody linguistic characteristics of spoken language, and propose a novel SignLLM framework to transform sign videos into a language-like representation for improved readability by off-the-shelf LLMs. SignLLM comprises two key modules: (1) The Vector-Quantized Visual Sign module converts sign videos into a sequence of discrete character-level sign tokens, and (2) the Codebook Reconstruction and Alignment module converts these character-level tokens into word-level sign representations using an optimal transport formulation. A sign-text alignment loss further bridges the gap between sign and text tokens, enhancing semantic compatibility. We achieve state-of-the-art gloss-free results on two widely-used SLT benchmarks.
What's in Your "Safe" Data?: Identifying Benign Data that Breaks Safety
Authors: Luxi He, Mengzhou Xia, Peter Henderson
Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Cryptography and Security (cs.CR)
Abstract
Current Large Language Models (LLMs), even those tuned for safety and alignment, are susceptible to jailbreaking. Some have found that just further fine-tuning an aligned model with benign data (i.e., data without harmful content) surprisingly leads to substantial degradation in safety. We delve into the data-centric aspects of why benign fine-tuning inadvertently contributes to jailbreaking. First, we represent fine-tuning data through two lenses: representation and gradient spaces. Furthermore, we propose a bi-directional anchoring method that prioritizes data points that are close to harmful examples and distant from benign ones. By doing so, our approach effectively identifies subsets of benign data that are more likely to degrade the model's safety after fine-tuning. Training on just 100 of these seemingly benign datapoints can lead to the fine-tuned model affirmatively responding to > 70% of tested harmful requests, compared to < 20% after fine-tuning on randomly selected data. We further find that selected data are often in the form of lists and bullet points, or math questions.
Evaluating Text-to-Visual Generation with Image-to-Text Generation
Authors: Zhiqiu Lin, Deepak Pathak, Baiqi Li, Jiayao Li, Xide Xia, Graham Neubig, Pengchuan Zhang, Deva Ramanan
Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG); Multimedia (cs.MM)
Abstract
Despite significant progress in generative AI, comprehensive evaluation remains challenging because of the lack of effective metrics and standardized benchmarks. For instance, the widely-used CLIPScore measures the alignment between a (generated) image and text prompt, but it fails to produce reliable scores for complex prompts involving compositions of objects, attributes, and relations. One reason is that text encoders of CLIP can notoriously act as a "bag of words", conflating prompts such as "the horse is eating the grass" with "the grass is eating the horse". To address this, we introduce the VQAScore, which uses a visual-question-answering (VQA) model to produce an alignment score by computing the probability of a "Yes" answer to a simple "Does this figure show '{text}'?" question. Though simpler than prior art, VQAScore computed with off-the-shelf models produces state-of-the-art results across many (8) image-text alignment benchmarks. We also compute VQAScore with an in-house model that follows best practices in the literature. For example, we use a bidirectional image-question encoder that allows image embeddings to depend on the question being asked (and vice versa). Our in-house model, CLIP-FlanT5, outperforms even the strongest baselines that make use of the proprietary GPT-4V. Interestingly, although we train with only images, VQAScore can also align text with video and 3D models. VQAScore allows researchers to benchmark text-to-visual generation using complex texts that capture the compositional structure of real-world prompts. We introduce GenAI-Bench, a more challenging benchmark with 1,600 compositional text prompts that require parsing scenes, objects, attributes, relationships, and high-order reasoning like comparison and logic. GenAI-Bench also offers over 15,000 human ratings for leading image and video generation models such as Stable Diffusion, DALL-E 3, and Gen2.
Keyword: aligning
Aurora-M: The First Open Source Multilingual Language Model Red-teamed according to the U.S. Executive Order
Authors: Taishi Nakamura, Mayank Mishra, Simone Tedeschi, Yekun Chai, Jason T Stillerman, Felix Friedrich, Prateek Yadav, Tanmay Laud, Vu Minh Chien, Terry Yue Zhuo, Diganta Misra, Ben Bogin, Xuan-Son Vu, Marzena Karpinska, Arnav Varma Dantuluri, Wojciech Kusa, Tommaso Furlanello, Rio Yokota, Niklas Muennighoff, Suhas Pai, Tosin Adewumi, Veronika Laippala, Xiaozhe Yao, Adalberto Junior, Alpay Ariyak, Aleksandr Drozd, Jordan Clive, Kshitij Gupta, Liangyu Chen, Qi Sun, Ken Tsui, Noah Persaud, Nour Fahmy, Tianlong Chen, Mohit Bansal, Nicolo Monti, Tai Dang, Ziyang Luo, Tien-Tung Bui, Roberto Navigli, Virendra Mehta, Matthew Blumberg, Victor May, Huu Nguyen, Sampo Pyysalo
Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Abstract
Pretrained language models underpin several AI applications, but their high computational cost for training limits accessibility. Initiatives such as BLOOM and StarCoder aim to democratize access to pretrained models for collaborative community development. However, such existing models face challenges: limited multilingual capabilities, continual pretraining causing catastrophic forgetting, whereas pretraining from scratch is computationally expensive, and compliance with AI safety and development laws. This paper presents Aurora-M, a 15B parameter multilingual open-source model trained on English, Finnish, Hindi, Japanese, Vietnamese, and code. Continually pretrained from StarCoderPlus on 435 billion additional tokens, Aurora-M surpasses 2 trillion tokens in total training token count. It is the first open-source multilingual model fine-tuned on human-reviewed safety instructions, thus aligning its development not only with conventional red-teaming considerations, but also with the specific concerns articulated in the Biden-Harris Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence. Aurora-M is rigorously evaluated across various tasks and languages, demonstrating robustness against catastrophic forgetting and outperforming alternatives in multilingual settings, particularly in safety evaluations. To promote responsible open-source LLM development, Aurora-M and its variants are released at https://huggingface.co/collections/aurora-m/aurora-m-models-65fdfdff62471e09812f5407 .
Regularized Best-of-N Sampling to Mitigate Reward Hacking for Language Model Alignment
Abstract
Best-of-N (BoN) sampling with a reward model has been shown to be an effective strategy for aligning Large Language Models (LLMs) to human preferences at the time of decoding. BoN sampling is susceptible to a problem known as reward hacking. Because the reward model is an imperfect proxy for the true objective, over-optimizing its value can compromise its performance on the true objective. A common solution to prevent reward hacking in preference learning techniques is to optimize a reward using proximity regularization (e.g., KL regularization), which ensures that the language model remains close to the reference model. In this research, we propose Regularized Best-of-N (RBoN), a variant of BoN that aims to mitigate reward hacking by incorporating a proximity term in response selection, similar to preference learning techniques. We evaluate two variants of RBoN on the AlpacaFarm dataset and find that they outperform BoN, especially when the proxy reward model has a low correlation with the true objective.
Keyword: align
A Likelihood Ratio Test of Genetic Relationship among Languages
Abstract
Lexical resemblances among a group of languages indicate that the languages could be genetically related, i.e., they could have descended from a common ancestral language. However, such resemblances can arise by chance and, hence, need not always imply an underlying genetic relationship. Many tests of significance based on permutation of wordlists and word similarity measures appeared in the past to determine the statistical significance of such relationships. We demonstrate that although existing tests may work well for bilateral comparisons, i.e., on pairs of languages, they are either infeasible by design or are prone to yield false positives when applied to groups of languages or language families. To this end, inspired by molecular phylogenetics, we propose a likelihood ratio test to determine if given languages are related based on the proportion of invariant character sites in the aligned wordlists applied during tree inference. Further, we evaluate some language families and show that the proposed test solves the problem of false positives. Finally, we demonstrate that the test supports the existence of macro language families such as Nostratic and Macro-Mayan.
Token-Efficient Leverage Learning in Large Language Models
Authors: Yuanhao Zeng, Min Wang, Yihang Wang, Yingxia Shao
Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Abstract
Large Language Models (LLMs) have excelled in various tasks but perform better in high-resource scenarios, which presents challenges in low-resource scenarios. Data scarcity and the inherent difficulty of adapting LLMs to specific tasks compound the challenge. To address the twin hurdles, we introduce \textbf{Leverage Learning}. We present a streamlined implement of this methodology called Token-Efficient Leverage Learning (TELL). TELL showcases the potential of Leverage Learning, demonstrating effectiveness across various LLMs and low-resource tasks, ranging from $10^4$ to $10^6$ tokens. It reduces task data requirements by up to nearly an order of magnitude compared to conventional Supervised Fine-Tuning (SFT) while delivering competitive performance. With the same amount of task data, TELL leads in improving task performance compared to SFT. We discuss the mechanism of Leverage Learning, suggesting it aligns with quantization hypothesis and explore its promising potential through empirical testing.
Evaluating the Factuality of Large Language Models using Large-Scale Knowledge Graphs
Abstract
The advent of Large Language Models (LLMs) has significantly transformed the AI landscape, enhancing machine learning and AI capabilities. Factuality issue is a critical concern for LLMs, as they may generate factually incorrect responses. In this paper, we propose GraphEval to evaluate an LLM's performance using a substantially large test dataset. Specifically, the test dataset is retrieved from a large knowledge graph with more than 10 million facts without expensive human efforts. Unlike conventional methods that evaluate LLMs based on generated responses, GraphEval streamlines the evaluation process by creating a judge model to estimate the correctness of the answers given by the LLM. Our experiments demonstrate that the judge model's factuality assessment aligns closely with the correctness of the LLM's generated outputs, while also substantially reducing evaluation costs. Besides, our findings offer valuable insights into LLM performance across different metrics and highlight the potential for future improvements in ensuring the factual integrity of LLM outputs. The code is publicly available at https://github.com/xz-liu/GraphEval.
Enhancing Content-based Recommendation via Large Language Model
Abstract
In real-world applications, users express different behaviors when they interact with different items, including implicit click/like interactions, and explicit comments/reviews interactions. Nevertheless, almost all recommender works are focused on how to describe user preferences by the implicit click/like interactions, to find the synergy of people. For the content-based explicit comments/reviews interactions, some works attempt to utilize them to mine the semantic knowledge to enhance recommender models. However, they still neglect the following two points: (1) The content semantic is a universal world knowledge; how do we extract the multi-aspect semantic information to empower different domains? (2) The user/item ID feature is a fundamental element for recommender models; how do we align the ID and content semantic feature space? In this paper, we propose a `plugin' semantic knowledge transferring method \textbf{LoID}, which includes two major components: (1) LoRA-based large language model pretraining to extract multi-aspect semantic information; (2) ID-based contrastive objective to align their feature spaces. We conduct extensive experiments with SOTA baselines on real-world datasets, the detailed results demonstrating significant improvements of our method LoID.
Keyword: vision language
There is no result
Keyword: vision-language
Open-Vocabulary Federated Learning with Multimodal Prototyping
Authors: Huimin Zeng, Zhenrui Yue, Dong Wang
Subjects: Computation and Language (cs.CL); Computer Vision and Pattern Recognition (cs.CV)
Abstract
Existing federated learning (FL) studies usually assume the training label space and test label space are identical. However, in real-world applications, this assumption is too ideal to be true. A new user could come up with queries that involve data from unseen classes, and such open-vocabulary queries would directly defect such FL systems. Therefore, in this work, we explicitly focus on the under-explored open-vocabulary challenge in FL. That is, for a new user, the global server shall understand her/his query that involves arbitrary unknown classes. To address this problem, we leverage the pre-trained vision-language models (VLMs). In particular, we present a novel adaptation framework tailored for VLMs in the context of FL, named as Federated Multimodal Prototyping (Fed-MP). Fed-MP adaptively aggregates the local model weights based on light-weight client residuals, and makes predictions based on a novel multimodal prototyping mechanism. Fed-MP exploits the knowledge learned from the seen classes, and robustifies the adapted VLM to unseen categories. Our empirical evaluation on various datasets validates the effectiveness of Fed-MP.
Do Vision-Language Models Understand Compound Nouns?
Abstract
Open-vocabulary vision-language models (VLMs) like CLIP, trained using contrastive loss, have emerged as a promising new paradigm for text-to-image retrieval. However, do VLMs understand compound nouns (CNs) (e.g., lab coat) as well as they understand nouns (e.g., lab)? We curate Compun, a novel benchmark with 400 unique and commonly used CNs, to evaluate the effectiveness of VLMs in interpreting CNs. The Compun benchmark challenges a VLM for text-to-image retrieval where, given a text prompt with a CN, the task is to select the correct image that shows the CN among a pair of distractor images that show the constituent nouns that make up the CN. Next, we perform an in-depth analysis to highlight CLIPs' limited understanding of certain types of CNs. Finally, we present an alternative framework that moves beyond hand-written templates for text prompts widely used by CLIP-like models. We employ a Large Language Model to generate multiple diverse captions that include the CN as an object in the scene described by the caption. Our proposed method improves CN understanding of CLIP by 8.25% on Compun. Code and benchmark are available at: https://github.com/sonalkum/Compun
Keyword: language-vision
There is no result
Keyword: phrase-grounding
There is no result
Keyword: phrase grounding
There is no result
Keyword: reference expression comprehension
There is no result
Keyword: chest
A Survey of using Large Language Models for Generating Infrastructure as Code
Abstract
Infrastructure as Code (IaC) is a revolutionary approach which has gained significant prominence in the Industry. IaC manages and provisions IT infrastructure using machine-readable code by enabling automation, consistency across the environments, reproducibility, version control, error reduction and enhancement in scalability. However, IaC orchestration is often a painstaking effort which requires specialised skills as well as a lot of manual effort. Automation of IaC is a necessity in the present conditions of the Industry and in this survey, we study the feasibility of applying Large Language Models (LLM) to address this problem. LLMs are large neural network-based models which have demonstrated significant language processing abilities and shown to be capable of following a range of instructions within a broad scope. Recently, they have also been adapted for code understanding and generation tasks successfully, which makes them a promising choice for the automatic generation of IaC configurations. In this survey, we delve into the details of IaC, usage of IaC in different platforms, their challenges, LLMs in terms of code-generation aspects and the importance of LLMs in IaC along with our own experiments. Finally, we conclude by presenting the challenges in this area and highlighting the scope for future research.
Keyword: x-ray
There is no result
Keyword: clinical
Small Language Models Learn Enhanced Reasoning Skills from Medical Textbooks
Authors: Hyunjae Kim, Hyeon Hwang, Jiwoo Lee, Sihyeon Park, Dain Kim, Taewhoo Lee, Chanwoong Yoon, Jiwoong Sohn, Donghee Choi, Jaewoo Kang
Abstract
While recent advancements in commercial large language models (LM) have shown promising results in medical tasks, their closed-source nature poses significant privacy and security concerns, hindering their widespread use in the medical field. Despite efforts to create open-source models, their limited parameters often result in insufficient multi-step reasoning capabilities required for solving complex medical problems. To address this, we introduce Meerkat-7B, a novel medical AI system with 7 billion parameters. Meerkat-7B was trained using our new synthetic dataset consisting of high-quality chain-of-thought reasoning paths sourced from 18 medical textbooks, along with diverse instruction-following datasets. Our system achieved remarkable accuracy across seven medical benchmarks, surpassing GPT-3.5 by 13.1%, as well as outperforming the previous best 7B models such as MediTron-7B and BioMistral-7B by 13.4% and 9.8%, respectively. Notably, it surpassed the passing threshold of the United States Medical Licensing Examination (USMLE) for the first time for a 7B-parameter model. Additionally, our system offered more detailed free-form responses to clinical queries compared to existing 7B and 13B models, approaching the performance level of GPT-3.5. This significantly narrows the performance gap with large LMs, showcasing its effectiveness in addressing complex medical challenges.
Edinburgh Clinical NLP at SemEval-2024 Task 2: Fine-tune your model unless you have access to GPT-4
Authors: Aryo Pradipta Gema, Giwon Hong, Pasquale Minervini, Luke Daines, Beatrice Alex
Abstract
The NLI4CT task assesses Natural Language Inference systems in predicting whether hypotheses entail or contradict evidence from Clinical Trial Reports. In this study, we evaluate various Large Language Models (LLMs) with multiple strategies, including Chain-of-Thought, In-Context Learning, and Parameter-Efficient Fine-Tuning (PEFT). We propose a PEFT method to improve the consistency of LLMs by merging adapters that were fine-tuned separately using triplet and language modelling objectives. We found that merging the two PEFT adapters improves the F1 score (+0.0346) and consistency (+0.152) of the LLMs. However, our novel methods did not produce more accurate results than GPT-4 in terms of faithfulness and consistency. Averaging the three metrics, GPT-4 ranks joint-first in the competition with 0.8328. Finally, our contamination analysis with GPT-4 indicates that there was no test data leakage.
Extracting Social Determinants of Health from Pediatric Patient Notes Using Large Language Models: Novel Corpus and Methods
Authors: Yujuan Fu, Giridhar Kaushik Ramachandran, Nicholas J Dobbins, Namu Park, Michael Leu, Abby R. Rosenberg, Kevin Lybarger, Fei Xia, Ozlem Uzuner, Meliha Yetisgen
Abstract
Social determinants of health (SDoH) play a critical role in shaping health outcomes, particularly in pediatric populations where interventions can have long-term implications. SDoH are frequently studied in the Electronic Health Record (EHR), which provides a rich repository for diverse patient data. In this work, we present a novel annotated corpus, the Pediatric Social History Annotation Corpus (PedSHAC), and evaluate the automatic extraction of detailed SDoH representations using fine-tuned and in-context learning methods with Large Language Models (LLMs). PedSHAC comprises annotated social history sections from 1,260 clinical notes obtained from pediatric patients within the University of Washington (UW) hospital system. Employing an event-based annotation scheme, PedSHAC captures ten distinct health determinants to encompass living and economic stability, prior trauma, education access, substance use history, and mental health with an overall annotator agreement of 81.9 F1. Our proposed fine-tuning LLM-based extractors achieve high performance at 78.4 F1 for event arguments. In-context learning approaches with GPT-4 demonstrate promise for reliable SDoH extraction with limited annotated examples, with extraction performance at 82.3 F1 for event triggers.
LLM-RadJudge: Achieving Radiologist-Level Evaluation for X-Ray Report Generation
Abstract
Evaluating generated radiology reports is crucial for the development of radiology AI, but existing metrics fail to reflect the task's clinical requirements. This study proposes a novel evaluation framework using large language models (LLMs) to compare radiology reports for assessment. We compare the performance of various LLMs and demonstrate that, when using GPT-4, our proposed metric achieves evaluation consistency close to that of radiologists. Furthermore, to reduce costs and improve accessibility, making this method practical, we construct a dataset using LLM evaluation results and perform knowledge distillation to train a smaller model. The distilled model achieves evaluation capabilities comparable to GPT-4. Our framework and distilled model offer an accessible and efficient evaluation method for radiology report generation, facilitating the development of more clinically relevant models. The model will be further open-sourced and accessible.
Large Language Models are Capable of Offering Cognitive Reappraisal, if Guided
Authors: Hongli Zhan, Allen Zheng, Yoon Kyung Lee, Jina Suh, Junyi Jessy Li, Desmond C. Ong
Abstract
Large language models (LLMs) have offered new opportunities for emotional support, and recent work has shown that they can produce empathic responses to people in distress. However, long-term mental well-being requires emotional self-regulation, where a one-time empathic response falls short. This work takes a first step by engaging with cognitive reappraisals, a strategy from psychology practitioners that uses language to targetedly change negative appraisals that an individual makes of the situation; such appraisals is known to sit at the root of human emotional experience. We hypothesize that psychologically grounded principles could enable such advanced psychology capabilities in LLMs, and design RESORT which consists of a series of reappraisal constitutions across multiple dimensions that can be used as LLM instructions. We conduct a first-of-its-kind expert evaluation (by clinical psychologists with M.S. or Ph.D. degrees) of an LLM's zero-shot ability to generate cognitive reappraisal responses to medium-length social media messages asking for support. This fine-grained evaluation showed that even LLMs at the 7B scale guided by RESORT are capable of generating empathic responses that can help users reappraise their situations.
TWIN-GPT: Digital Twins for Clinical Trials via Large Language Model
Authors: Yue Wang, Yingzhou Lu, Yinlong Xu, Zihan Ma, Hongxia Xu, Bang Du, Honghao Gao, Jian Wu
Subjects: Machine Learning (cs.LG); Computation and Language (cs.CL); Methodology (stat.ME)
Abstract
Recently, there has been a burgeoning interest in virtual clinical trials, which simulate real-world scenarios and hold the potential to significantly enhance patient safety, expedite development, reduce costs, and contribute to the broader scientific knowledge in healthcare. Existing research often focuses on leveraging electronic health records (EHRs) to support clinical trial outcome prediction. Yet, trained with limited clinical trial outcome data, existing approaches frequently struggle to perform accurate predictions. Some research has attempted to generate EHRs to augment model development but has fallen short in personalizing the generation for individual patient profiles. Recently, the emergence of large language models has illuminated new possibilities, as their embedded comprehensive clinical knowledge has proven beneficial in addressing medical issues. In this paper, we propose a large language model-based digital twin creation approach, called TWIN-GPT. TWIN-GPT can establish cross-dataset associations of medical information given limited data, generating unique personalized digital twins for different patients, thereby preserving individual patient characteristics. Comprehensive experiments show that using digital twins created by TWIN-GPT can boost clinical trial outcome prediction, exceeding various previous prediction approaches. Besides, we also demonstrate that TWIN-GPT can generate high-fidelity trial data that closely approximate specific patients, aiding in more accurate result predictions in data-scarce situations. Moreover, our study provides practical evidence for the application of digital twins in healthcare, highlighting its potential significance.
Keyword: biomedical
On-the-fly Definition Augmentation of LLMs for Biomedical NER
Authors: Monica Munnangi, Sergey Feldman, Byron C Wallace, Silvio Amir, Tom Hope, Aakanksha Naik
Abstract
Despite their general capabilities, LLMs still struggle on biomedical NER tasks, which are difficult due to the presence of specialized terminology and lack of training data. In this work we set out to improve LLM performance on biomedical NER in limited data settings via a new knowledge augmentation approach which incorporates definitions of relevant concepts on-the-fly. During this process, to provide a test bed for knowledge augmentation, we perform a comprehensive exploration of prompting strategies. Our experiments show that definition augmentation is useful for both open source and closed LLMs. For example, it leads to a relative improvement of 15\% (on average) in GPT-4 performance (F1) across all (six) of our test datasets. We conduct extensive ablations and analyses to demonstrate that our performance improvements stem from adding relevant definitional knowledge. We find that careful prompting strategies also improve LLM performance, allowing them to outperform fine-tuned language models in few-shot settings. To facilitate future research in this direction, we release our code at https://github.com/allenai/beacon.
Keyword: radiology
There is no result
Keyword: radiography
There is no result
Keyword: medical
Classification and Clustering of Sentence-Level Embeddings of Scientific Articles Generated by Contrastive Learning
Authors: Gustavo Bartz Guedes, Ana Estela Antunes da Silva
Abstract
Scientific articles are long text documents organized into sections, each describing aspects of the research. Analyzing scientific production has become progressively challenging due to the increase in the number of available articles. Within this scenario, our approach consisted of fine-tuning transformer language models to generate sentence-level embeddings from scientific articles, considering the following labels: background, objective, methods, results, and conclusion. We trained our models on three datasets with contrastive learning. Two datasets are from the article's abstracts in the computer science and medical domains. Also, we introduce PMC-Sents-FULL, a novel dataset of sentences extracted from the full texts of medical articles. We compare the fine-tuned and baseline models in clustering and classification tasks to evaluate our approach. On average, clustering agreement measures values were five times higher. For the classification measures, in the best-case scenario, we had an average improvement in F1-micro of 30.73\%. Results show that fine-tuning sentence transformers with contrastive learning and using the generated embeddings in downstream tasks is a feasible approach to sentence classification in scientific articles. Our experiment codes are available on GitHub.
Keyword: active-learning
There is no result
Keyword: active learning
There is no result
Keyword: chexpert
There is no result
Keyword: vision
From Robustness to Improved Generalization and Calibration in Pre-trained Language Models
Abstract
Enhancing generalization and uncertainty quantification in pre-trained language models (PLMs) is crucial for their effectiveness and reliability. Building on machine learning research that established the importance of robustness for improving generalization, we investigate the role of representation smoothness, achieved via Jacobian and Hessian regularization, in enhancing PLM performance. Although such regularization methods have proven effective in computer vision, their application in natural language processing (NLP), where PLM inputs are derived from a discrete domain, poses unique challenges. We introduce a novel two-phase regularization approach, JacHess, which minimizes the norms of the Jacobian and Hessian matrices within PLM intermediate representations relative to their inputs. Our evaluation using the GLUE benchmark demonstrates that JacHess significantly improves in-domain generalization and calibration in PLMs, outperforming unregularized fine-tuning and other similar regularization methods.
SentiCSE: A Sentiment-aware Contrastive Sentence Embedding Framework with Sentiment-guided Textual Similarity
Authors: Jaemin Kim, Yohan Na, Kangmin Kim, Sang Rak Lee, Dong-Kyu Chae
Abstract
Recently, sentiment-aware pre-trained language models (PLMs) demonstrate impressive results in downstream sentiment analysis tasks. However, they neglect to evaluate the quality of their constructed sentiment representations; they just focus on improving the fine-tuning performance, which overshadows the representation quality. We argue that without guaranteeing the representation quality, their downstream performance can be highly dependent on the supervision of the fine-tuning data rather than representation quality. This problem would make them difficult to foray into other sentiment-related domains, especially where labeled data is scarce. We first propose Sentiment-guided Textual Similarity (SgTS), a novel metric for evaluating the quality of sentiment representations, which is designed based on the degree of equivalence in sentiment polarity between two sentences. We then propose SentiCSE, a novel Sentiment-aware Contrastive Sentence Embedding framework for constructing sentiment representations via combined word-level and sentence-level objectives, whose quality is guaranteed by SgTS. Qualitative and quantitative comparison with the previous sentiment-aware PLMs shows the superiority of our work. Our code is available at: https://github.com/nayohan/SentiCSE
A General and Efficient Training for Transformer via Token Expansion
Authors: Wenxuan Huang, Yunhang Shen, Jiao Xie, Baochang Zhang, Gaoqi He, Ke Li, Xing Sun, Shaohui Lin
Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Computer Vision and Pattern Recognition (cs.CV)
Abstract
The remarkable performance of Vision Transformers (ViTs) typically requires an extremely large training cost. Existing methods have attempted to accelerate the training of ViTs, yet typically disregard method universality with accuracy dropping. Meanwhile, they break the training consistency of the original transformers, including the consistency of hyper-parameters, architecture, and strategy, which prevents them from being widely applied to different Transformer networks. In this paper, we propose a novel token growth scheme Token Expansion (termed ToE) to achieve consistent training acceleration for ViTs. We introduce an "initialization-expansion-merging" pipeline to maintain the integrity of the intermediate feature distribution of original transformers, preventing the loss of crucial learnable information in the training process. ToE can not only be seamlessly integrated into the training and fine-tuning process of transformers (e.g., DeiT and LV-ViT), but also effective for efficient training frameworks (e.g., EfficientTrain), without twisting the original training hyper-parameters, architecture, and introducing additional training strategies. Extensive experiments demonstrate that ToE achieves about 1.3x faster for the training of ViTs in a lossless manner, or even with performance gains over the full-token training baselines. Code is available at https://github.com/Osilly/TokenExpansion .
CausalChaos! Dataset for Comprehensive Causal Action Question Answering Over Longer Causal Chains Grounded in Dynamic Visual Scenes
Authors: Ting En Lam, Yuhan Chen, Elston Tan, Eric Peh, Ruirui Chen, Paritosh Parmar, Basura Fernando
Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG)
Abstract
Causal video question answering (QA) has garnered increasing interest, yet existing datasets often lack depth in causal reasoning analysis. To address this gap, we capitalize on the unique properties of cartoons and construct CausalChaos!, a novel, challenging causal Why-QA dataset built upon the iconic "Tom and Jerry" cartoon series. With thoughtful questions and multi-level answers, our dataset contains much longer causal chains embedded in dynamic interactions and visuals, at the same time principles of animation allows animators to create well-defined, unambiguous causal relationships. These factors allow models to solve more challenging, yet well-defined causal relationships. We also introduce hard negative mining, including CausalConfusion version. While models perform well, there is much room for improvement, especially, on open-ended answers. We identify more advanced/explicit causal relationship modeling and joint modeling of vision and language as the immediate areas for future efforts to focus upon. Along with the other complementary datasets, our new challenging dataset will pave the way for these developments in the field. We will release our dataset, codes, and models to help future efforts in this domain.
Keyword: visual
Noise-Aware Training of Layout-Aware Language Models
Authors: Ritesh Sarkhel, Xiaoqi Ren, Lauro Beltrao Costa, Guolong Su, Vincent Perot, Yanan Xie, Emmanouil Koukoumidis, Arnab Nandi
Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Abstract
A visually rich document (VRD) utilizes visual features along with linguistic cues to disseminate information. Training a custom extractor that identifies named entities from a document requires a large number of instances of the target document type annotated at textual and visual modalities. This is an expensive bottleneck in enterprise scenarios, where we want to train custom extractors for thousands of different document types in a scalable way. Pre-training an extractor model on unlabeled instances of the target document type, followed by a fine-tuning step on human-labeled instances does not work in these scenarios, as it surpasses the maximum allowable training time allocated for the extractor. We address this scenario by proposing a Noise-Aware Training method or NAT in this paper. Instead of acquiring expensive human-labeled documents, NAT utilizes weakly labeled documents to train an extractor in a scalable way. To avoid degradation in the model's quality due to noisy, weakly labeled samples, NAT estimates the confidence of each training sample and incorporates it as uncertainty measure during training. We train multiple state-of-the-art extractor models using NAT. Experiments on a number of publicly available and in-house datasets show that NAT-trained models are not only robust in performance -- it outperforms a transfer-learning baseline by up to 6% in terms of macro-F1 score, but it is also more label-efficient -- it reduces the amount of human-effort required to obtain comparable performance by up to 73%.
MIPS at SemEval-2024 Task 3: Multimodal Emotion-Cause Pair Extraction in Conversations with Multimodal Language Models
Abstract
This paper presents our winning submission to Subtask 2 of SemEval 2024 Task 3 on multimodal emotion cause analysis in conversations. We propose a novel Multimodal Emotion Recognition and Multimodal Emotion Cause Extraction (MER-MCE) framework that integrates text, audio, and visual modalities using specialized emotion encoders. Our approach sets itself apart from top-performing teams by leveraging modality-specific features for enhanced emotion understanding and causality inference. Experimental evaluation demonstrates the advantages of our multimodal approach, with our submission achieving a competitive weighted F1 score of 0.3435, ranking third with a margin of only 0.0339 behind the 1st team and 0.0025 behind the 2nd team. Project: https://github.com/MIPS-COLT/MER-MCE.git
AURORA: Navigating UI Tarpits via Automated Neural Screen Understanding
Authors: Safwat Ali Khan, Wenyu Wang, Yiran Ren, Bin Zhu, Jiangfan Shi, Alyssa McGowan, Wing Lam, Kevin Moran
Subjects: Software Engineering (cs.SE); Computation and Language (cs.CL); Computer Vision and Pattern Recognition (cs.CV); Human-Computer Interaction (cs.HC)
Abstract
Nearly a decade of research in software engineering has focused on automating mobile app testing to help engineers in overcoming the unique challenges associated with the software platform. Much of this work has come in the form of Automated Input Generation tools (AIG tools) that dynamically explore app screens. However, such tools have repeatedly been demonstrated to achieve lower-than-expected code coverage - particularly on sophisticated proprietary apps. Prior work has illustrated that a primary cause of these coverage deficiencies is related to so-called tarpits, or complex screens that are difficult to navigate. In this paper, we take a critical step toward enabling AIG tools to effectively navigate tarpits during app exploration through a new form of automated semantic screen understanding. We introduce AURORA, a technique that learns from the visual and textual patterns that exist in mobile app UIs to automatically detect common screen designs and navigate them accordingly. The key idea of AURORA is that there are a finite number of mobile app screen designs, albeit with subtle variations, such that the general patterns of different categories of UI designs can be learned. As such, AURORA employs a multi-modal, neural screen classifier that is able to recognize the most common types of UI screen designs. After recognizing a given screen, it then applies a set of flexible and generalizable heuristics to properly navigate the screen. We evaluated AURORA both on a set of 12 apps with known tarpits from prior work, and on a new set of five of the most popular apps from the Google Play store. Our results indicate that AURORA is able to effectively navigate tarpit screens, outperforming prior approaches that avoid tarpits by 19.6% in terms of method coverage. The improvements can be attributed to AURORA's UI design classification and heuristic navigation techniques.
IsoBench: Benchmarking Multimodal Foundation Models on Isomorphic Representations
Abstract
Current foundation models exhibit impressive capabilities when prompted either with text only or with both image and text inputs. But do their capabilities change depending on the input modality? In this work, we propose $\textbf{IsoBench}$, a benchmark dataset containing problems from four major areas: math, science, algorithms, and games. Each example is presented with multiple $\textbf{isomorphic representations}$ of inputs, such as visual, textual, and mathematical presentations. IsoBench provides fine-grained feedback to diagnose performance gaps caused by the form of the representation. Across various foundation models, we observe that on the same problem, models have a consistent preference towards textual representations. Most prominently, when evaluated on all IsoBench problems, Claude-3 Opus performs 28.7 points worse when provided with images instead of text; similarly, GPT-4 Turbo is 18.7 points worse and Gemini Pro is 14.9 points worse. Finally, we present two prompting techniques, $\textit{IsoCombination}$ and $\textit{IsoScratchPad}$, which improve model performance by considering combinations of, and translations between, different input representations.
Keyword: visio-linguistic
There is no result
Keyword: cross-modal
There is no result
Keyword: modality
UniMEEC: Towards Unified Multimodal Emotion Recognition and Emotion Cause
Abstract
Multimodal emotion recognition in conversation (MERC) and multimodal emotion-cause pair extraction (MECPE) has recently garnered significant attention. Emotions are the expression of affect or feelings; responses to specific events, thoughts, or situations are known as emotion causes. Both are like two sides of a coin, collectively describing human behaviors and intents. However, most existing works treat MERC and MECPE as separate tasks, which may result in potential challenges in integrating emotion and cause in real-world applications. In this paper, we propose a Unified Multimodal Emotion recognition and Emotion-Cause analysis framework (UniMEEC) to explore the causality and complementarity between emotion and emotion cause. Concretely, UniMEEC reformulates the MERC and MECPE tasks as two mask prediction problems, enhancing the interaction between emotion and cause. Meanwhile, UniMEEC shares the prompt learning among modalities for probing modality-specific knowledge from the Pre-trained model. Furthermore, we propose a task-specific hierarchical context aggregation to control the information flow to the task. Experiment results on four public benchmark datasets verify the model performance on MERC and MECPE tasks and achieve consistent improvements compared with state-of-the-art methods.
Scaling Properties of Speech Language Models
Authors: Santiago Cuervo, Ricard Marxer
Subjects: Audio and Speech Processing (eess.AS); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Neural and Evolutionary Computing (cs.NE)
Abstract
Speech Language Models (SLMs) aim to learn language from raw audio, without textual resources. Despite significant advances, our current models exhibit weak syntax and semantic abilities. However, if the scaling properties of neural language models hold for the speech modality, these abilities will improve as the amount of compute used for training increases. In this paper, we use models of this scaling behavior to estimate the scale at which our current methods will yield a SLM with the English proficiency of text-based Large Language Models (LLMs). We establish a strong correlation between pre-training loss and downstream syntactic and semantic performance in SLMs and LLMs, which results in predictable scaling of linguistic performance. We show that the linguistic performance of SLMs scales up to three orders of magnitude more slowly than that of text-based LLMs. Additionally, we study the benefits of synthetic data designed to boost semantic understanding and the effects of coarser speech tokenization.
Keyword: modalities
LITE: Modeling Environmental Ecosystems with Multimodal Large Language Models
Abstract
The modeling of environmental ecosystems plays a pivotal role in the sustainable management of our planet. Accurate prediction of key environmental variables over space and time can aid in informed policy and decision-making, thus improving people's livelihood. Recently, deep learning-based methods have shown promise in modeling the spatial-temporal relationships for predicting environmental variables. However, these approaches often fall short in handling incomplete features and distribution shifts, which are commonly observed in environmental data due to the substantial cost of data collection and malfunctions in measuring instruments. To address these issues, we propose LITE -- a multimodal large language model for environmental ecosystems modeling. Specifically, LITE unifies different environmental variables by transforming them into natural language descriptions and line graph images. Then, LITE utilizes unified encoders to capture spatial-temporal dynamics and correlations in different modalities. During this step, the incomplete features are imputed by a sparse Mixture-of-Experts framework, and the distribution shift is handled by incorporating multi-granularity information from past observations. Finally, guided by domain instructions, a language model is employed to fuse the multimodal representations for the prediction. Our experiments demonstrate that LITE significantly enhances performance in environmental spatial-temporal prediction across different domains compared to the best baseline, with a 41.25% reduction in prediction error. This justifies its effectiveness. Our data and code are available at https://github.com/hrlics/LITE.
An image speaks a thousand words, but can everyone listen? On translating images for cultural relevance
Authors: Simran Khanuja, Sathyanarayanan Ramamoorthy, Yueqi Song, Graham Neubig
Subjects: Computation and Language (cs.CL); Computer Vision and Pattern Recognition (cs.CV)
Abstract
Given the rise of multimedia content, human translators increasingly focus on culturally adapting not only words but also other modalities such as images to convey the same meaning. While several applications stand to benefit from this, machine translation systems remain confined to dealing with language in speech and text. In this work, we take a first step towards translating images to make them culturally relevant. First, we build three pipelines comprising state-of-the-art generative models to do the task. Next, we build a two-part evaluation dataset: i) concept: comprising 600 images that are cross-culturally coherent, focusing on a single concept per image, and ii) application: comprising 100 images curated from real-world applications. We conduct a multi-faceted human evaluation of translated images to assess for cultural relevance and meaning preservation. We find that as of today, image-editing models fail at this task, but can be improved by leveraging LLMs and retrievers in the loop. Best pipelines can only translate 5% of images for some countries in the easier concept dataset and no translation is successful for some countries in the application dataset, highlighting the challenging nature of the task. Our code and data is released here: https://github.com/simran-khanuja/image-transcreation.
Keyword: multi-modal
There is no result
Keyword: multimodal
Against The Achilles' Heel: A Survey on Red Teaming for Generative Models
Abstract
Generative models are rapidly gaining popularity and being integrated into everyday applications, raising concerns over their safety issues as various vulnerabilities are exposed. Faced with the problem, the field of red teaming is experiencing fast-paced growth, which highlights the need for a comprehensive organization covering the entire pipeline and addressing emerging topics for the community. Our extensive survey, which examines over 120 papers, introduces a taxonomy of fine-grained attack strategies grounded in the inherent capabilities of language models. Additionally, we have developed the searcher framework that unifies various automatic red teaming approaches. Moreover, our survey covers novel areas including multimodal attacks and defenses, risks around multilingual models, overkill of harmless queries, and safety of downstream applications. We hope this survey can provide a systematic perspective on the field and unlock new areas of research.
WavLLM: Towards Robust and Adaptive Speech Large Language Model
Abstract
The recent advancements in large language models (LLMs) have revolutionized the field of natural language processing, progressively broadening their scope to multimodal perception and generation. However, effectively integrating listening capabilities into LLMs poses significant challenges, particularly with respect to generalizing across varied contexts and executing complex auditory tasks. In this work, we introduce WavLLM, a robust and adaptive speech large language model with dual encoders, and a prompt-aware LoRA weight adapter, optimized by a two-stage curriculum learning approach. Leveraging dual encoders, we decouple different types of speech information, utilizing a Whisper encoder to process the semantic content of speech, and a WavLM encoder to capture the unique characteristics of the speaker's identity. Within the curriculum learning framework, WavLLM first builds its foundational capabilities by optimizing on mixed elementary single tasks, followed by advanced multi-task training on more complex tasks such as combinations of the elementary tasks. To enhance the flexibility and adherence to different tasks and instructions, a prompt-aware LoRA weight adapter is introduced in the second advanced multi-task training stage. We validate the proposed model on universal speech benchmarks including tasks such as ASR, ST, SV, ER, and also apply it to specialized datasets like Gaokao English listening comprehension set for SQA, and speech Chain-of-Thought (CoT) evaluation set. Experiments demonstrate that the proposed model achieves state-of-the-art performance across a range of speech tasks on the same model size, exhibiting robust generalization capabilities in executing complex tasks using CoT approach. Furthermore, our model successfully completes Gaokao tasks without specialized training. The codes, models, audio, and Gaokao evaluation set can be accessed at \url{aka.ms/wavllm}.
How Can Large Language Models Enable Better Socially Assistive Human-Robot Interaction: A Brief Survey
Authors: Zhonghao Shi, Ellen Landrum, Amy O' Connell, Mina Kian, Leticia Pinto-Alva, Kaleen Shrestha, Xiaoyuan Zhu, Maja J Matarić
Subjects: Human-Computer Interaction (cs.HC); Computation and Language (cs.CL); Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
Abstract
Socially assistive robots (SARs) have shown great success in providing personalized cognitive-affective support for user populations with special needs such as older adults, children with autism spectrum disorder (ASD), and individuals with mental health challenges. The large body of work on SAR demonstrates its potential to provide at-home support that complements clinic-based interventions delivered by mental health professionals, making these interventions more effective and accessible. However, there are still several major technical challenges that hinder SAR-mediated interactions and interventions from reaching human-level social intelligence and efficacy. With the recent advances in large language models (LLMs), there is an increased potential for novel applications within the field of SAR that can significantly expand the current capabilities of SARs. However, incorporating LLMs introduces new risks and ethical concerns that have not yet been encountered, and must be carefully be addressed to safely deploy these more advanced systems. In this work, we aim to conduct a brief survey on the use of LLMs in SAR technologies, and discuss the potentials and risks of applying LLMs to the following three major technical challenges of SAR: 1) natural language dialog; 2) multimodal understanding; 3) LLMs as robot policies.
Keyword: alignment
Is Factuality Decoding a Free Lunch for LLMs? Evaluation on Knowledge Editing Benchmark
DeFT: Flash Tree-attention with IO-Awareness for Efficient Tree-search-based LLM Inference
Dialectical Alignment: Resolving the Tension of 3H and Security Threats of LLMs
you may be attackedto the LLMs' context window.Comparing Bad Apples to Good Oranges: Aligning Large Language Models via Joint Preference Optimization
EvoCodeBench: An Evolving Code Generation Benchmark Aligned with Real-World Code Repositories
Extensive Self-Contrast Enables Feedback-Free Language Model Alignment
Bailong: Bilingual Transfer Learning based on QLoRA and Zip-tie Embedding
A Survey on Multilingual Large Language Models: Corpora, Alignment, and Bias
ChatGLM-RLHF: Practices of Aligning Large Language Models with Human Feedback
Prior Constraints-based Reward Model Training for Aligning Large Language Models
Generating Faithful and Complete Hospital-Course Summaries from the Electronic Health Record
UniArk: Improving Generalisation and Consistency for Factual Knowledge Extraction through Debiasing
Design as Desired: Utilizing Visual Question Answering for Multimodal Pre-training
LLaMA-Excitor: General Instruction Tuning via Indirect Feature Interaction
LLMs are Good Sign Language Translators
What's in Your "Safe" Data?: Identifying Benign Data that Breaks Safety
Evaluating Text-to-Visual Generation with Image-to-Text Generation
Keyword: aligning
Aurora-M: The First Open Source Multilingual Language Model Red-teamed according to the U.S. Executive Order
Regularized Best-of-N Sampling to Mitigate Reward Hacking for Language Model Alignment
Keyword: align
A Likelihood Ratio Test of Genetic Relationship among Languages
Token-Efficient Leverage Learning in Large Language Models
Evaluating the Factuality of Large Language Models using Large-Scale Knowledge Graphs
Enhancing Content-based Recommendation via Large Language Model
Keyword: vision language
There is no result
Keyword: vision-language
Open-Vocabulary Federated Learning with Multimodal Prototyping
Do Vision-Language Models Understand Compound Nouns?
Keyword: language-vision
There is no result
Keyword: phrase-grounding
There is no result
Keyword: phrase grounding
There is no result
Keyword: reference expression comprehension
There is no result
Keyword: chest
A Survey of using Large Language Models for Generating Infrastructure as Code
Keyword: x-ray
There is no result
Keyword: clinical
Small Language Models Learn Enhanced Reasoning Skills from Medical Textbooks
Edinburgh Clinical NLP at SemEval-2024 Task 2: Fine-tune your model unless you have access to GPT-4
Extracting Social Determinants of Health from Pediatric Patient Notes Using Large Language Models: Novel Corpus and Methods
LLM-RadJudge: Achieving Radiologist-Level Evaluation for X-Ray Report Generation
Large Language Models are Capable of Offering Cognitive Reappraisal, if Guided
TWIN-GPT: Digital Twins for Clinical Trials via Large Language Model
Keyword: biomedical
On-the-fly Definition Augmentation of LLMs for Biomedical NER
Keyword: radiology
There is no result
Keyword: radiography
There is no result
Keyword: medical
Classification and Clustering of Sentence-Level Embeddings of Scientific Articles Generated by Contrastive Learning
Keyword: active-learning
There is no result
Keyword: active learning
There is no result
Keyword: chexpert
There is no result
Keyword: vision
From Robustness to Improved Generalization and Calibration in Pre-trained Language Models
SentiCSE: A Sentiment-aware Contrastive Sentence Embedding Framework with Sentiment-guided Textual Similarity
A General and Efficient Training for Transformer via Token Expansion
CausalChaos! Dataset for Comprehensive Causal Action Question Answering Over Longer Causal Chains Grounded in Dynamic Visual Scenes
Keyword: visual
Noise-Aware Training of Layout-Aware Language Models
MIPS at SemEval-2024 Task 3: Multimodal Emotion-Cause Pair Extraction in Conversations with Multimodal Language Models
AURORA: Navigating UI Tarpits via Automated Neural Screen Understanding
IsoBench: Benchmarking Multimodal Foundation Models on Isomorphic Representations
Keyword: visio-linguistic
There is no result
Keyword: cross-modal
There is no result
Keyword: modality
UniMEEC: Towards Unified Multimodal Emotion Recognition and Emotion Cause
Scaling Properties of Speech Language Models
Keyword: modalities
LITE: Modeling Environmental Ecosystems with Multimodal Large Language Models
An image speaks a thousand words, but can everyone listen? On translating images for cultural relevance
Keyword: multi-modal
There is no result
Keyword: multimodal
Against The Achilles' Heel: A Survey on Red Teaming for Generative Models
WavLLM: Towards Robust and Adaptive Speech Large Language Model
How Can Large Language Models Enable Better Socially Assistive Human-Robot Interaction: A Brief Survey