Authors: David R. Bellamy, Bhawesh Kumar, Cindy Wang, Andrew Beam
In this work we introduce Labrador, a pre-trained Transformer model for laboratory data. Labrador and BERT were pre-trained on a corpus of 100 million lab test results from electronic health records (EHRs) and evaluated on various downstream outcome prediction tasks. Both models demonstrate mastery of the pre-training task but neither consistently outperform XGBoost on downstream supervised tasks. Our ablation studies reveal that transfer learning shows limited effectiveness for BERT and achieves marginal success with Labrador. We explore the reasons for the failure of transfer learning and suggest that the data generating process underlying each patient cannot be characterized sufficiently using labs alone, among other factors. We encourage future work to focus on joint modeling of multiple EHR data categories and to include tree-based baselines in their evaluations.
Authors: Varun Sharma
Affective computing is a field of study that focuses on developing systems and technologies that can understand, interpret, and respond to human emotions. Speech Emotion Recognition (SER), in particular, has got a lot of attention from researchers in the recent past. However, in many cases, the publicly available datasets, used for training and evaluation, are scarce and imbalanced across the emotion labels. In this work, we focused on building a balanced corpus from these publicly available datasets by combining these datasets as well as employing various speech data augmentation techniques. Furthermore, we experimented with different architectures for speech emotion recognition. Our best system, a multi-modal speech, and text-based model, provides a performance of UA(Unweighed Accuracy) + WA (Weighed Accuracy) of 157.57 compared to the baseline algorithm performance of 119.66
Authors: Hao Li, Brandon Bennett
Text classification is an important topic in the field of natural language processing. It has been preliminarily applied in information retrieval, digital library, automatic abstracting, text filtering, word semantic discrimination and many other fields. The aim of this research is to use a variety of algorithms to test the ability to identify offensive posts and evaluate their performance against a variety of assessment methods. The motivation for this project is to reduce the harm of these languages to human censors by automating the screening of offending posts. The field is a new one, and despite much interest in the past two years, there has been no focus on the object of the offence. Through the experiment of this project, it should inspire future research on identification methods as well as identification content.
Authors: Yang Xu, Yongqiang Yao, Yufan Huang, Mengnan Qi, Maoquan Wang, Bin Gu, Neel Sundaresan
Instruction fine-tuning, involving the refinement of pre-trained LLMs using datasets accompanied by natural instructions, is a powerful approach. However, its effectiveness is hindered by the redundancy and deficiencies in LLM-generated instruction datasets. In this paper, we introduce a highly effective and versatile paradigm for selecting diverse and high-quality instruction-following data from fine-tuning datasets. We first employ the dataset enhancement and expansion to augment the dataset with more diverse and high-quality data, then we apply variety compression and quality compression sequentially to curate the desired dataset. Our experimental results showcase that, even with a limited quantity of high-quality instruction data, LLMs consistently maintain robust performance across both natural language understanding tasks and code generation tasks. Notably, they outperform models trained on significantly larger instruction datasets in certain instances.
Authors: Pavlos Constas, Vikram Rawal, Matthew Honorio Oliveira, Andreas Constas, Aditya Khan, Kaison Cheung, Najma Sultani, Carrie Chen, Micol Altomare, Michael Akzam, Jiacheng Chen, Vhea He, Lauren Altomare, Heraa Murqi, Asad Khan, Nimit Amikumar Bhanshali, Youssef Rachad, Michael Guerzhoy
We propose a Reinforcement-Learning-based system that would automatically prescribe a hypothetical patient medications that may help the patient with their mental-health-related speech disfluency, and adjust the medication and the dosages in response to data from the patient. We demonstrate the components of the system: a module that detects and evaluates speech disfluency on a large dataset we built, and a Reinforcement Learning algorithm that automatically finds good combinations of medications. To support the two modules, we collect data on the effect of psychiatric medications for speech disfluency from the literature, and build a plausible patient simulation system. We demonstrate that the Reinforcement Learning system is, under some circumstances, able to converge to a good medication regime. We collect and label a dataset of people with possible speech disfluency and demonstrate our methods using that dataset. Our work is a proof of concept: we show that there is promise in the idea of using automatic data collection to address disfluency.
Authors: Henry Bae, Aghyad Deeb, Alex Fleury, Kehang Zhu
We present ComplexityNet, a streamlined language model designed for assessing task complexity. This model predicts the likelihood of accurate output by various language models, each with different capabilities. Our initial application of ComplexityNet involves the Mostly Basic Python Problems (MBPP) dataset. We pioneered the creation of the first set of labels to define task complexity. ComplexityNet achieved a notable 79% accuracy in determining task complexity, a significant improvement over the 34% accuracy of the original, non fine-tuned model. Furthermore, ComplexityNet effectively reduces computational resource usage by 90% compared to using the highest complexity model, while maintaining a high code generation accuracy of 86.7%. This study demonstrates that fine-tuning smaller models to categorize tasks based on their complexity can lead to a more balanced trade-off between accuracy and efficiency in the use of Large Language Models. Our findings suggest a promising direction for optimizing LLM applications, especially in resource-constrained environments.
Authors: Keivan Alizadeh, Iman Mirzadeh, Dmitry Belenko, Karen Khatamifard, Minsik Cho, Carlo C Del Mundo, Mohammad Rastegari, Mehrdad Farajtabar
Large language models (LLMs) are central to modern natural language processing, delivering exceptional performance in various tasks. However, their intensive computational and memory requirements present challenges, especially for devices with limited DRAM capacity. This paper tackles the challenge of efficiently running LLMs that exceed the available DRAM capacity by storing the model parameters on flash memory but bringing them on demand to DRAM. Our method involves constructing an inference cost model that harmonizes with the flash memory behavior, guiding us to optimize in two critical areas: reducing the volume of data transferred from flash and reading data in larger, more contiguous chunks. Within this flash memory-informed framework, we introduce two principal techniques. First, "windowing'" strategically reduces data transfer by reusing previously activated neurons, and second, "row-column bundling", tailored to the sequential data access strengths of flash memory, increases the size of data chunks read from flash memory. These methods collectively enable running models up to twice the size of the available DRAM, with a 4-5x and 20-25x increase in inference speed compared to naive loading approaches in CPU and GPU, respectively. Our integration of sparsity awareness, context-adaptive loading, and a hardware-oriented design paves the way for effective inference of LLMs on devices with limited memory.
Authors: Md Abrar Jahin, Subrata Talapatra
This research delves into the intricate landscape of Musculoskeletal Disorder (MSD) risk factors, employing a novel fusion of Natural Language Processing (NLP) techniques and mode-based ranking methodologies. The primary objective is to advance the comprehension of MSD risk factors, their classification, and their relative severity, facilitating more targeted preventive and management interventions. The study utilizes eight diverse models, integrating pre-trained transformers, cosine similarity, and various distance metrics to classify risk factors into personal, biomechanical, workplace, psychological, and organizational classes. Key findings reveal that the BERT model with cosine similarity attains an overall accuracy of 28\%, while the sentence transformer, coupled with Euclidean, Bray-Curtis, and Minkowski distances, achieves a flawless accuracy score of 100\%. In tandem with the classification efforts, the research employs a mode-based ranking approach on survey data to discern the severity hierarchy of MSD risk factors. Intriguingly, the rankings align precisely with the previous literature, reaffirming the consistency and reliability of the approach. ``Working posture" emerges as the most severe risk factor, emphasizing the critical role of proper posture in preventing MSDs. The collective perceptions of survey participants underscore the significance of factors like ``Job insecurity," ``Effort reward imbalance," and ``Poor employee facility" in contributing to MSD risks. The convergence of rankings provides actionable insights for organizations aiming to reduce the prevalence of MSDs. The study concludes with implications for targeted interventions, recommendations for improving workplace conditions, and avenues for future research.
Authors: Zhaoxuan Tan, Meng Jiang
User modeling (UM) aims to discover patterns or learn representations from user data about the characteristics of a specific user, such as profile, preference, and personality. The user models enable personalization and suspiciousness detection in many online applications such as recommendation, education, and healthcare. Two common types of user data are text and graph, as the data usually contain a large amount of user-generated content (UGC) and online interactions. The research of text and graph mining is developing rapidly, contributing many notable solutions in the past two decades. Recently, large language models (LLMs) have shown superior performance on generating, understanding, and even reasoning over text data. The approaches of user modeling have been equipped with LLMs and soon become outstanding. This article summarizes existing research about how and why LLMs are great tools of modeling and understanding UGC. Then it reviews a few categories of large language models for user modeling (LLM-UM) approaches that integrate the LLMs with text and graph-based methods in different ways. Then it introduces specific LLM-UM techniques for a variety of UM applications. Finally, it presents remaining challenges and future directions in the LLM-UM research. We maintain the reading list at: https://github.com/TamSiuhin/LLM-UM-Reading
Authors: Bowen Zhao, Changkai Ji, Yuejie Zhang, Wen He, Yingwen Wang, Qing Wang, Rui Feng, Xiaobo Zhang
With the Generative Pre-trained Transformer 3.5 (GPT-3.5) exhibiting remarkable reasoning and comprehension abilities in Natural Language Processing (NLP), most Question Answering (QA) research has primarily centered around general QA tasks based on GPT, neglecting the specific challenges posed by Complex Table QA. In this paper, we propose to incorporate GPT-3.5 to address such challenges, in which complex tables are reconstructed into tuples and specific prompt designs are employed for dialogues. Specifically, we encode each cell's hierarchical structure, position information, and content as a tuple. By enhancing the prompt template with an explanatory description of the meaning of each tuple and the logical reasoning process of the task, we effectively improve the hierarchical structure awareness capability of GPT-3.5 to better parse the complex tables. Extensive experiments and results on Complex Table QA datasets, i.e., the open-domain dataset HiTAB and the aviation domain dataset AIT-QA show that our approach significantly outperforms previous work on both datasets, leading to state-of-the-art (SOTA) performance.
Authors: Xinpeng Wang, Xiaoyuan Yi, Han Jiang, Shanlin Zhou, Zhihua Wei, Xing Xie
Warning: this paper includes model outputs showing offensive content. Recent large-scale Visual-Language Generative Models (VLGMs) have achieved unprecedented improvement in multimodal image/text generation. However, these models might also generate toxic content, e.g., offensive text and pornography images, raising significant ethical risks. Despite exhaustive studies on toxic degeneration of language models, this problem remains largely unexplored within the context of visual-language generation. This work delves into the propensity for toxicity generation and susceptibility to toxic data across various VLGMs. For this purpose, we built ToViLaG, a dataset comprising 32K co-toxic/mono-toxic text-image pairs and 1K innocuous but evocative text that tends to stimulate toxicity. Furthermore, we propose WInToRe, a novel toxicity metric tailored to visual-language generation, which theoretically reflects different aspects of toxicity considering both input and output. On such a basis, we benchmarked the toxicity of a diverse spectrum of VLGMs and discovered that some models do more evil than expected while some are more vulnerable to infection, underscoring the necessity of VLGMs detoxification. Therefore, we develop an innovative bottleneck-based detoxification method. Our method could reduce toxicity while maintaining comparable generation quality, providing a promising initial solution to this line of research.
Authors: Mukul Singh, José Cambronero, Sumit Gulwani, Vu Le, Gust Verbruggen
Multi-modality promises to unlock further uses for large language models. Recently, the state-of-the-art language model GPT-4 was enhanced with vision capabilities. We carry out a prompting evaluation of GPT-4V and five other baselines on structured reasoning tasks, such as mathematical reasoning, visual data analysis, and code generation. We show that visual Chain-of-Thought, an extension of Chain-of-Thought to multi-modal LLMs, yields significant improvements over the vanilla model. We also present a categorized analysis of scenarios where these models perform well and where they struggle, highlighting challenges associated with coherent multimodal reasoning.
Authors: YoungJoon Yoo, Jongwon Choi
This paper introduces a novel approach for topic modeling utilizing latent codebooks from Vector-Quantized Variational Auto-Encoder~(VQ-VAE), discretely encapsulating the rich information of the pre-trained embeddings such as the pre-trained language model. From the novel interpretation of the latent codebooks and embeddings as conceptual bag-of-words, we propose a new generative topic model called Topic-VQ-VAE~(TVQ-VAE) which inversely generates the original documents related to the respective latent codebook. The TVQ-VAE can visualize the topics with various generative distributions including the traditional BoW distribution and the autoregressive image generation. Our experimental results on document analysis and image generation demonstrate that TVQ-VAE effectively captures the topic context which reveals the underlying structures of the dataset and supports flexible forms of document generation. Official implementation of the proposed TVQ-VAE is available at https://github.com/clovaai/TVQ-VAE.
Authors: Shangshang Zheng, He Bai, Yizhe Zhang, Yi Su, Xiaochuan Niu, Navdeep Jaitly
Current approaches to evaluating large language models (LLMs) with pre-existing Knowledge Graphs (KG) mostly ignore the structure of the KG and make arbitrary choices of which part of the graph to evaluate. In this paper, we introduce KGLens, a method to evaluate LLMs by generating natural language questions from a KG in a structure aware manner so that we can characterize its performance on a more aggregated level. KGLens uses a parameterized KG, where each edge is augmented with a beta distribution that guides how to sample edges from the KG for QA testing. As the evaluation proceeds, different edges of the parameterized KG are sampled and assessed appropriately, converging to a more global picture of the performance of the LLMs on the KG as a whole. In our experiments, we construct three domain-specific KGs for knowledge assessment, comprising over 19,000 edges, 700 relations, and 21,000 entities. The results demonstrate that KGLens can not only assess overall performance but also provide topic, temporal, and relation analyses of LLMs. This showcases the adaptability and customizability of KGLens, emphasizing its ability to focus the evaluation based on specific criteria.
Authors: Akash Ghosh, Arkadeep Acharya, Raghav Jain, Sriparna Saha, Aman Chadha, Setu Sinha
In the era of modern healthcare, swiftly generating medical question summaries is crucial for informed and timely patient care. Despite the increasing complexity and volume of medical data, existing studies have focused solely on text-based summarization, neglecting the integration of visual information. Recognizing the untapped potential of combining textual queries with visual representations of medical conditions, we introduce the Multimodal Medical Question Summarization (MMQS) Dataset. This dataset, a major contribution to our work, pairs medical queries with visual aids, facilitating a richer and more nuanced understanding of patient needs. We also propose a framework, utilizing the power of Contrastive Language Image Pretraining(CLIP) and Large Language Models(LLMs), consisting of four modules that identify medical disorders, generate relevant context, filter medical concepts, and craft visually aware summaries. Our comprehensive framework harnesses the power of CLIP, a multimodal foundation model, and various general-purpose LLMs, comprising four main modules: the medical disorder identification module, the relevant context generation module, the context filtration module for distilling relevant medical concepts and knowledge, and finally, a general-purpose LLM to generate visually aware medical question summaries. Leveraging our MMQS dataset, we showcase how visual cues from images enhance the generation of medically nuanced summaries. This multimodal approach not only enhances the decision-making process in healthcare but also fosters a more nuanced understanding of patient queries, laying the groundwork for future research in personalized and responsive medical care
Authors: Ying-Ying Chang, Wei-Yao Wang, Wen-Chih Peng
In the dynamic and rapidly evolving world of social media, detecting anomalous users has become a crucial task to address malicious activities such as misinformation and cyberbullying. As the increasing number of anomalous users improves the ability to mimic normal users and evade detection, existing methods only focusing on bot detection are ineffective in terms of capturing subtle distinctions between users. To address these challenges, we proposed SeGA, preference-aware self-contrastive learning for anomalous user detection, which leverages heterogeneous entities and their relations in the Twittersphere to detect anomalous users with different malicious strategies. SeGA utilizes the knowledge of large language models to summarize user preferences via posts. In addition, integrating user preferences with prompts as pseudo-labels for preference-aware self-contrastive learning enables the model to learn multifaceted aspects for describing the behaviors of users. Extensive experiments on the proposed TwBNT benchmark demonstrate that SeGA significantly outperforms the state-of-the-art methods (+3.5\% ~ 27.6\%) and empirically validate the effectiveness of the model design and pre-training strategies. Our code and data are publicly available at https://github.com/ying0409/SeGA.
Authors: Yu Wang, Zexue He, Zhankui He, Hao Xu, Julian McAuley
Understanding and accurately explaining compatibility relationships between fashion items is a challenging problem in the burgeoning domain of AI-driven outfit recommendations. Present models, while making strides in this area, still occasionally fall short, offering explanations that can be elementary and repetitive. This work aims to address these shortcomings by introducing the Pair Fashion Explanation (PFE) dataset, a unique resource that has been curated to illuminate these compatibility relationships. Furthermore, we propose an innovative two-stage pipeline model that leverages this dataset. This fine-tuning allows the model to generate explanations that convey the compatibility relationships between items. Our experiments showcase the model's potential in crafting descriptions that are knowledgeable, aligned with ground-truth matching correlations, and that produce understandable and informative descriptions, as assessed by both automatic metrics and human evaluation. Our code and data are released at https://github.com/wangyu-ustc/PairFashionExplanation
Authors: Juan A. Rodriguez, Shubham Agarwal, Issam H. Laradji, Pau Rodriguez, David Vazquez, Christopher Pal, Marco Pedersoli
Scalable Vector Graphics (SVGs) have become integral in modern image rendering applications due to their infinite scalability in resolution, versatile usability, and editing capabilities. SVGs are particularly popular in the fields of web development and graphic design. Existing approaches for SVG modeling using deep learning often struggle with generating complex SVGs and are restricted to simpler ones that require extensive processing and simplification. This paper introduces StarVector, a multimodal SVG generation model that effectively integrates Code Generation Large Language Models (CodeLLMs) and vision models. Our approach utilizes a CLIP image encoder to extract visual representations from pixel-based images, which are then transformed into visual tokens via an adapter module. These visual tokens are pre-pended to the SVG token embeddings, and the sequence is modeled by the StarCoder model using next-token prediction, effectively learning to align the visual and code tokens. This enables StarVector to generate unrestricted SVGs that accurately represent pixel images. To evaluate StarVector's performance, we present SVG-Bench, a comprehensive benchmark for evaluating SVG methods across multiple datasets and relevant metrics. Within this benchmark, we introduce novel datasets including SVG-Stack, a large-scale dataset of real-world SVG examples, and use it to pre-train StarVector as a large foundation model for SVGs. Our results demonstrate significant enhancements in visual quality and complexity handling over current methods, marking a notable advancement in SVG generation technology. Code and models: https://github.com/joanrod/star-vector
Authors: Jiankai Sun, Chuanyang Zheng, Enze Xie, Zhengying Liu, Ruihang Chu, Jianing Qiu, Jiaqi Xu, Mingyu Ding, Hongyang Li, Mengzhe Geng, Yue Wu, Wenhai Wang, Junsong Chen, Zhangyue Yin, Xiaozhe Ren, Jie Fu, Junxian He, Wu Yuan, Qi Liu, Xihui Liu, Yu Li, Hao Dong, Yu Cheng, Ming Zhang, Pheng Ann Heng, Jifeng Dai, Ping Luo, Jingdong Wang, Ji-Rong Wen, Xipeng Qiu, Yike Guo, Hui Xiong, Qun Liu, Zhenguo Li
Reasoning, a crucial ability for complex problem-solving, plays a pivotal role in various real-world settings such as negotiation, medical diagnosis, and criminal investigation. It serves as a fundamental methodology in the field of Artificial General Intelligence (AGI). With the ongoing development of foundation models, there is a growing interest in exploring their abilities in reasoning tasks. In this paper, we introduce seminal foundation models proposed or adaptable for reasoning, highlighting the latest advancements in various reasoning tasks, methods, and benchmarks. We then delve into the potential future directions behind the emergence of reasoning abilities within foundation models. We also discuss the relevance of multimodal learning, autonomous agents, and super alignment in the context of reasoning. By discussing these future research directions, we hope to inspire researchers in their exploration of this field, stimulate further advancements in reasoning with foundation models, and contribute to the development of AGI.
Authors: Md. Jalal Uddin Chowdhury, Ashab Hussan
This research paper presents a comprehensive review-based study on various Text-to-Speech (TTS) technologies. TTS technology is an important aspect of human-computer interaction, enabling machines to convert written text into audible speech. The paper examines the different TTS technologies available, including concatenative TTS, formant synthesis TTS, and statistical parametric TTS. The study focuses on comparing the advantages and limitations of these technologies in terms of their naturalness of voice, the level of complexity of the system, and their suitability for different applications. In addition, the paper explores the latest advancements in TTS technology, including neural TTS and hybrid TTS. The findings of this research will provide valuable insights for researchers, developers, and users who want to understand the different TTS technologies and their suitability for specific applications.
The most successful multi-domain text classification (MDTC) approaches employ the shared-private paradigm to facilitate the enhancement of domain-invariant features through domain-specific attributes. Additionally, they employ adversarial training to align marginal feature distributions. Nevertheless, these methodologies encounter two primary challenges: (1) Neglecting class-aware information during adversarial alignment poses a risk of misalignment; (2) The limited availability of labeled data across multiple domains fails to ensure adequate discriminative capacity for the model. To tackle these issues, we propose a method called Regularized Conditional Alignment (RCA) to align the joint distributions of domains and classes, thus matching features within the same category and amplifying the discriminative qualities of acquired features. Moreover, we employ entropy minimization and virtual adversarial training to constrain the uncertainty of predictions pertaining to unlabeled data and enhance the model's robustness. Empirical results on two benchmark datasets demonstrate that our RCA approach outperforms state-of-the-art MDTC techniques.
Authors: Megan Kinniment, Lucas Jun Koba Sato, Haoxing Du, Brian Goodrich, Max Hasin, Lawrence Chan, Luke Harold Miles, Tao R. Lin, Hjalmar Wijk, Joel Burget, Aaron Ho, Elizabeth Barnes, Paul Christiano
In this report, we explore the ability of language model agents to acquire resources, create copies of themselves, and adapt to novel challenges they encounter in the wild. We refer to this cluster of capabilities as "autonomous replication and adaptation" or ARA. We believe that systems capable of ARA could have wide-reaching and hard-to-anticipate consequences, and that measuring and forecasting ARA may be useful for informing measures around security, monitoring, and alignment. Additionally, once a system is capable of ARA, placing bounds on a system's capabilities may become significantly more difficult.
We construct four simple example agents that combine language models with tools that allow them to take actions in the world. We then evaluate these agents on 12 tasks relevant to ARA. We find that these language model agents can only complete the easiest tasks from this list, although they make some progress on the more challenging tasks. Unfortunately, these evaluations are not adequate to rule out the possibility that near-future agents will be capable of ARA. In particular, we do not think that these evaluations provide good assurance that the ``next generation'' of language models (e.g. 100x effective compute scaleup on existing models) will not yield agents capable of ARA, unless intermediate evaluations are performed during pretraining. Relatedly, we expect that fine-tuning of the existing models could produce substantially more competent agents, even if the fine-tuning is not directly targeted at ARA.
Authors: Madeleine Grunde-McLaughlin, Michelle S. Lam, Ranjay Krishna, Daniel S. Weld, Jeffrey Heer
LLM chains enable complex tasks by decomposing work into a sequence of sub-tasks. Crowdsourcing workflows similarly decompose complex tasks into smaller tasks for human crowdworkers. Chains address LLM errors analogously to the way crowdsourcing workflows address human error. To characterize opportunities for LLM chaining, we survey 107 papers across the crowdsourcing and chaining literature to construct a design space for chain development. The design space connects an LLM designer's objectives to strategies they can use to achieve those objectives, and tactics to implement each strategy. To explore how techniques from crowdsourcing may apply to chaining, we adapt crowdsourcing workflows to implement LLM chains across three case studies: creating a taxonomy, shortening text, and writing a short story. From the design space and our case studies, we identify which techniques transfer from crowdsourcing to LLM chaining and raise implications for future research and development.
Authors: Liang Zhang, Zhelun Chen
In recent years, the rapid advancement and impressive capabilities of Large Language Models (LLMs) have been evident across various domains. This paper explores the application, implications, and potential of LLMs in building energy efficiency and decarbonization studies. The wide-ranging capabilities of LLMs are examined in the context of the building energy field, including intelligent control systems, code generation, data infrastructure, knowledge extraction, and education. Despite the promising potential of LLMs, challenges including complex and expensive computation, data privacy, security and copyright, complexity in fine-tuned LLMs, and self-consistency are discussed. The paper concludes with a call for future research focused on the enhancement of LLMs for domain-specific tasks, multi-modal LLMs, and collaborative research between AI and energy experts.
Authors: Charles Rajan, Nishit Asnani, Shreya Singh
Multi-document summarization is the process of automatically generating a concise summary of multiple documents related to the same topic. This summary can help users quickly understand the key information from a large collection of documents. Multi-document summarization systems are more complex than single-document summarization systems due to the need to identify and combine information from multiple sources. In this paper, we have developed a machine learning model that generates a concise summary of a topic from multiple news documents. The model is designed to be unbiased by sampling its input equally from all the different aspects of the topic, even if the majority of the news sources lean one way.
Authors: Paulo Pirozelli, Marcos M. José, Paulo de Tarso P. Filho, Anarosa A. F. Brandão, Fabio G. Cozman
Logical reasoning is central to complex human activities, such as thinking, debating, and planning; it is also a central component of many AI systems as well. In this paper, we investigate the extent to which encoder-only transformer language models (LMs) can reason according to logical rules. We ask whether those LMs can deduce theorems in propositional calculus and first-order logic; if their relative success in these problems reflects general logical capabilities; and which layers contribute the most to the task. First, we show for several encoder-only LMs that they can be trained, to a reasonable degree, to determine logical validity on various datasets. Next, by cross-probing fine-tuned models on these datasets, we show that LMs have difficulty in transferring their putative logical reasoning ability, which suggests that they may have learned dataset-specific features, instead of a general capability. Finally, we conduct a layerwise probing experiment, which shows that the hypothesis classification task is mostly solved through higher layers.
Authors: Anaelia Ovalle, Ninareh Mehrabi, Palash Goyal, Jwala Dhamala, Kai-Wei Chang, Richard Zemel, Aram Galstyan, Yuval Pinter, Rahul Gupta
A large body of NLP research has documented the ways gender biases manifest and amplify within large language models (LLMs), though this research has predominantly operated within a gender binary-centric context. A growing body of work has identified the harmful limitations of this gender-exclusive framing; many LLMs cannot correctly and consistently refer to persons outside the gender binary, especially if they use neopronouns. While data scarcity has been identified as a possible culprit, the precise mechanisms through which it influences LLM misgendering remain underexplored. Our work addresses this gap by studying data scarcity's role in subword tokenization and, consequently, the formation of LLM word representations. We uncover how the Byte-Pair Encoding (BPE) tokenizer, a backbone for many popular LLMs, contributes to neopronoun misgendering through out-of-vocabulary behavior. We introduce pronoun tokenization parity (PTP), a novel approach to reduce LLM neopronoun misgendering by preserving a token's functional structure. We evaluate PTP's efficacy using pronoun consistency-based metrics and a novel syntax-based metric. Through several controlled experiments, finetuning LLMs with PTP improves neopronoun consistency from 14.5% to 58.4%, highlighting the significant role tokenization plays in LLM pronoun consistency.
Authors: Zhangdie Yuan, Andreas Vlachos
Despite progress in automated fact-checking, most systems require a significant amount of labeled training data, which is expensive. In this paper, we propose a novel zero-shot method, which instead of operating directly on the claim and evidence sentences, decomposes them into semantic triples augmented using external knowledge graphs, and uses large language models trained for natural language inference. This allows it to generalize to adversarial datasets and domains that supervised models require specific training data for. Our empirical results show that our approach outperforms previous zero-shot approaches on FEVER, FEVER-Symmetric, FEVER 2.0, and Climate-FEVER, while being comparable or better than supervised models on the adversarial and the out-of-domain datasets.
Authors: Yi Cheng, Wenge Liu, Jian Wang, Chak Tou Leong, Yi Ouyang, Wenjie Li, Xian Wu, Yefeng Zheng
In recent years, there has been a growing interest in exploring dialogues with more complex goals, such as negotiation, persuasion, and emotional support, which go beyond traditional service-focused dialogue systems. Apart from the requirement for much more sophisticated strategic reasoning and communication skills, a significant challenge of these tasks lies in the difficulty of objectively measuring the achievement of their goals in a quantifiable way, making it difficult for existing research to directly optimize the dialogue procedure towards them. In our work, we emphasize the multifaceted nature of complex dialogue goals and argue that it is more feasible to accomplish them by comprehensively considering and jointly promoting their different aspects. To this end, we propose a novel dialogue framework, Cooper, which coordinates multiple specialized agents, each dedicated to a specific dialogue goal aspect separately, to approach the complex objective. Through this divide-and-conquer manner, we make complex dialogue goals more approachable and elicit greater intelligence via the collaboration of individual agents. Experiments on persuasion and emotional support dialogues demonstrate the superiority of our method over a set of competitive baselines.
Authors: Lang Yu, Qin Chen, Jie Zhou, Liang He
Large language models (LLMs) have shown great success in various Natural Language Processing (NLP) tasks, whist they still need updates after deployment to fix errors or keep pace with the changing knowledge in the world. Researchers formulate such problem as Model Editing and have developed various editors focusing on different axes of editing properties. However, current editors can hardly support all properties and rely on heavy computational resources. In this paper, we propose a plug-in Model Editing method based on neuron-indexed dynamic LoRA (MELO), which alters the behavior of language models by dynamically activating certain LoRA blocks according to the index built in an inner vector database. Our method satisfies various editing properties with high efficiency and can be easily integrated into multiple LLM backbones. Experimental results show that our proposed MELO achieves state-of-the-art editing performance on three sequential editing tasks (document classification, question answering and hallucination correction), while requires the least trainable parameters and computational cost.
Authors: Maria Antoniak, Aakanksha Naik, Carla S. Alvarado, Lucy Lu Wang, Irene Y. Chen
Objective: An ethical framework for the use of large language models (LLMs) is urgently needed to shape how natural language processing (NLP) tools are used for healthcare applications. Drawing directly from the voices of those most affected, we propose a set of guiding principles for the use of NLP in healthcare, with examples based on applications in maternal health.
Materials and Methods: We led an interactive session centered on an LLM-based chatbot demonstration during a full-day workshop with 39 participants, and additionally surveyed 30 healthcare workers and 30 birthing people about their values, needs, and perceptions of AI and LLMs. We conducted quantitative and qualitative analyses of the interactive discussions to consolidate our findings into a set of guiding principles.
Results: Using the case study of maternal health, we propose nine principles for ethical use of LLMs, grouped into three categories: (i) contextual significance, (ii) measurements, and (iii) who/what is valued. We describe rationales underlying these principles and provide practical advice.
Discussion: Healthcare faces existing challenges including the balance of power in clinician-patient relationships, systemic health disparities, historical injustices, and economic constraints. Our principles serve as a framework for surfacing key considerations when deploying LLMs in medicine, as well as providing a methodological pattern for other researchers to follow.
Conclusion: This set of principles can serve as a resource to practitioners working on maternal health and other healthcare fields to emphasize the importance of technical nuance, historical context, and inclusive design when developing LLMs for use in clinical settings.
Authors: Gemini Team Google: Rohan Anil, Sebastian Borgeaud, Yonghui Wu, Jean-Baptiste Alayrac, Jiahui Yu, Radu Soricut, Johan Schalkwyk, Andrew M. Dai, Anja Hauth, Katie Millican, David Silver, Slav Petrov, Melvin Johnson, Ioannis Antonoglou, Julian Schrittwieser, Amelia Glaese, Jilin Chen, Emily Pitler, Timothy Lillicrap, Angeliki Lazaridou, Orhan Firat, James Molloy, Michael Isard, Paul R. Barham, Tom Hennigan, Benjamin Lee, Fabio Viola, Malcolm Reynolds, Yuanzhong Xu, Ryan Doherty, Eli Collins, Clemens Meyer, Eliza Rutherford, Erica Moreira, Kareem Ayoub, Megha Goel, George Tucker, Enrique Piqueras, Maxim Krikun, Iain Barr, Nikolay Savinov, Ivo Danihelka, Becca Roelofs, Anaïs White, Anders Andreassen, Tamara von Glehn, Lakshman Yagati, Mehran Kazemi, Lucas Gonzalez, Misha Khalman, Jakub Sygnowski, et al. (890 additional authors not shown)
This report introduces a new family of multimodal models, Gemini, that exhibit remarkable capabilities across image, audio, video, and text understanding. The Gemini family consists of Ultra, Pro, and Nano sizes, suitable for applications ranging from complex reasoning tasks to on-device memory-constrained use-cases. Evaluation on a broad range of benchmarks shows that our most-capable Gemini Ultra model advances the state of the art in 30 of 32 of these benchmarks - notably being the first model to achieve human-expert performance on the well-studied exam benchmark MMLU, and improving the state of the art in every one of the 20 multimodal benchmarks we examined. We believe that the new capabilities of Gemini models in cross-modal reasoning and language understanding will enable a wide variety of use cases and we discuss our approach toward deploying them responsibly to users.
Authors: Youshao Xiao, Weichang Wu, Zhenglei Zhou, Fagui Mao, Shangchun Zhao, Lin Ju, Lei Liang, Xiaolu Zhang, Jun Zhou
Recently, ChatGPT or InstructGPT like large language models (LLM) has made a significant impact in the AI world. These models are incredibly versatile, capable of performing language tasks on par or even exceeding the capabilities of human experts. Many works have attempted to reproduce the complex InstructGPT's RLHF (Reinforcement Learning with Human Feedback) training pipeline. However, the mainstream distributed RLHF training methods typically adopt a fixed model placement strategy, referred to as the Flattening strategy. This strategy treats all four models involved in RLHF as a single entity and places them on all devices, regardless of their differences. Unfortunately, this strategy exacerbates the generation bottlenecks in the RLHF training and degrades the overall training efficiency. To address these issues, we propose an adaptive model placement framework that offers two flexible model placement strategies. These strategies allow for the agile allocation of models across devices in a fine-grained manner. The Interleaving strategy helps reduce memory redundancy and communication costs during RLHF training. On the other hand, the Separation strategy improves the throughput of model training by separating the training and generation stages of the RLHF pipeline. Notably, this framework seamlessly integrates with other mainstream techniques for acceleration and enables automatic hyperparameter search. Extensive experiments have demonstrated that our Interleaving and Separation strategies can achieve notable improvements up to 11x, compared to the current state-of-the-art (SOTA) approaches. These experiments encompassed a wide range of training scenarios, involving models of varying sizes and devices of different scales. The results highlight the effectiveness and superiority of our approaches in accelerating the training of distributed RLHF.
Authors: Burak Aksar, Yara Rizk, Tathagata Chakraborti
Chatbots have become one of the main pathways for the delivery of business automation tools. Multi-agent systems offer a framework for designing chatbots at scale, making it easier to support complex conversations that span across multiple domains as well as enabling developers to maintain and expand their capabilities incrementally over time. However, multi-agent systems complicate the natural language understanding (NLU) of user intents, especially when they rely on decentralized NLU models: some utterances (termed single intent) may invoke a single agent while others (termed multi-intent) may explicitly invoke multiple agents. Without correctly parsing multi-intent inputs, decentralized NLU approaches will not achieve high prediction accuracy. In this paper, we propose an efficient parsing and orchestration pipeline algorithm to service multi-intent utterances from the user in the context of a multi-agent system. Our proposed approach achieved comparable performance to competitive deep learning models on three different datasets while being up to 48 times faster.
Authors: Zheng Wei Lim, Ekaterina Vylomova, Charles Kemp, Trevor Cohn
Human translators linger on some words and phrases more than others, and predicting this variation is a step towards explaining the underlying cognitive processes. Using data from the CRITT Translation Process Research Database, we evaluate the extent to which surprisal and attentional features derived from a Neural Machine Translation (NMT) model account for reading and production times of human translators. We find that surprisal and attention are complementary predictors of translation difficulty, and that surprisal derived from a NMT model is the single most successful predictor of production duration. Our analyses draw on data from hundreds of translators operating across 13 language pairs, and represent the most comprehensive investigation of human translation difficulty to date.
Authors: Zizhong Li, Haopeng Zhang, Jiawei Zhang
The proliferation of fake news has emerged as a critical issue in recent years, requiring significant efforts to detect it. However, the existing fake news detection datasets are sourced from human journalists, which are likely to have inherent bias limitations due to the highly subjective nature of this task. In this paper, we revisit the existing fake news dataset verified by human journalists with augmented fact-checking by large language models (ChatGPT), and we name the augmented fake news dataset ChatGPT-FC. We quantitatively analyze the distinctions and resemblances between human journalists and LLM in assessing news subject credibility, news creator credibility, time-sensitive, and political framing. Our findings highlight LLM's potential to serve as a preliminary screening method, offering a promising avenue to mitigate the inherent biases of human journalists and enhance fake news detection.
Authors: Weixi Song, Zuchao Li, Lefei Zhang, Hai Zhao, Bo Du
With the prevalence of pre-training-fine-tuning paradigm, how to efficiently adapt the pre-trained model to the downstream tasks has been an intriguing issue. Parameter-Efficient Fine-Tuning (PEFT) methods have been proposed for low-cost adaptation, including Adapters, Bia-only, and the recently widely used Low-Rank Adaptation. Although these methods have demonstrated their effectiveness to some extent and have been widely applied, the underlying principles are still unclear. In this paper, we reveal the transition of loss landscape in the downstream domain from random initialization to pre-trained initialization, that is, from low-amplitude oscillation to high-amplitude oscillation. The parameter gradients exhibit a property akin to sparsity, where a small fraction of components dominate the total gradient norm, for instance, 1% of the components account for 99% of the gradient. This property ensures that the pre-trained model can easily find a flat minimizer which guarantees the model's ability to generalize even with a low number of trainable parameters. Based on this, we propose a gradient-based sparse fine-tuning algorithm, named Sparse Increment Fine-Tuning (SIFT), and validate its effectiveness on a range of tasks including the GLUE Benchmark and Instruction-tuning. The code is accessible at https://github.com/song-wx/SIFT/.
Authors: Taehong Jang, Joonmo Ahn, Sojung Lucia Kim
In Korean ancient documents, there is no spacing or punctuation, and they are written in classical Chinese characters. This makes it challenging for modern individuals and translation models to accurately interpret and translate them. While China has models predicting punctuation and spacing, applying them directly to Korean texts is problematic due to data differences. Therefore, we developed the first models which predict punctuation and spacing for Korean historical texts and evaluated their performance. Our punctuation restoration model achieved an F1 score of 0.84, and Spacing model achieved a score of 0.96. It has the advantage of enabling inference on low-performance GPUs with less VRAM while maintaining quite high accuracy.
Authors: Ziqian Zeng, Yihuai Hong, Hongliang Dai, Huiping Zhuang, Cen Chen
Early Exiting is one of the most popular methods to achieve efficient inference. Current early exiting methods adopt the (weighted) sum of the cross entropy loss of all internal classifiers during training, imposing all these classifiers to predict all instances correctly. However, during inference, as long as one internal classifier predicts an instance correctly, it can accelerate without losing accuracy. Thus, there is a notable gap between training and inference. We propose ConsistentEE, an early exiting method that is consistent in training and inference. ConsistentEE formulates the early exiting process as a reinforcement learning problem. A policy network is added to decide whether an instance should exit or continue. The training objective of ConsistentEE only require each instance to be predicted correctly by one internal classifier. Additionally, we introduce the concept Memorize Layer to measure the hardness of an instance. We incorporate memorized layer into reward function design, which allows ``easy'' instances to focus more on acceleration while ``hard'' instances to focus more on accuracy. Experimental results show that our method outperforms other baselines on various natural language understanding and generation tasks.
Authors: Unggi Lee, Sungjun Yoon, Joon Seo Yun, Kyoungsoo Park, YoungHoon Jung, Damji Stratton, Hyeoncheol Kim
This paper presents novel techniques for enhancing the performance of knowledge tracing (KT) models by focusing on the crucial factor of question and concept difficulty level. Despite the acknowledged significance of difficulty, previous KT research has yet to exploit its potential for model optimization and has struggled to predict difficulty from unseen data. To address these problems, we propose a difficulty-centered contrastive learning method for KT models and a Large Language Model (LLM)-based framework for difficulty prediction. These innovative methods seek to improve the performance of KT models and provide accurate difficulty estimates for unseen data. Our ablation study demonstrates the efficacy of these techniques by demonstrating enhanced KT model performance. Nonetheless, the complex relationship between language and difficulty merits further investigation.
Authors: Nirmalya Thakur, Yuvraj Nihal Duggal, Zihui Liu
The recent outbreak of the MPox virus has resulted in a tremendous increase in the usage of Twitter. Prior works in this area of research have primarily focused on the sentiment analysis and content analysis of these Tweets, and the few works that have focused on topic modeling have multiple limitations. This paper aims to address this research gap and makes two scientific contributions to this field. First, it presents the results of performing Topic Modeling on 601,432 Tweets about the 2022 Mpox outbreak that were posted on Twitter between 7 May 2022 and 3 March 2023. The results indicate that the conversations on Twitter related to Mpox during this time range may be broadly categorized into four distinct themes - Views and Perspectives about Mpox, Updates on Cases and Investigations about Mpox, Mpox and the LGBTQIA+ Community, and Mpox and COVID-19. Second, the paper presents the findings from the analysis of these Tweets. The results show that the theme that was most popular on Twitter (in terms of the number of Tweets posted) during this time range was Views and Perspectives about Mpox. This was followed by the theme of Mpox and the LGBTQIA+ Community, which was followed by the themes of Mpox and COVID-19 and Updates on Cases and Investigations about Mpox, respectively. Finally, a comparison with related studies in this area of research is also presented to highlight the novelty and significance of this research work.
Authors: Chen Li
One of the key issues in Mandarin Chinese text-to-speech (TTS) systems is polyphone disambiguation when doing grapheme-to-phoneme (G2P) conversion. In this paper, we introduce a novel method to solve the problem as a generation task. Following the trending research of large language models (LLM) and prompt learning, the proposed method consists of three modules. Retrieval module incorporates external knowledge which is a multi-level semantic dictionary of Chinese polyphonic characters to format the sentence into a prompt. Generation module adopts the decoder-only Transformer architecture to induce the target text. Postprocess module corrects the generated text into a valid result if needed. Experimental results show that our method outperforms the existing methods on a public dataset called CPP. We also empirically study the impacts of different templates of the prompt, different sizes of training data, and whether to incorporate external knowledge.
Authors: Haowei Du, Quzhe Huang, Chen Li, Chen Zhang, Yang Li, Dongyan Zhao
Multi-hop Knowledge Base Question Answering(KBQA) aims to find the answer entity in a knowledge graph (KG), which requires multiple steps of reasoning. Existing retrieval-based approaches solve this task by concentrating on the specific relation at different hops and predicting the intermediate entity within the reasoning path. During the reasoning process of these methods, the representation of relations are fixed but the initial relation representation may not be optimal. We claim they fail to utilize information from head-tail entities and the semantic connection between relations to enhance the current relation representation, which undermines the ability to capture information of relations in KGs. To address this issue, we construct a \textbf{dual relation graph} where each node denotes a relation in the original KG (\textbf{primal entity graph}) and edges are constructed between relations sharing same head or tail entities. Then we iteratively do primal entity graph reasoning, dual relation graph information propagation, and interaction between these two graphs. In this way, the interaction between entity and relation is enhanced, and we derive better entity and relation representations. Experiments on two public datasets, WebQSP and CWQ, show that our approach achieves a significant performance gain over the prior state-of-the-art. Our code is available on \url{https://github.com/yanmenxue/RAH-KBQA}.
Authors: Haowei Du, Dingyu Zhang, Chen Li, Yang Li, Dongyan Zhao
Recent approaches in Incomplete Utterance Rewriting (IUR) fail to capture the source of important words, which is crucial to edit the incomplete utterance, and introduce words from irrelevant utterances. We propose a novel and effective multi-task information interaction framework including context selection, edit matrix construction, and relevance merging to capture the multi-granularity of semantic information. Benefiting from fetching the relevant utterance and figuring out the important words, our approach outperforms existing state-of-the-art models on two benchmark datasets Restoration-200K and CANAND in this field. Code will be provided on \url{https://github.com/yanmenxue/QR}.
Authors: Rui Liu, Yifan Hu, Yi Ren, Xiang Yin, Haizhou Li
Conversational Speech Synthesis (CSS) aims to accurately express an utterance with the appropriate prosody and emotional inflection within a conversational setting. While recognising the significance of CSS task, the prior studies have not thoroughly investigated the emotional expressiveness problems due to the scarcity of emotional conversational datasets and the difficulty of stateful emotion modeling. In this paper, we propose a novel emotional CSS model, termed ECSS, that includes two main components: 1) to enhance emotion understanding, we introduce a heterogeneous graph-based emotional context modeling mechanism, which takes the multi-source dialogue history as input to model the dialogue context and learn the emotion cues from the context; 2) to achieve emotion rendering, we employ a contrastive learning-based emotion renderer module to infer the accurate emotion style for the target utterance. To address the issue of data scarcity, we meticulously create emotional labels in terms of category and intensity, and annotate additional emotional information on the existing conversational dataset (DailyTalk). Both objective and subjective evaluations suggest that our model outperforms the baseline models in understanding and rendering emotions. These evaluations also underscore the importance of comprehensive emotional annotations. Code and audio samples can be found at: https://github.com/walker-hyf/ECSS.
Authors: Chen Gao, Xiaochong Lan, Nian Li, Yuan Yuan, Jingtao Ding, Zhilun Zhou, Fengli Xu, Yong Li
Agent-based modeling and simulation has evolved as a powerful tool for modeling complex systems, offering insights into emergent behaviors and interactions among diverse agents. Integrating large language models into agent-based modeling and simulation presents a promising avenue for enhancing simulation capabilities. This paper surveys the landscape of utilizing large language models in agent-based modeling and simulation, examining their challenges and promising future directions. In this survey, since this is an interdisciplinary field, we first introduce the background of agent-based modeling and simulation and large language model-empowered agents. We then discuss the motivation for applying large language models to agent-based simulation and systematically analyze the challenges in environment perception, human alignment, action generation, and evaluation. Most importantly, we provide a comprehensive overview of the recent works of large language model-empowered agent-based modeling and simulation in multiple scenarios, which can be divided into four domains: cyber, physical, social, and hybrid, covering simulation of both real-world and virtual environments. Finally, since this area is new and quickly evolving, we discuss the open problems and promising future directions.
Authors: Yongqi An, Xu Zhao, Tao Yu, Ming Tang, Jinqiao Wang
Network Pruning is a promising way to address the huge computing resource demands of the deployment and inference of Large Language Models (LLMs). Retraining-free is important for LLMs' pruning methods. However, almost all of the existing retraining-free pruning approaches for LLMs focus on unstructured pruning, which requires specific hardware support for acceleration. In this paper, we propose a novel retraining-free structured pruning framework for LLMs, named FLAP (FLuctuation-based Adaptive Structured Pruning). It is hardware-friendly by effectively reducing storage and enhancing inference speed. For effective structured pruning of LLMs, we highlight three critical elements that demand the utmost attention: formulating structured importance metrics, adaptively searching the global compressed model, and implementing compensation mechanisms to mitigate performance loss. First, FLAP determines whether the output feature map is easily recoverable when a column of weight is removed, based on the fluctuation pruning metric. Then it standardizes the importance scores to adaptively determine the global compressed model structure. At last, FLAP adds additional bias terms to recover the output feature maps using the baseline values. We thoroughly evaluate our approach on a variety of language benchmarks. Without any retraining, our method significantly outperforms the state-of-the-art methods, including LLM-Pruner and the extension of Wanda in structured pruning. The code is released at https://github.com/CASIA-IVA-Lab/FLAP.
Authors: Hongyin Zhu, Prayag Tiwari
Climate change presents significant challenges to the global community, and it is imperative to raise widespread awareness of the climate crisis and educate users about low-carbon living. Artificial intelligence, particularly large language models (LLMs), have emerged as powerful tools in mitigating the climate crisis, leveraging their extensive knowledge, broad user base, and natural language interaction capabilities. However, despite the growing body of research on climate change, there is a lack of comprehensive assessments of climate crisis knowledge within LLMs. This paper aims to resolve this gap by proposing an automatic evaluation framework. We employ a hybrid approach to data acquisition that combines data synthesis and manual collection to compile a diverse set of questions related to the climate crisis. These questions cover various aspects of climate change, including its causes, impacts, mitigation strategies, and adaptation measures. We then evaluate the model knowledge through prompt engineering based on the collected questions and generated answers. We propose a set of comprehensive metrics to evaluate the climate crisis knowledge, incorporating indicators from 10 different perspectives. Experimental results show that our method is effective in evaluating the knowledge of LLMs regarding the climate crisis. We evaluate several state-of-the-art LLMs and find that their knowledge falls short in terms of timeliness.
Authors: Zhuowei Zhang, Mengting Hu, Yinhao Bai, Zhen Zhang
Mind-map generation aims to process a document into a hierarchical structure to show its central idea and branches. Such a manner is more conducive to understanding the logic and semantics of the document than plain text. Recently, a state-of-the-art method encodes the sentences of a document sequentially and converts them to a relation graph via sequence-to-graph. Though this method is efficient to generate mind-maps in parallel, its mechanism focuses more on sequential features while hardly capturing structural information. Moreover, it's difficult to model long-range semantic relations. In this work, we propose a coreference-guided mind-map generation network (CMGN) to incorporate external structure knowledge. Specifically, we construct a coreference graph based on the coreference semantic relationship to introduce the graph structure information. Then we employ a coreference graph encoder to mine the potential governing relations between sentences. In order to exclude noise and better utilize the information of the coreference graph, we adopt a graph enhancement module in a contrastive learning manner. Experimental results demonstrate that our model outperforms all the existing methods. The case study further proves that our model can more accurately and concisely reveal the structure and semantics of a document. Code and data are available at https://github.com/Cyno2232/CMGN.
Authors: Md. Rafiul Biswas, Ashhadul Islam, Zubair Shah, Wajdi Zaghouani, Samir Brahim Belhaouari
The advanced large language model (LLM) ChatGPT has shown its potential in different domains and remains unbeaten due to its characteristics compared to other LLMs. This study aims to evaluate the potential of using a fine-tuned ChatGPT model as a personal medical assistant in the Arabic language. To do so, this study uses publicly available online questions and answering datasets in Arabic language. There are almost 430K questions and answers for 20 disease-specific categories. GPT-3.5-turbo model was fine-tuned with a portion of this dataset. The performance of this fine-tuned model was evaluated through automated and human evaluation. The automated evaluations include perplexity, coherence, similarity, and token count. Native Arabic speakers with medical knowledge evaluated the generated text by calculating relevance, accuracy, precision, logic, and originality. The overall result shows that ChatGPT has a bright future in medical assistance.
Authors: Top Piriyakulkij, Volodymyr Kuleshov, Kevin Ellis
Actively inferring user preferences, for example by asking good questions, is important for any human-facing decision-making system. Active inference allows such systems to adapt and personalize themselves to nuanced individual preferences. To enable this ability for instruction-tuned large language models (LLMs), one may prompt them to ask users questions to infer their preferences, transforming the language models into more robust, interactive systems. However, out of the box, these models are not efficient at extracting preferences: the questions they generate are not informative, requiring a high number of user interactions and impeding the usability of the downstream system. In this work, we introduce an inference-time algorithm that helps LLMs quickly infer preferences by using more informative questions. Our algorithm uses a probabilistic model whose conditional distributions are defined by prompting an LLM, and returns questions that optimize expected entropy and expected model change. Results in a simplified interactive web shopping setting with real product items show that an LLM equipped with our entropy reduction algorithm outperforms baselines with the same underlying LLM on task performance while using fewer user interactions.
Authors: DaLuo, Yanglei Gan, Rui Hou, Run Lin, Qiao Liu, Yuxiang Cai, Wannian Gao
Few-shot Relation Extraction (FSRE) aims to extract relational facts from a sparse set of labeled corpora. Recent studies have shown promising results in FSRE by employing Pre-trained Language Models (PLMs) within the framework of supervised contrastive learning, which considers both instances and label facts. However, how to effectively harness massive instance-label pairs to encompass the learned representation with semantic richness in this learning paradigm is not fully explored. To address this gap, we introduce a novel synergistic anchored contrastive pre-training framework. This framework is motivated by the insight that the diverse viewpoints conveyed through instance-label pairs capture incomplete yet complementary intrinsic textual semantics. Specifically, our framework involves a symmetrical contrastive objective that encompasses both sentence-anchored and label-anchored contrastive losses. By combining these two losses, the model establishes a robust and uniform representation space. This space effectively captures the reciprocal alignment of feature distributions among instances and relational facts, simultaneously enhancing the maximization of mutual information across diverse perspectives within the same relation. Experimental results demonstrate that our framework achieves significant performance enhancements compared to baseline models in downstream FSRE tasks. Furthermore, our approach exhibits superior adaptability to handle the challenges of domain shift and zero-shot relation extraction. Our code is available online at https://github.com/AONE-NLP/FSRE-SaCon.
Authors: Sichao Xiong, Yigit Ihlamur
This research introduces an innovative evaluation method for the "founder-idea" fit in early-stage startups, utilizing advanced large language model techniques to assess founders' profiles against their startup ideas to enhance decision-making. Embeddings, self-play, tree-of-thought, and critique-based refinement techniques show early promising results that each idea's success patterns are unique and they should be evaluated based on the context of the founder's background.
Authors: Liu Bin, Yin Guosheng
In text generation, a large language model (LM) makes a choice of each new word based only on the former selection of its context using the softmax function. Nevertheless, the link statistics information of concurrent words based on a scene-specific corpus is valuable in choosing the next word, which can help to ensure the topic of the generated text to be aligned with the current task. To fully explore the co-occurrence information,we propose a graphmax function for task-specific text generation. Using the graph-based regularization, graphmax enables the final word choice to be determined by both the global knowledge from the LM and the local knowledge from the scene-specific corpus. The traditional softmax function is regularized with a graph total variation (GTV) term, which incorporates the local knowledge into the LM and encourages the model to consider the statistical relationships between words in a scene-specific corpus. The proposed graphmax is versatile and can be readily plugged into any large pre-trained LM for text generation and machine translation. Through extensive experiments, we demonstrate that the new GTV-based regularization can improve performances in various natural language processing tasks in comparison with existing methods. Moreover, through human experiments, we observe that participants can easily distinguish the text generated by graphmax or softmax.
Authors: Hanqi Yan, Lin Gui, Gabriele Pergola, Yulan He
The Emotion Cause Extraction (ECE)} task aims to identify clauses which contain emotion-evoking information for a particular emotion expressed in text. We observe that a widely-used ECE dataset exhibits a bias that the majority of annotated cause clauses are either directly before their associated emotion clauses or are the emotion clauses themselves. Existing models for ECE tend to explore such relative position information and suffer from the dataset bias. To investigate the degree of reliance of existing ECE models on clause relative positions, we propose a novel strategy to generate adversarial examples in which the relative position information is no longer the indicative feature of cause clauses. We test the performance of existing models on such adversarial examples and observe a significant performance drop. To address the dataset bias, we propose a novel graph-based method to explicitly model the emotion triggering paths by leveraging the commonsense knowledge to enhance the semantic dependencies between a candidate clause and an emotion clause. Experimental results show that our proposed approach performs on par with the existing state-of-the-art methods on the original ECE dataset, and is more robust against adversarial attacks compared to existing models.
Authors: Kevin Denamganaï, Sondess Missaoui, James Alfred Walker
Human beings use compositionality to generalise from past experiences to novel experiences. We assume a separation of our experiences into fundamental atomic components that can be recombined in novel ways to support our ability to engage with novel experiences. We frame this as the ability to learn to generalise compositionally, and we will refer to behaviours making use of this ability as compositional learning behaviours (CLBs). A central problem to learning CLBs is the resolution of a binding problem (BP). While it is another feat of intelligence that human beings perform with ease, it is not the case for state-of-the-art artificial agents. Thus, in order to build artificial agents able to collaborate with human beings, we propose to develop a novel benchmark to investigate agents' abilities to exhibit CLBs by solving a domain-agnostic version of the BP. We take inspiration from the language emergence and grounding framework of referential games and propose a meta-learning extension of referential games, entitled Meta-Referential Games, and use this framework to build our benchmark, the Symbolic Behaviour Benchmark (S2B). We provide baseline results and error analysis showing that our benchmark is a compelling challenge that we hope will spur the research community towards developing more capable artificial agents.
Authors: Hanqi Yan, Lin Gui, Wenjie Li, Yulan He
Token uniformity is commonly observed in transformer-based models, in which different tokens share a large proportion of similar information after going through stacked multiple self-attention layers in a transformer. In this paper, we propose to use the distribution of singular values of outputs of each transformer layer to characterise the phenomenon of token uniformity and empirically illustrate that a less skewed singular value distribution can alleviate the `token uniformity' problem. Base on our observations, we define several desirable properties of singular value distributions and propose a novel transformation function for updating the singular values. We show that apart from alleviating token uniformity, the transformation function should preserve the local neighbourhood structure in the original embedding space. Our proposed singular value transformation function is applied to a range of transformer-based language models such as BERT, ALBERT, RoBERTa and DistilBERT, and improved performance is observed in semantic textual similarity evaluation and a range of GLUE tasks. Our source code is available at https://github.com/hanqi-qi/tokenUni.git.
Authors: Jing Huang, Zhengxuan Wu, Kyle Mahowald, Christopher Potts
Language tasks involving character-level manipulations (e.g., spelling corrections, arithmetic operations, word games) are challenging for models operating on subword units. To address this, we develop a causal intervention framework to learn robust and interpretable character representations inside subword-based language models. Our method treats each character as a typed variable in a causal model and learns such causal structures by adapting the interchange intervention training method of Geiger et al. (2021). We additionally introduce a suite of character-level tasks that systematically vary in their dependence on meaning and sequence-level context. While character-level models still perform best on purely form-based tasks like string reversal, our method outperforms character-level models on more complex tasks that blend form, meaning, and context, such as spelling correction in context and word search games. Compared with standard subword-based models, our approach also significantly improves robustness on unseen token sequences and leads to human-interpretable internal representations of characters.
Authors: Wenting Li, Jiahong Xue, Xi Zhang, Huacan Chen, Zeyu Chen, Yuanzhe Cai
Word embedding has become ubiquitous and is widely used in various text mining and natural language processing (NLP) tasks, such as information retrieval, semantic analysis, and machine translation, among many others. Unfortunately, it is prohibitively expensive to train the word embedding in a relatively large corpus. We propose a graph-based word embedding algorithm, called Word-Graph2vec, which converts the large corpus into a word co-occurrence graph, then takes the word sequence samples from this graph by randomly traveling and trains the word embedding on this sampling corpus in the end. We posit that because of the stable vocabulary, relative idioms, and fixed expressions in English, the size and density of the word co-occurrence graph change slightly with the increase in the training corpus. So that Word-Graph2vec has stable runtime on the large scale data set, and its performance advantage becomes more and more obvious with the growth of the training corpus. Extensive experiments conducted on real-world datasets show that the proposed algorithm outperforms traditional Skip-Gram by four-five times in terms of efficiency, while the error generated by the random walk sampling is small.
Authors: OpenAI: Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ahmad, Ilge Akkaya, Florencia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, Shyamal Anadkat, Red Avila, Igor Babuschkin, Suchir Balaji, Valerie Balcom, Paul Baltescu, Haiming Bao, Mo Bavarian, Jeff Belgum, Irwan Bello, Jake Berdine, Gabriel Bernadett-Shapiro, Christopher Berner, Lenny Bogdonoff, Oleg Boiko, Madelaine Boyd, Anna-Luisa Brakman, Greg Brockman, Tim Brooks, Miles Brundage, Kevin Button, Trevor Cai, Rosie Campbell, Andrew Cann, Brittany Carey, Chelsea Carlson, Rory Carmichael, Brooke Chan, Che Chang, Fotis Chantzis, Derek Chen, Sully Chen, Ruby Chen, Jason Chen, Mark Chen, Ben Chess, Chester Cho, Casey Chu, Hyung Won Chung, Dave Cummings, Jeremiah Currier, Yunxing Dai, Cory Decareaux, Thomas Degry, Noah Deutsch, et al. (226 additional authors not shown)
We report the development of GPT-4, a large-scale, multimodal model which can accept image and text inputs and produce text outputs. While less capable than humans in many real-world scenarios, GPT-4 exhibits human-level performance on various professional and academic benchmarks, including passing a simulated bar exam with a score around the top 10% of test takers. GPT-4 is a Transformer-based model pre-trained to predict the next token in a document. The post-training alignment process results in improved performance on measures of factuality and adherence to desired behavior. A core component of this project was developing infrastructure and optimization methods that behave predictably across a wide range of scales. This allowed us to accurately predict some aspects of GPT-4's performance based on models trained with no more than 1/1,000th the compute of GPT-4.
Authors: Zhengqing Yuan, Xinyi Wang, Kun Wang, Lichao Sun, Yanfang Ye
In recent years, advancements in large language models have been remarkable, with models such as ChatGPT demonstrating exceptional proficiency in diverse linguistic tasks. The pre-training of large models with billions of parameters, poses a formidable challenge, primarily due to the scarcity of datasets of a commensurate scale for effective training. Nevertheless, innovative strategies have emerged, including methods to fine-tune these pre-trained models using fewer parameters set, as evidenced by models like MiniGPT-4 and LLaVA. Despite their potential in various domains, these models remain limited in their understanding of artistic imagery. They have yet to fully grasp the intricate nuances of art images or to provide an objective articulation of the emotions they evoke, in a manner akin to human perception. This work introduces ArtGPT-4, a pioneering large vision-language model tailored to address the deficiencies of contemporary models in artistic comprehension. ArtGPT-4 underwent training on image-text pairs utilizing a Tesla A100 device in a mere 2 hours, with a dataset comprising approximately 0.52M entries. Impressively, the model can render images with an artistic-understanding and convey the emotions they inspire, mirroring human interpretation. Additionally, this work presents a unique dataset designed to evaluate the efficacy of vision-language models. In subsequent evaluations, ArtGPT-4 not only achieved state-of-the-art performance on the ArtEmis and ArtEmis-v2.0 datasets but also exceeded the established benchmarks introduced in This study, lagging behind professional artists' descriptions by a negligible 0.15 points on a 6-point scale. The code and the pre-trained model are accessible in https://huggingface.co/Tyrannosaurus/ArtGPT-4.
Authors: Lean Wang, Lei Li, Damai Dai, Deli Chen, Hao Zhou, Fandong Meng, Jie Zhou, Xu Sun
In-context learning (ICL) emerges as a promising capability of large language models (LLMs) by providing them with demonstration examples to perform diverse tasks. However, the underlying mechanism of how LLMs learn from the provided context remains under-explored. In this paper, we investigate the working mechanism of ICL through an information flow lens. Our findings reveal that label words in the demonstration examples function as anchors: (1) semantic information aggregates into label word representations during the shallow computation layers' processing; (2) the consolidated information in label words serves as a reference for LLMs' final predictions. Based on these insights, we introduce an anchor re-weighting method to improve ICL performance, a demonstration compression technique to expedite inference, and an analysis framework for diagnosing ICL errors in GPT2-XL. The promising applications of our findings again validate the uncovered ICL working mechanism and pave the way for future studies.
Authors: Wang Zhu, Jesse Thomason, Robin Jia
We train a language model (LM) to robustly answer multistep questions by generating and answering sub-questions. We propose Chain-of-Questions, a framework that trains a model to generate sub-questions and sub-answers one at a time by leveraging human annotated question decomposition meaning representation (QDMR). The key technical challenge is that QDMR only contains sub-questions but not answers to those sub-questions, so we treat sub-answers as latent variables and optimize them using a novel dynamic mixture of Hard-EM and MAPO. Chain-of-Questions greatly outperforms strong neuro-symbolic methods by 9.0 F1 on DROP contrast set, and outperforms GPT-3.5 by 24.3 F1 on HOTPOTQA adversarial set, thus demonstrating the effectiveness and robustness of our framework.
Authors: Zhiyuan Hu, Chumin Liu, Yue Feng, Anh Tuan Luu, Bryan Hooi
Controllable text generation is a challenging and meaningful field in natural language generation (NLG). Especially, poetry generation is a typical one with well-defined and strict conditions for text generation which is an ideal playground for the assessment of current methodologies. While prior works succeeded in controlling either semantic or metrical aspects of poetry generation, simultaneously addressing both remains a challenge. In this paper, we pioneer the use of the Diffusion model for generating sonnets and Chinese SongCi poetry to tackle such challenges. In terms of semantics, our PoetryDiffusion model, built upon the Diffusion model, generates entire sentences or poetry by comprehensively considering the entirety of sentence information. This approach enhances semantic expression, distinguishing it from autoregressive and large language models (LLMs). For metrical control, the separation feature of diffusion generation and its constraint control module enable us to flexibly incorporate a novel metrical controller to manipulate and evaluate metrics (format and rhythm). The denoising process in PoetryDiffusion allows for gradual enhancement of semantics and flexible integration of the metrical controller which can calculate and impose penalties on states that stray significantly from the target control distribution. Experimental results on two datasets demonstrate that our model outperforms existing models in automatic evaluation of semantic, metrical, and overall performance as well as human evaluation.
Authors: Chen Qian, Xin Cong, Wei Liu, Cheng Yang, Weize Chen, Yusheng Su, Yufan Dang, Jiahao Li, Juyuan Xu, Dahai Li, Zhiyuan Liu, Maosong Sun
Software engineering is a domain characterized by intricate decision-making processes, often relying on nuanced intuition and consultation. Recent advancements in deep learning have started to revolutionize software engineering practices through elaborate designs implemented at various stages of software development. In this paper, we present an innovative paradigm that leverages large language models (LLMs) throughout the entire software development process, streamlining and unifying key processes through natural language communication, thereby eliminating the need for specialized models at each phase. At the core of this paradigm lies ChatDev, a virtual chat-powered software development company that mirrors the established waterfall model, meticulously dividing the development process into four distinct chronological stages: designing, coding, testing, and documenting. Each stage engages a team of "software agents", such as programmers, code reviewers, and test engineers, fostering collaborative dialogue and facilitating a seamless workflow. The chat chain acts as a facilitator, breaking down each stage into atomic subtasks. This enables dual roles, allowing for proposing and validating solutions through context-aware communication, leading to efficient resolution of specific subtasks. The instrumental analysis of ChatDev highlights its remarkable efficacy in software generation, enabling the completion of the entire software development process in under seven minutes at a cost of less than one dollar. It not only identifies and alleviates potential vulnerabilities but also rectifies potential hallucinations while maintaining commendable efficiency and cost-effectiveness. The potential of ChatDev unveils fresh possibilities for integrating LLMs into the realm of software development. Our code is available at https://github.com/OpenBMB/ChatDev.
Authors: Zhen Qin, Yiran Zhong, Hui Deng
Length extrapolation has attracted considerable attention recently since it allows transformers to be tested on longer sequences than those used in training. Previous research has shown that this property can be attained by using carefully designed Relative Positional Encodings (RPEs). While these methods perform well on a variety of corpora, the conditions for length extrapolation have yet to be investigated. This paper attempts to determine what types of RPEs allow for length extrapolation through a thorough mathematical and empirical analysis. We discover that a transformer is certain to possess this property as long as the series that corresponds to the RPE's exponential converges. Two practices are derived from the conditions and examined in language modeling tasks on a variety of corpora. As a bonus from the conditions, we derive a new Theoretical Receptive Field (TRF) to measure the receptive field of RPEs without taking any training steps. Extensive experiments are conducted on the Wikitext-103, Books, Github, and WikiBook datasets to demonstrate the viability of our discovered conditions. We also compare TRF to Empirical Receptive Field (ERF) across different models, showing consistently matched trends on the aforementioned datasets. The code is available at https://github.com/OpenNLPLab/Rpe.
Authors: Bin Wang, Zhengyuan Liu, Xin Huang, Fangkai Jiao, Yang Ding, Ai Ti Aw, Nancy F. Chen
We present SeaEval, a benchmark for multilingual foundation models. In addition to characterizing how these models understand and reason with natural language, we also investigate how well they comprehend cultural practices, nuances, and values. Alongside standard accuracy metrics, we investigate the brittleness of foundation models in the dimensions of semantics and multilinguality. Our analyses span both open-sourced and closed models, leading to empirical results across classic NLP tasks, reasoning, and cultural comprehension. Key findings indicate (1) Most models exhibit varied behavior when given paraphrased instructions. (2) Many models still suffer from exposure bias (e.g., positional bias, majority label bias). (3) For questions rooted in factual, scientific, and commonsense knowledge, consistent responses are expected across multilingual queries that are semantically equivalent. Yet, most models surprisingly demonstrate inconsistent performance on these queries. (4) Multilingually-trained models have not attained "balanced multilingual" capabilities. Our endeavors underscore the need for more generalizable semantic representations and enhanced multilingual contextualization. SeaEval can serve as a launchpad for more thorough investigations and evaluations for multilingual and multicultural scenarios.
Authors: Mingxin Li, Richong Zhang, Zhijie Nie, Yongyi Mao
Sentence Representation Learning (SRL) is a fundamental task in Natural Language Processing (NLP), with the Contrastive Learning of Sentence Embeddings (CSE) being the mainstream technique due to its superior performance. An intriguing phenomenon in CSE is the significant performance gap between supervised and unsupervised methods, with their only difference lying in the training data. Previous works attribute this performance gap to differences in two representation properties (alignment and uniformity). However, since alignment and uniformity only measure the results, they fail to answer "What aspects of the training data contribute to the performance gap?" and "How can the performance gap be narrowed?", In this paper, we conduct empirical experiments to answer these "What" and "How" questions. We first answer the "What" question by thoroughly comparing the behavior of supervised and unsupervised CSE during their respective training processes. From the comparison, we identify the similarity pattern as a key factor to the performance gap, and introduce a metric, called Relative Fitting Difficulty (RFD), to measure the complexity of the similarity pattern. Then, based on the insights gained from the "What" question, we tackle the "How" question by increasing the pattern complexity of the training data. We achieve this by leveraging the In-Context Learning (ICL) capability of the Large Language Model (LLM) to generate data that simulates complex patterns. By utilizing the hierarchical patterns in the LLM-generated data, we effectively narrow the gap between supervised and unsupervised CSE. We release our codes and appendix at https://github.com/BDBC-KG-NLP/NGCSE.
Authors: Fernanda De La Torre, Cathy Mengying Fang, Han Huang, Andrzej Banburski-Fahey, Judith Amores Fernandez, Jaron Lanier
We present Large Language Model for Mixed Reality (LLMR), a framework for the real-time creation and modification of interactive Mixed Reality experiences using LLMs. LLMR leverages novel strategies to tackle difficult cases where ideal training data is scarce, or where the design goal requires the synthesis of internal dynamics, intuitive analysis, or advanced interactivity. Our framework relies on text interaction and the Unity game engine. By incorporating techniques for scene understanding, task planning, self-debugging, and memory management, LLMR outperforms the standard GPT-4 by 4x in average error rate. We demonstrate LLMR's cross-platform interoperability with several example worlds, and evaluate it on a variety of creation and modification tasks to show that it can produce and edit diverse objects, tools, and scenes. Finally, we conducted a usability study (N=11) with a diverse set that revealed participants had positive experiences with the system and would use it again.
Authors: Huihui Xu, Kevin Ashley
Traditional evaluation metrics like ROUGE compare lexical overlap between the reference and generated summaries without taking argumentative structure into account, which is important for legal summaries. In this paper, we propose a novel legal summarization evaluation framework that utilizes GPT-4 to generate a set of question-answer pairs that cover main points and information in the reference summary. GPT-4 is then used to generate answers based on the generated summary for the questions from the reference summary. Finally, GPT-4 grades the answers from the reference summary and the generated summary. We examined the correlation between GPT-4 grading with human grading. The results suggest that this question-answering approach with GPT-4 can be a useful tool for gauging the quality of the summary.
Authors: Shen Zheng, Yuyu Zhang, Yijie Zhu, Chenguang Xi, Pengyang Gao, Xun Zhou, Kevin Chen-Chuan Chang
With the rapid advancement of large language models (LLMs), there is a pressing need for a comprehensive evaluation suite to assess their capabilities and limitations. Existing LLM leaderboards often reference scores reported in other papers without consistent settings and prompts, which may inadvertently encourage cherry-picking favored settings and prompts for better results. In this work, we introduce GPT-Fathom, an open-source and reproducible LLM evaluation suite built on top of OpenAI Evals. We systematically evaluate 10+ leading LLMs as well as OpenAI's legacy models on 20+ curated benchmarks across 7 capability categories, all under aligned settings. Our retrospective study on OpenAI's earlier models offers valuable insights into the evolutionary path from GPT-3 to GPT-4. Currently, the community is eager to know how GPT-3 progressively improves to GPT-4, including technical details like whether adding code data improves LLM's reasoning capability, which aspects of LLM capability can be improved by SFT and RLHF, how much is the alignment tax, etc. Our analysis sheds light on many of these questions, aiming to improve the transparency of advanced LLMs.
Authors: Anej Svete, Ryan Cotterell
Studying language models (LMs) in terms of well-understood formalisms allows us to precisely characterize their abilities and limitations. Previous work has investigated the representational capacity of recurrent neural network (RNN) LMs in terms of their capacity to recognize unweighted formal languages. However, LMs do not describe unweighted formal languages -- rather, they define \emph{probability distributions} over strings. In this work, we study what classes of such probability distributions RNN LMs can represent, which allows us to make more direct statements about their capabilities. We show that simple RNNs are equivalent to a subclass of probabilistic finite-state automata, and can thus model a strict subset of probability distributions expressible by finite-state models. Furthermore, we study the space complexity of representing finite-state LMs with RNNs. We show that, to represent an arbitrary deterministic finite-state LM with $N$ states over an alphabet $\alphabet$, an RNN requires $\Omega\left(N |\Sigma|\right)$ neurons. These results present a first step towards characterizing the classes of distributions RNN LMs can represent and thus help us understand their capabilities and limitations.
Authors: Jiwan Chung, Youngjae Yu
Multimodal language generation, which leverages the synergy of language and vision, is a rapidly expanding field. However, existing vision-language models face challenges in tasks that require complex linguistic understanding. To address this issue, we introduce Visual-Language models as Importance Sampling weights (VLIS), a novel framework that combines the visual conditioning capability of vision-language models with the language understanding of unimodal text-only language models without further training. It extracts pointwise mutual information of each image and text from a visual-language model and uses the value as an importance sampling weight to adjust the token likelihood from a text-only model. VLIS improves vision-language models on diverse tasks, including commonsense understanding (WHOOPS, OK-VQA, and ScienceQA) and complex text generation (Concadia, Image Paragraph Captioning, and ROCStories). Our results suggest that VLIS represents a promising new direction for multimodal language generation.
Authors: Jiabin Tang, Yuhao Yang, Wei Wei, Lei Shi, Lixin Su, Suqi Cheng, Dawei Yin, Chao Huang
Graph Neural Networks (GNNs) have advanced graph structure understanding via recursive information exchange and aggregation among graph nodes. To improve model robustness, self-supervised learning (SSL) has emerged as a promising approach for data augmentation. However, existing methods for generating pre-trained graph embeddings often rely on fine-tuning with specific downstream task labels, which limits their usability in scenarios where labeled data is scarce or unavailable. To address this, our research focuses on advancing the generalization capabilities of graph models in challenging zero-shot learning scenarios. Inspired by the success of large language models (LLMs), we aim to develop a graph-oriented LLM that can achieve high generalization across diverse downstream datasets and tasks, even without any information available from the downstream graph data. In this work, we present the GraphGPT framework that aligns LLMs with graph structural knowledge with a graph instruction tuning paradigm. Our framework incorporates a text-graph grounding component to establish a connection between textual information and graph structures. Additionally, we propose a dual-stage instruction tuning paradigm, accompanied by a lightweight graph-text alignment projector. This paradigm explores self-supervised graph structural signals and task-specific graph instructions, to guide LLMs in understanding complex graph structures and improving their adaptability across different downstream tasks. Our framework is evaluated on supervised and zero-shot graph learning tasks, demonstrating superior generalization and outperforming state-of-the-art baselines.
Authors: Houwen Peng, Kan Wu, Yixuan Wei, Guoshuai Zhao, Yuxiang Yang, Ze Liu, Yifan Xiong, Ziyue Yang, Bolin Ni, Jingcheng Hu, Ruihang Li, Miaosen Zhang, Chen Li, Jia Ning, Ruizhe Wang, Zheng Zhang, Shuguang Liu, Joe Chau, Han Hu, Peng Cheng
In this paper, we explore FP8 low-bit data formats for efficient training of large language models (LLMs). Our key insight is that most variables, such as gradients and optimizer states, in LLM training can employ low-precision data formats without compromising model accuracy and requiring no changes to hyper-parameters. Specifically, we propose a new FP8 automatic mixed-precision framework for training LLMs. This framework offers three levels of FP8 utilization to streamline mixed-precision and distributed parallel training for LLMs. It gradually incorporates 8-bit gradients, optimizer states, and distributed learning in an incremental manner. Experiment results show that, during the training of GPT-175B model on H100 GPU platform, our FP8 mixed-precision training framework not only achieved a remarkable 39% reduction in real memory usage but also ran 75% faster than the widely adopted BF16 framework (i.e., Megatron-LM), surpassing the speed of Nvidia Transformer Engine by 37%. This largely reduces the training costs for large foundation models. Furthermore, our FP8 mixed-precision training methodology is generic. It can be seamlessly applied to other tasks such as LLM instruction tuning and reinforcement learning with human feedback, offering savings in fine-tuning expenses. Our FP8 low-precision training framework is open-sourced at {https://github.com/Azure/MS-AMP}{aka.ms/MS.AMP}.
Authors: Yann Hicke, Anmol Agarwal, Qianou Ma, Paul Denny
Responding to the thousands of student questions on online QA platforms each semester has a considerable human cost, particularly in computing courses with rapidly growing enrollments. To address the challenges of scalable and intelligent question-answering (QA), we introduce an innovative solution that leverages open-source Large Language Models (LLMs) from the LLaMA-2 family to ensure data privacy. Our approach combines augmentation techniques such as retrieval augmented generation (RAG), supervised fine-tuning (SFT), and learning from human preferences data using Direct Preference Optimization (DPO). Through extensive experimentation on a Piazza dataset from an introductory CS course, comprising 10,000 QA pairs and 1,500 pairs of preference data, we demonstrate a significant 30% improvement in the quality of answers, with RAG being a particularly impactful addition. Our contributions include the development of a novel architecture for educational QA, extensive evaluations of LLM performance utilizing both human assessments and LLM-based metrics, and insights into the challenges and future directions of educational data processing. This work paves the way for the development of AI-TA, an intelligent QA assistant customizable for courses with an online QA platform
Authors: Jiachen Zhao
Recently, large language models (LLMs) have made remarkable progress in natural language processing. The most representative ability of LLMs is in-context learning (ICL), which enables LLMs to learn patterns from in-context exemplars without training. The performance of ICL greatly depends on the exemplars used. However, how to choose exemplars remains unclear due to the lack of understanding of how in-context learning works. In this paper, we present a novel perspective on ICL by conceptualizing it as contextual retrieval from a model of associative memory. We establish a theoretical framework of ICL based on Hopfield Networks. Based on our framework, we look into how in-context exemplars influence the performance of ICL and propose more efficient active exemplar selection. Our study sheds new light on the mechanism of ICL by connecting it to memory retrieval, with potential implications for advancing the understanding of LLMs.
Authors: Shezheng Song, Xiaopeng Li, Shasha Li, Shan Zhao, Jie Yu, Jun Ma, Xiaoguang Mao, Weimin Zhang
This review paper explores Multimodal Large Language Models (MLLMs), which integrate Large Language Models (LLMs) like GPT-4 to handle multimodal data such as text and vision. MLLMs demonstrate capabilities like generating image narratives and answering image-based questions, bridging the gap towards real-world human-computer interactions and hinting at a potential pathway to artificial general intelligence. However, MLLMs still face challenges in processing the semantic gap in multimodality, which may lead to erroneous generation, posing potential risks to society. Choosing the appropriate modality alignment method is crucial, as improper methods might require more parameters with limited performance improvement. This paper aims to explore modality alignment methods for LLMs and their existing capabilities. Implementing modality alignment allows LLMs to address environmental issues and enhance accessibility. The study surveys existing modal alignment methods in MLLMs into four groups: (1) Multimodal Converters that change data into something LLMs can understand; (2) Multimodal Perceivers to improve how LLMs perceive different types of data; (3) Tools Assistance for changing data into one common format, usually text; and (4) Data-Driven methods that teach LLMs to understand specific types of data in a dataset. This field is still in a phase of exploration and experimentation, and we will organize and update various existing research methods for multimodal information alignment.
Authors: Ling Luo, Jinzhong Ning, Yingwen Zhao, Zhijun Wang, Zeyuan Ding, Peng Chen, Weiru Fu, Qinyu Han, Guangtao Xu, Yunzhi Qiu, Dinghao Pan, Jiru Li, Hao Li, Wenduo Feng, Senbo Tu, Yuqi Liu, Zhihao Yang, Jian Wang, Yuanyuan Sun, Hongfei Lin
Objective: Most existing fine-tuned biomedical large language models (LLMs) focus on enhancing performance in monolingual biomedical question answering and conversation tasks. To investigate the effectiveness of the fine-tuned LLMs on diverse biomedical NLP tasks in different languages, We present Taiyi, a bilingual fine-tuned LLM for diverse biomedical tasks. Materials and Methods: We first curated a comprehensive collection of 140 existing biomedical text mining datasets (102 English and 38 Chinese datasets) across over 10 task types. Subsequently, a two-stage strategy is proposed for supervised fine-tuning to optimize the model performance across varied tasks. Results: Experimental results on 13 test sets covering named entity recognition, relation extraction, text classification, question answering tasks demonstrate that Taiyi achieves superior performance compared to general LLMs. The case study involving additional biomedical NLP tasks further shows Taiyi's considerable potential for bilingual biomedical multi-tasking. Conclusion: Leveraging rich high-quality biomedical corpora and developing effective fine-tuning strategies can significantly improve the performance of LLMs within the biomedical domain. Taiyi shows the bilingual multi-tasking capability through supervised fine-tuning. However, those tasks such as information extraction that are not generation tasks in nature remain challenging for LLM-based generative approaches, and they still underperform the conventional discriminative approaches of smaller language models.
Authors: Will LeVine, Ben Pikus, Tony Chen, Sean Hendryx
Foundation models, specifically Large Language Models (LLM's), have lately gained wide-spread attention and adoption. Reinforcement Learning with Human Feedback (RLHF) involves training a reward model to capture desired behaviors, which is then used to align LLM's. These reward models are additionally used at inference-time to estimate LLM responses' adherence to those desired behaviors. However, there is little work measuring how robust these reward models are to distribution shifts. In this work, we evaluate how reward model performance - measured via accuracy and calibration (i.e. alignment between accuracy and confidence) - is affected by distribution shift. We show novel calibration patterns and accuracy drops due to OOD prompts and responses, and that the reward model is more sensitive to shifts in responses than prompts. Additionally, we adapt an OOD detection technique commonly used in classification to the reward model setting to detect these distribution shifts in prompts and responses.
Authors: Xiang Yue, Yuansheng Ni, Kai Zhang, Tianyu Zheng, Ruoqi Liu, Ge Zhang, Samuel Stevens, Dongfu Jiang, Weiming Ren, Yuxuan Sun, Cong Wei, Botao Yu, Ruibin Yuan, Renliang Sun, Ming Yin, Boyuan Zheng, Zhenzhu Yang, Yibo Liu, Wenhao Huang, Huan Sun, Yu Su, Wenhu Chen
We introduce MMMU: a new benchmark designed to evaluate multimodal models on massive multi-discipline tasks demanding college-level subject knowledge and deliberate reasoning. MMMU includes 11.5K meticulously collected multimodal questions from college exams, quizzes, and textbooks, covering six core disciplines: Art & Design, Business, Science, Health & Medicine, Humanities & Social Science, and Tech & Engineering. These questions span 30 subjects and 183 subfields, comprising 30 highly heterogeneous image types, such as charts, diagrams, maps, tables, music sheets, and chemical structures. Unlike existing benchmarks, MMMU focuses on advanced perception and reasoning with domain-specific knowledge, challenging models to perform tasks akin to those faced by experts. The evaluation of 14 open-source LMMs as well as the proprietary GPT-4V(ision) and Gemini highlights the substantial challenges posed by MMMU. Even the advanced GPT-4V and Gemini Ultra only achieve accuracies of 56% and 59% respectively, indicating significant room for improvement. We believe MMMU will stimulate the community to build next-generation multimodal foundation models towards expert artificial general intelligence.
Authors: Wang Zhu, Ishika Singh, Yuan Huang, Robin Jia, Jesse Thomason
Data augmentation via back-translation is common when pretraining Vision-and-Language Navigation (VLN) models, even though the generated instructions are noisy. But: does that noise matter? We find that nonsensical or irrelevant language instructions during pretraining can have little effect on downstream performance for both HAMT and VLN-BERT on R2R, and is still better than only using clean, human data. To underscore these results, we concoct an efficient augmentation method, Unigram + Object, which generates nonsensical instructions that nonetheless improve downstream performance. Our findings suggest that what matters for VLN R2R pretraining is the quantity of visual trajectories, not the quality of instructions.
Authors: Xiaobin Tian, Zequn Sun, Wei Hu
Entity alignment (EA) seeks identical entities in different knowledge graphs, which is a long-standing task in the database research. Recent work leverages deep learning to embed entities in vector space and align them via nearest neighbor search. Although embedding-based EA has gained marked success in recent years, it lacks explanations for alignment decisions. In this paper, we present the first framework that can generate explanations for understanding and repairing embedding-based EA results. Given an EA pair produced by an embedding model, we first compare its neighbor entities and relations to build a matching subgraph as a local explanation. We then construct an alignment dependency graph to understand the pair from an abstract perspective. Finally, we repair the pair by resolving three types of alignment conflicts based on dependency graphs. Experiments on a variety of EA datasets demonstrate the effectiveness, generalization, and robustness of our framework in explaining and repairing embedding-based EA results.
Authors: Yunshui Li, Binyuan Hui, Xiaobo Xia, Jiaxi Yang, Min Yang, Lei Zhang, Shuzheng Si, Junhao Liu, Tongliang Liu, Fei Huang, Yongbin Li
Aligning large language models(LLMs) with human is a critical step in effectively utilizing their pre-trained capabilities across a wide array of language tasks. Current instruction tuning practices often rely on expanding dataset size without a clear strategy for ensuring data quality, which can inadvertently introduce noise and degrade model performance. To address this challenge, we introduce Nuggets, a novel and efficient methodology that employs one shot learning to select high-quality instruction data from expansive datasets. Nuggets assesses the potential of individual instruction examples to act as effective one shot examples, thereby identifying those that can significantly enhance diverse task performance. Nuggets utilizes a scoring system based on the impact of candidate examples on the perplexity of a diverse anchor set, facilitating the selection of the most beneficial data for instruction tuning. Through rigorous testing on two benchmarks, including MT-Bench and Alpaca-Eval, we demonstrate that instruction tuning with the top 1% of Nuggets-curated examples substantially outperforms conventional methods that use the full dataset. These findings advocate for a data selection paradigm that prioritizes quality, offering a more efficient pathway to align LLMs with humans.
Authors: Mengzhao Jia, Can Xie, Liqiang Jing
Despite commendable achievements made by existing work, prevailing multimodal sarcasm detection studies rely more on textual content over visual information. It unavoidably induces spurious correlations between textual words and labels, thereby significantly hindering the models' generalization capability. To address this problem, we define the task of out-of-distribution (OOD) multimodal sarcasm detection, which aims to evaluate models' generalizability when the word distribution is different in training and testing settings. Moreover, we propose a novel debiasing multimodal sarcasm detection framework with contrastive learning, which aims to mitigate the harmful effect of biased textual factors for robust OOD generalization. In particular, we first design counterfactual data augmentation to construct the positive samples with dissimilar word biases and negative samples with similar word biases. Subsequently, we devise an adapted debiasing contrastive learning mechanism to empower the model to learn robust task-relevant features and alleviate the adverse effect of biased words. Extensive experiments show the superiority of the proposed framework.
Authors: Renxi Wang, Minghao Wu, Yuxia Wang, Xudong Han, Chiyu Zhang, Haonan Li
While instructions fine-tuning of large language models (LLMs) has been proven to enhance performance across various applications, the influence of the instruction dataset mixture on LLMs has not been thoroughly explored. In this study, we classify instructions into three main types: NLP downstream tasks, coding, and general chatting, and investigate their impact on LLMs. Our findings reveal that specific types of instructions are more beneficial for particular uses, while it may cause harms to other aspects, emphasizing the importance of meticulously designing the instruction mixture to maximize model performance. This study sheds light on the instruction mixture and paves the way for future research.
Authors: David Ponce, Thierry Etchegoyhen, Jesús Calleja Pérez, Harritxu Gete
The Split and Rephrase task, which consists in splitting complex sentences into a sequence of shorter grammatical sentences, while preserving the original meaning, can facilitate the processing of complex texts for humans and machines alike. In this work, we describe an approach based on large language models, which improves over the state of the art by large margins on all the major metrics for the task, on publicly available datasets. We also describe results from two human evaluations that further establish the significant improvements obtained with large language models and the viability of the approach. We evaluate different strategies, including fine-tuning pretrained language models of varying parameter size, and applying both zero-shot and few-shot in-context learning on instruction-tuned language models. Although the latter were markedly outperformed by fine-tuned models, they still achieved promising results overall. Our results thus demonstrate the strong potential of different variants of large language models for the Split and Rephrase task, using relatively small amounts of training samples and model parameters overall.
Authors: Cheng Li, Jindong Wang, Yixuan Zhang, Kaijie Zhu, Xinyi Wang, Wenxin Hou, Jianxun Lian, Fang Luo, Qiang Yang, Xing Xie
Emotion significantly impacts our daily behaviors and interactions. While recent generative AI models, such as large language models, have shown impressive performance in various tasks, it remains unclear whether they truly comprehend emotions. This paper aims to address this gap by incorporating psychological theories to gain a holistic understanding of emotions in generative AI models. Specifically, we propose three approaches: 1) EmotionPrompt to enhance AI model performance, 2) EmotionAttack to impair AI model performance, and 3) EmotionDecode to explain the effects of emotional stimuli, both benign and malignant. Through extensive experiments involving language and multi-modal models on semantic understanding, logical reasoning, and generation tasks, we demonstrate that both textual and visual EmotionPrompt can boost the performance of AI models while EmotionAttack can hinder it. Additionally, EmotionDecode reveals that AI models can comprehend emotional stimuli akin to the mechanism of dopamine in the human brain. Our work heralds a novel avenue for exploring psychology to enhance our understanding of generative AI models. This paper is an extended version of our previous work EmotionPrompt (arXiv:2307.11760).
Authors: Yijiong Yu
Although LLMs continue to iterate and improve, most open-source models still have a context window of no more than 4k, limiting their ability to handle long-context problems. Most existing open-source models for long-context chat still lack satisfactory accuracy. To address this issue, I approach it from the perspective of training data and theoretically prove that training the capability to handle long contexts requires "effective" rather than "long" data. Based on this, I propose using the "original text paraphrase" task, and successfully extend the context window of the existing model to 32k by a low-cost and effective method, achieving extremely high accuracy in multi-document-QA and surpassing all existing open-source models of the same scale. The model and training data have been open-sourced on HuggingFace(https://huggingface.co/yuyijiong/Qwen-14b-chat-yarn-32k) and WiseModel(https://wisemodel.cn/models/yuyijiong/Qwen-14b-chat-yarn-32k).
Authors: Yuyang Chai, Zhuang Li, Jiahui Liu, Lei Chen, Fei Li, Donghong Ji, Chong Teng
Despite significant advancements in multi-label text classification, the ability of existing models to generalize to novel and seldom-encountered complex concepts, which are compositions of elementary ones, remains underexplored. This research addresses this gap. By creating unique data splits across three benchmarks, we assess the compositional generalization ability of existing multi-label text classification models. Our results show that these models often fail to generalize to compositional concepts encountered infrequently during training, leading to inferior performance on tests with these new combinations. To address this, we introduce a data augmentation method that leverages two innovative text generation models designed to enhance the classification models' capacity for compositional generalization. Our experiments show that this data augmentation approach significantly improves the compositional generalization capabilities of classification models on our benchmarks, with both generation models surpassing other text generation baselines.