Authors: Tejit Pabari, Beth Tellman, Giannis Karamanolakis, Mitchell Thomas, Max Mauerman, Eugene Wu, Upmanu Lall, Marco Tedesco, Michael S Steckler, Paolo Colosio, Daniel E Osgood, Melody Braun, Jens de Bruijn, Shammun Islam
Floods cause large losses to property, life, and livelihoods across the world every year, hindering sustainable development. Safety nets to help absorb financial shocks in disasters, such as insurance, are often unavailable in regions of the world most vulnerable to floods, like Bangladesh. Index-based insurance has emerged as an affordable solution, which considers weather data or information from satellites to create a "flood index" that should correlate with the damage insured. However, existing flood event databases are often incomplete, and satellite sensors are not reliable under extreme weather conditions (e.g., because of clouds), which limits the spatial and temporal resolution of current approaches for index-based insurance.
In this work, we explore a novel approach for supporting satellite-based flood index insurance by extracting high-resolution spatio-temporal information from news media. First, we publish a dataset consisting of 40,000 news articles covering flood events in Bangladesh by 10 prominent news sources, and inundated area estimates for each division in Bangladesh collected from a satellite radar sensor. Second, we show that keyword-based models are not adequate for this novel application, while context-based classifiers cover complex and implicit flood related patterns. Third, we show that time series extracted from news media have substantial correlation Spearman's rho$=0.70 with satellite estimates of inundated area. Our work demonstrates that news media is a promising source for improving the temporal resolution and expanding the spatial coverage of the available flood damage data.
Authors: Huan Wang, Yan-Fu Li, Min Xie
Prognostics and health management (PHM) is essential for industrial operation and maintenance, focusing on predicting, diagnosing, and managing the health status of industrial systems. The emergence of the ChatGPT-Like large-scale language model (LLM) has begun to lead a new round of innovation in the AI field. It has extensively promoted the level of intelligence in various fields. Therefore, it is also expected further to change the application paradigm in industrial PHM and promote PHM to become intelligent. Although ChatGPT-Like LLMs have rich knowledge reserves and powerful language understanding and generation capabilities, they lack domain-specific expertise, significantly limiting their practicability in PHM applications. To this end, this study explores the ChatGPT-Like LLM empowered by the local knowledge base (LKB) in industrial PHM to solve the above limitations. In addition, we introduce the method and steps of combining the LKB with LLMs, including LKB preparation, LKB vectorization, prompt engineering, etc. Experimental analysis of real cases shows that combining the LKB with ChatGPT-Like LLM can significantly improve its performance and make ChatGPT-Like LLMs more accurate, relevant, and able to provide more insightful information. This can promote the development of ChatGPT-Like LLMs in industrial PHM and promote their efficiency and quality.
Authors: Buvarp Gohsh, Woods Ali, Anders Michael
The intricate hierarchical structure of syntax is fundamental to the intricate and systematic nature of human language. This study investigates the premise that language models, specifically their attention distributions, can encapsulate syntactic dependencies. We introduce Dynamic Syntax Mapping (DSM), an innovative approach for the agnostic induction of these structures. Our method diverges from traditional syntax models which rely on predefined annotation schemata. Instead, we focus on a core characteristic inherent in dependency relations: syntactic substitutability. This concept refers to the interchangeability of words within the same syntactic category at either end of a dependency. By leveraging this property, we generate a collection of syntactically invariant sentences, which serve as the foundation for our parsing framework. Our findings reveal that the use of an increasing array of substitutions notably enhances parsing precision on natural language data. Specifically, in the context of long-distance subject-verb agreement, DSM exhibits a remarkable advancement over prior methodologies. Furthermore, DSM's adaptability is demonstrated through its successful application in varied parsing scenarios, underscoring its broad applicability.
Authors: Yuhao Chen, Chloe Wong, Hanwen Yang, Juan Aguenza, Sai Bhujangari, Benthan Vu, Xun Lei, Amisha Prasad, Manny Fluss, Eric Phuong, Minghao Liu, James Davis
This study critically evaluates the mathematical proficiency of OpenAI's language model, ChatGPT, by juxtaposing its default computational capabilities against the efficiency of three prescriptive methods: strategic prompting, persona implementation, and the Chain of Thought approach. The evaluation harnessed the diverse and extensive problem sets from the MATH, GSM8K, and MMLU data-sets, which encompassing a broad spectrum of mathematical conundrums and levels of complexity. A sophisticated grading script was designed to determine the efficacy of these interventions in enhancing the model's mathematical precision. Contrary to expectations, our empirical analysis revealed that none of the trialed methods substantially improved ChatGPT's baseline performance. In some cases, these interventions inadvertently disrupted the model's response generation. This investigation concluded that while the pursuit of innovative strategies for augmenting language model performance remains crucial, the specific methods examined within this study did not induce significant improvements in ChatGPT's computational aptitude. These findings underscore the importance of further comprehensive research and exploration of novel techniques to enhance the precision and dependability of such models across diverse domains.
Authors: Abhinav Arun, Dipendra Singh Mal, Mehul Soni, Tomohiro Sawada
Recent advancements in deep learning have led to the development of powerful language models (LMs) that excel in various tasks. Despite these achievements, there is still room for improvement, particularly in enhancing reasoning abilities and incorporating multimodal data. This report investigates the potential impact of combining Chain-of-Thought (CoT) reasoning and Visual Question Answering (VQA) techniques to improve LM's accuracy in solving multiple-choice questions. By employing TextVQA and ScienceQA datasets, we assessed the effectiveness of three text embedding methods and three visual embedding approaches. Our experiments aimed to fill the gap in current research by investigating the combined impact of CoT and VQA, contributing to the understanding of how these techniques can improve the reasoning capabilities of state-of-the-art models like GPT-4. Results from our experiments demonstrated the potential of these approaches in enhancing LM's reasoning and question-answering capabilities, providing insights for further research and development in the field, and paving the way for more accurate and reliable AI systems that can handle complex reasoning tasks across multiple modalities.
Authors: Zhen Tan, Tianlong Chen, Zhenyu Zhang, Huan Liu
Large Language Models (LLMs) have achieved unprecedented breakthroughs in various natural language processing domains. However, the enigmatic ``black-box'' nature of LLMs remains a significant challenge for interpretability, hampering transparent and accountable applications. While past approaches, such as attention visualization, pivotal subnetwork extraction, and concept-based analyses, offer some insight, they often focus on either local or global explanations within a single dimension, occasionally falling short in providing comprehensive clarity. In response, we propose a novel methodology anchored in sparsity-guided techniques, aiming to provide a holistic interpretation of LLMs. Our framework, termed SparseCBM, innovatively integrates sparsity to elucidate three intertwined layers of interpretation: input, subnetwork, and concept levels. In addition, the newly introduced dimension of interpretable inference-time intervention facilitates dynamic adjustments to the model during deployment. Through rigorous empirical evaluations on real-world datasets, we demonstrate that SparseCBM delivers a profound understanding of LLM behaviors, setting it apart in both interpreting and ameliorating model inaccuracies. Codes are provided in supplements.
Authors: Xingfang Wu, Heng Li, Nobukazu Yoshioka, Hironori Washizaki, Foutse Khomh
One goal of technical online communities is to help developers find the right answer in one place. A single question can be asked in different ways with different wordings, leading to the existence of duplicate posts on technical forums. The question of how to discover and link duplicate posts has garnered the attention of both developer communities and researchers. For example, Stack Overflow adopts a voting-based mechanism to mark and close duplicate posts. However, addressing these constantly emerging duplicate posts in a timely manner continues to pose challenges. Therefore, various approaches have been proposed to detect duplicate posts on technical forum posts automatically. The existing methods suffer from limitations either due to their reliance on handcrafted similarity metrics which can not sufficiently capture the semantics of posts, or their lack of supervision to improve the performance. Additionally, the efficiency of these methods is hindered by their dependence on pair-wise feature generation, which can be impractical for large amount of data. In this work, we attempt to employ and refine the GPT-3 embeddings for the duplicate detection task. We assume that the GPT-3 embeddings can accurately represent the semantics of the posts. In addition, by training a Siamese-based network based on the GPT-3 embeddings, we obtain a latent embedding that accurately captures the duplicate relation in technical forum posts. Our experiment on a benchmark dataset confirms the effectiveness of our approach and demonstrates superior performance compared to baseline methods. When applied to the dataset we constructed with a recent Stack Overflow dump, our approach attains a Top-1, Top-5, and Top-30 accuracy of 23.1%, 43.9%, and 68.9%, respectively. With a manual study, we confirm our approach's potential of finding unlabelled duplicates on technical forums.
Authors: Jay Kejriwal, Stefan Benus, Lina M. Rojas-Barahona
Speakers tend to engage in adaptive behavior, known as entrainment, when they become similar to their interlocutor in various aspects of speaking. We present an unsupervised deep learning framework that derives meaningful representation from textual features for developing semantic entrainment. We investigate the model's performance by extracting features using different variations of the BERT model (DistilBERT and XLM-RoBERTa) and Google's universal sentence encoder (USE) embeddings on two human-human (HH) corpora (The Fisher Corpus English Part 1, Columbia games corpus) and one human-machine (HM) corpus (Voice Assistant Conversation Corpus (VACC)). In addition to semantic features we also trained DNN-based models utilizing two auditory embeddings (TRIpLet Loss network (TRILL) vectors, Low-level descriptors (LLD) features) and two units of analysis (Inter pausal unit and Turn). The results show that semantic entrainment can be assessed with our model, that models can distinguish between HH and HM interactions and that the two units of analysis for extracting acoustic features provide comparable findings.
Authors: Nishant Vishwamitra, Keyan Guo, Farhan Tajwar Romit, Isabelle Ondracek, Long Cheng, Ziming Zhao, Hongxin Hu
Online hate is an escalating problem that negatively impacts the lives of Internet users, and is also subject to rapid changes due to evolving events, resulting in new waves of online hate that pose a critical threat. Detecting and mitigating these new waves present two key challenges: it demands reasoning-based complex decision-making to determine the presence of hateful content, and the limited availability of training samples hinders updating the detection model. To address this critical issue, we present a novel framework called HATEGUARD for effectively moderating new waves of online hate. HATEGUARD employs a reasoning-based approach that leverages the recently introduced chain-of-thought (CoT) prompting technique, harnessing the capabilities of large language models (LLMs). HATEGUARD further achieves prompt-based zero-shot detection by automatically generating and updating detection prompts with new derogatory terms and targets in new wave samples to effectively address new waves of online hate. To demonstrate the effectiveness of our approach, we compile a new dataset consisting of tweets related to three recently witnessed new waves: the 2022 Russian invasion of Ukraine, the 2021 insurrection of the US Capitol, and the COVID-19 pandemic. Our studies reveal crucial longitudinal patterns in these new waves concerning the evolution of events and the pressing need for techniques to rapidly update existing moderation tools to counteract them. Comparative evaluations against state-of-the-art tools illustrate the superiority of our framework, showcasing a substantial 22.22% to 83.33% improvement in detecting the three new waves of online hate. Our work highlights the severe threat posed by the emergence of new waves of online hate and represents a paradigm shift in addressing this threat practically.
Authors: Dimitris Gkoumas, Bo Wang, Adam Tsakalidis, Maria Wolters, Arkaitz Zubiaga, Matthew Purver, Maria Liakata
Dementia affects cognitive functions of adults, including memory, language, and behaviour. Standard diagnostic biomarkers such as MRI are costly, whilst neuropsychological tests suffer from sensitivity issues in detecting dementia onset. The analysis of speech and language has emerged as a promising and non-intrusive technology to diagnose and monitor dementia. Currently, most work in this direction ignores the multi-modal nature of human communication and interactive aspects of everyday conversational interaction. Moreover, most studies ignore changes in cognitive status over time due to the lack of consistent longitudinal data. Here we introduce a novel fine-grained longitudinal multi-modal corpus collected in a natural setting from healthy controls and people with dementia over two phases, each spanning 28 sessions. The corpus consists of spoken conversations, a subset of which are transcribed, as well as typed and written thoughts and associated extra-linguistic information such as pen strokes and keystrokes. We present the data collection process and describe the corpus in detail. Furthermore, we establish baselines for capturing longitudinal changes in language across different modalities for two cohorts, healthy controls and people with dementia, outlining future research directions enabled by the corpus.
Authors: Ayush Kumar, Vijit Malik, Jithendra Vepa
Intent Detection is one of the core tasks of dialog systems. Few-shot Intent Detection is challenging due to limited number of annotated utterances for novel classes. Generalized Few-shot intent detection is more realistic but challenging setup which aims to discriminate the joint label space of both novel intents which have few examples each and existing intents consisting of enough labeled data. Large label spaces and fewer number of shots increase the complexity of the task. In this work, we employ a simple and effective method based on Natural Language Inference that leverages the semantics in the class-label names to learn and predict the novel classes. Our method achieves state-of-the-art results on 1-shot and 5-shot intent detection task with gains ranging from 2-8\% points in F1 score on four benchmark datasets. Our method also outperforms existing approaches on a more practical setting of generalized few-shot intent detection with gains up to 20% F1 score. We show that the suggested approach performs well across single and multi domain datasets with the number of class labels from as few as 7 to as high as 150.
Authors: Valentin Liévin, Christoffer Egeberg Hother, Andreas Geert Motzfeldt, Ole Winther
Although large language models (LLMs) often produce impressive outputs, it remains unclear how they perform in real-world scenarios requiring strong reasoning skills and expert domain knowledge. We set out to investigate whether close- and open-source models (GPT-3.5, LLama-2, etc.) can be applied to answer and reason about difficult real-world-based questions. We focus on three popular medical benchmarks (MedQA-USMLE, MedMCQA, and PubMedQA) and multiple prompting scenarios: Chain-of-Thought (CoT, think step-by-step), few-shot and retrieval augmentation. Based on an expert annotation of the generated CoTs, we found that InstructGPT can often read, reason and recall expert knowledge. Last, by leveraging advances in prompt engineering (few-shot and ensemble methods), we demonstrated that GPT-3.5 not only yields calibrated predictive distributions, but also reaches the passing score on three datasets: MedQA-USMLE 60.2%, MedMCQA 62.7% and PubMedQA 78.2%. Open-source models are closing the gap: Llama-2 70B also passed the MedQA-USMLE with 62.5% accuracy.
Authors: Yanran Chen, Steffen Eger
Recently proposed BERT-based evaluation metrics for text generation perform well on standard benchmarks but are vulnerable to adversarial attacks, e.g., relating to information correctness. We argue that this stems (in part) from the fact that they are models of semantic similarity. In contrast, we develop evaluation metrics based on Natural Language Inference (NLI), which we deem a more appropriate modeling. We design a preference-based adversarial attack framework and show that our NLI based metrics are much more robust to the attacks than the recent BERT-based metrics. On standard benchmarks, our NLI based metrics outperform existing summarization metrics, but perform below SOTA MT metrics. However, when combining existing metrics with our NLI metrics, we obtain both higher adversarial robustness (15%-30%) and higher quality metrics as measured on standard benchmarks (+5% to 30%).
Authors: Linyi Yang, Lifan Yuan, Leyang Cui, Wenyang Gao, Yue Zhang
Few-shot Named Entity Recognition (NER) is imperative for entity tagging in limited resource domains and thus received proper attention in recent years. Existing approaches for few-shot NER are evaluated mainly under in-domain settings. In contrast, little is known about how these inherently faithful models perform in cross-domain NER using a few labeled in-domain examples. This paper proposes a two-step rationale-centric data augmentation method to improve the model's generalization ability. Results on several datasets show that our model-agnostic method significantly improves the performance of cross-domain NER tasks compared to previous state-of-the-art methods, including the data augmentation and prompt-tuning methods. Our codes are available at https://github.com/lifan-yuan/FactMix.
Authors: Guy Dar, Mor Geva, Ankit Gupta, Jonathan Berant
Understanding Transformer-based models has attracted significant attention, as they lie at the heart of recent technological advances across machine learning. While most interpretability methods rely on running models over inputs, recent work has shown that a zero-pass approach, where parameters are interpreted directly without a forward/backward pass is feasible for some Transformer parameters, and for two-layer attention networks. In this work, we present a theoretical analysis where all parameters of a trained Transformer are interpreted by projecting them into the embedding space, that is, the space of vocabulary items they operate on. We derive a simple theoretical framework to support our arguments and provide ample evidence for its validity. First, an empirical analysis showing that parameters of both pretrained and fine-tuned models can be interpreted in embedding space. Second, we present two applications of our framework: (a) aligning the parameters of different models that share a vocabulary, and (b) constructing a classifier without training by ``translating'' the parameters of a fine-tuned classifier to parameters of a different model that was only pretrained. Overall, our findings open the door to interpretation methods that, at least in part, abstract away from model specifics and operate in the embedding space only.
Authors: Jacob Krantz, Shurjo Banerjee, Wang Zhu, Jason Corso, Peter Anderson, Stefan Lee, Jesse Thomason
We present Iterative Vision-and-Language Navigation (IVLN), a paradigm for evaluating language-guided agents navigating in a persistent environment over time. Existing Vision-and-Language Navigation (VLN) benchmarks erase the agent's memory at the beginning of every episode, testing the ability to perform cold-start navigation with no prior information. However, deployed robots occupy the same environment for long periods of time. The IVLN paradigm addresses this disparity by training and evaluating VLN agents that maintain memory across tours of scenes that consist of up to 100 ordered instruction-following Room-to-Room (R2R) episodes, each defined by an individual language instruction and a target path. We present discrete and continuous Iterative Room-to-Room (IR2R) benchmarks comprising about 400 tours each in 80 indoor scenes. We find that extending the implicit memory of high-performing transformer VLN agents is not sufficient for IVLN, but agents that build maps can benefit from environment persistence, motivating a renewed focus on map-building agents in VLN.
Authors: Yanran Chen, Steffen Eger
We consider the end-to-end abstract-to-title generation problem, exploring seven recent transformer based models (including ChatGPT) fine-tuned on more than 30k abstract-title pairs from NLP and machine learning (ML) venues. As an extension, we also consider the harder problem of generating humorous paper titles. For the latter, we compile the first large-scale humor annotated dataset for scientific papers in the NLP/ML domains, comprising almost ~2.6k titles. We evaluate all models using human and automatic metrics. Our human evaluation suggests that our best end-to-end system performs similarly to human authors (but arguably slightly worse). Generating funny titles is more difficult, however, and our automatic systems clearly underperform relative to humans and often learn dataset artefacts of humor. Finally, ChatGPT, without any fine-tuning, performs on the level of our best fine-tuned system.
Authors: Gyubok Lee, Hyeonji Hwang, Seongsu Bae, Yeonsu Kwon, Woncheol Shin, Seongjun Yang, Minjoon Seo, Jong-Yeup Kim, Edward Choi
We present a new text-to-SQL dataset for electronic health records (EHRs). The utterances were collected from 222 hospital staff members, including physicians, nurses, and insurance review and health records teams. To construct the QA dataset on structured EHR data, we conducted a poll at a university hospital and used the responses to create seed questions. We then manually linked these questions to two open-source EHR databases, MIMIC-III and eICU, and included various time expressions and held-out unanswerable questions in the dataset, which were also collected from the poll. Our dataset poses a unique set of challenges: the model needs to 1) generate SQL queries that reflect a wide range of needs in the hospital, including simple retrieval and complex operations such as calculating survival rate, 2) understand various time expressions to answer time-sensitive questions in healthcare, and 3) distinguish whether a given question is answerable or unanswerable. We believe our dataset, EHRSQL, can serve as a practical benchmark for developing and assessing QA models on structured EHR data and take a step further towards bridging the gap between text-to-SQL research and its real-life deployment in healthcare. EHRSQL is available at https://github.com/glee4810/EHRSQL.
Authors: Zhuoyuan Mao, Tetsuji Nakagawa
Large-scale language-agnostic sentence embedding models such as LaBSE (Feng et al., 2022) obtain state-of-the-art performance for parallel sentence alignment. However, these large-scale models can suffer from inference speed and computation overhead. This study systematically explores learning language-agnostic sentence embeddings with lightweight models. We demonstrate that a thin-deep encoder can construct robust low-dimensional sentence embeddings for 109 languages. With our proposed distillation methods, we achieve further improvements by incorporating knowledge from a teacher model. Empirical results on Tatoeba, United Nations, and BUCC show the effectiveness of our lightweight models. We release our lightweight language-agnostic sentence embedding models LEALLA on TensorFlow Hub.
Authors: Seonghyeon Ye, Hyeonbin Hwang, Sohee Yang, Hyeongu Yun, Yireun Kim, Minjoon Seo
In this paper, we present our finding that prepending a Task-Agnostic Prefix Prompt (TAPP) to the input improves the instruction-following ability of various Large Language Models (LLMs) during inference. TAPP is different from canonical prompts for LLMs in that it is a fixed prompt prepended to the beginning of every input regardless of the target task for zero-shot generalization. We observe that both base LLMs (i.e. not fine-tuned to follow instructions) and instruction-tuned models benefit from TAPP, resulting in 34.58% and 12.26% improvement on average, respectively. This implies that the instruction-following ability of LLMs can be improved during inference time with a fixed prompt constructed with simple heuristics. We hypothesize that TAPP assists language models to better estimate the output distribution by focusing more on the instruction of the target task during inference. In other words, such ability does not seem to be sufficiently activated in not only base LLMs but also many instruction-fine-tuned LLMs. All experiments are reproducible from https://github.com/seonghyeonye/TAPP.
Authors: Junjie Ye, Xuanting Chen, Nuo Xu, Can Zu, Zekai Shao, Shichun Liu, Yuhan Cui, Zeyang Zhou, Chao Gong, Yang Shen, Jie Zhou, Siming Chen, Tao Gui, Qi Zhang, Xuanjing Huang
GPT series models, such as GPT-3, CodeX, InstructGPT, ChatGPT, and so on, have gained considerable attention due to their exceptional natural language processing capabilities. However, despite the abundance of research on the difference in capabilities between GPT series models and fine-tuned models, there has been limited attention given to the evolution of GPT series models' capabilities over time. To conduct a comprehensive analysis of the capabilities of GPT series models, we select six representative models, comprising two GPT-3 series models (i.e., davinci and text-davinci-001) and four GPT-3.5 series models (i.e., code-davinci-002, text-davinci-002, text-davinci-003, and gpt-3.5-turbo). We evaluate their performance on nine natural language understanding (NLU) tasks using 21 datasets. In particular, we compare the performance and robustness of different models for each task under zero-shot and few-shot scenarios. Our extensive experiments reveal that the overall ability of GPT series models on NLU tasks does not increase gradually as the models evolve, especially with the introduction of the RLHF training strategy. While this strategy enhances the models' ability to generate human-like responses, it also compromises their ability to solve some tasks. Furthermore, our findings indicate that there is still room for improvement in areas such as model robustness.
Authors: Yingjian Liu, Jiang Li, Xiaoping Wang, Zhigang Zeng
Emotion Recognition in Conversation (ERC) has attracted growing attention in recent years as a result of the advancement and implementation of human-computer interface technologies. In this paper, we propose a novel approach to dependency modeling driven by Emotional Inertia and Contagion (EmotionIC) for ERC task. Our EmotionIC consists of three main components, i.e., Identity Masked Multi-Head Attention (IMMHA), Dialogue-based Gated Recurrent Unit (DiaGRU), and Skip-chain Conditional Random Field (SkipCRF). Compared to previous ERC models, EmotionIC can model a conversation more thoroughly at both the feature-extraction and classification levels. The proposed model attempts to integrate the advantages of attention- and recurrence-based methods at the feature-extraction level. Specifically, IMMHA is applied to capture identity-based global contextual dependencies, while DiaGRU is utilized to extract speaker- and temporal-aware local contextual information. At the classification level, SkipCRF can explicitly mine complex emotional flows from higher-order neighboring utterances in the conversation. Experimental results show that our method can significantly outperform the state-of-the-art models on four benchmark datasets. The ablation studies confirm that our modules can effectively model emotional inertia and contagion.
Authors: Jeffrey Lu, Ivan Rodriguez
The current trend in developing machine learning models for reading comprehension and logical reasoning tasks is focused on improving the models' abilities to understand and utilize logical rules. This work focuses on providing a novel loss function and accompanying model architecture that has more interpretable components than some other models by representing a common strategy employed by humans when given reading comprehension and logical reasoning tasks. Our strategy involves emphasizing relative accuracy over absolute accuracy and can theoretically produce the correct answer with incomplete knowledge. We examine the effectiveness of this strategy to solve reading comprehension and logical reasoning questions. The models were evaluated on the ReClor dataset, a challenging reading comprehension and logical reasoning benchmark. We propose the polytuplet loss function, which forces prioritization of learning the relative correctness of answer choices over learning the true accuracy of each choice. Our results indicate that models employing polytuplet loss outperform existing baseline models, though further research is required to quantify the benefits it may present.
Authors: Yonatan Bitton, Shlomi Cohen-Ganor, Ido Hakimi, Yoad Lewenberg, Roee Aharoni, Enav Weinreb
One of the exciting capabilities of recent language models for dialog is their ability to independently search for relevant information to ground a given dialog response. However, obtaining training data to teach models how to issue search queries is time and resource consuming. In this work, we propose q2d: an automatic data generation pipeline that generates information-seeking dialogs from questions. We prompt a large language model (PaLM) to create conversational versions of question answering datasets, and use it to improve query generation models that communicate with external search APIs to ground dialog responses. Unlike previous approaches which relied on human written dialogs with search queries, our method allows to automatically generate query-based grounded dialogs with better control and scale. Our experiments demonstrate that: (1) For query generation on the QReCC dataset, models trained on our synthetically-generated data achieve 90%--97% of the performance of models trained on the human-generated data; (2) We can successfully generate data for training dialog models in new domains without any existing dialog data as demonstrated on the multi-hop MuSiQue and Bamboogle QA datasets. (3) We perform a thorough analysis of the generated dialogs showing that humans find them of high quality and struggle to distinguish them from human-written dialogs.
Authors: Xin Cheng, Di Luo, Xiuying Chen, Lemao Liu, Dongyan Zhao, Rui Yan
With direct access to human-written reference as memory, retrieval-augmented generation has achieved much progress in a wide range of text generation tasks. Since better memory would typically prompt better generation~(we define this as primal problem). The traditional approach for memory retrieval involves selecting memory that exhibits the highest similarity to the input. However, this method is constrained by the quality of the fixed corpus from which memory is retrieved. In this paper, by exploring the duality of the primal problem: better generation also prompts better memory, we propose a novel framework, selfmem, which addresses this limitation by iteratively employing a retrieval-augmented generator to create an unbounded memory pool and using a memory selector to choose one output as memory for the subsequent generation round. This enables the model to leverage its own output, referred to as self-memory, for improved generation. We evaluate the effectiveness of selfmem on three distinct text generation tasks: neural machine translation, abstractive text summarization, and dialogue generation, under two generation paradigms: fine-tuned small model and few-shot LLM. Our approach achieves state-of-the-art results in four directions in JRC-Acquis, XSum (50.3 ROUGE-1), and BigPatent (62.9 ROUGE-1), demonstrating the potential of self-memory in enhancing retrieval-augmented generation models. Furthermore, we conduct thorough analyses of each component in the selfmem framework to identify bottlenecks and provide insights for future research.
Authors: Alexei Grinbaum, Laurynas Adomaitis
We suggest the implementation of the Dual Use Research of Concern (DURC) framework, originally designed for life sciences, to the domain of generative AI, with a specific focus on Large Language Models (LLMs). With its demonstrated advantages and drawbacks in biological research, we believe the DURC criteria can be effectively redefined for LLMs, potentially contributing to improved AI governance. Acknowledging the balance that must be struck when employing the DURC framework, we highlight its crucial political role in enhancing societal awareness of the impact of generative AI. As a final point, we offer a series of specific recommendations for applying the DURC approach to LLM research.
Authors: Michal Yarom, Yonatan Bitton, Soravit Changpinyo, Roee Aharoni, Jonathan Herzig, Oran Lang, Eran Ofek, Idan Szpektor
Automatically determining whether a text and a corresponding image are semantically aligned is a significant challenge for vision-language models, with applications in generative text-to-image and image-to-text tasks. In this work, we study methods for automatic text-image alignment evaluation. We first introduce SeeTRUE: a comprehensive evaluation set, spanning multiple datasets from both text-to-image and image-to-text generation tasks, with human judgements for whether a given text-image pair is semantically aligned. We then describe two automatic methods to determine alignment: the first involving a pipeline based on question generation and visual question answering models, and the second employing an end-to-end classification approach by finetuning multimodal pretrained models. Both methods surpass prior approaches in various text-image alignment tasks, with significant improvements in challenging cases that involve complex composition or unnatural images. Finally, we demonstrate how our approaches can localize specific misalignments between an image and a given text, and how they can be used to automatically re-rank candidates in text-to-image generation.
Authors: Joshua Ainslie, James Lee-Thorp, Michiel de Jong, Yury Zemlyanskiy, Federico Lebrón, Sumit Sanghai
Multi-query attention (MQA), which only uses a single key-value head, drastically speeds up decoder inference. However, MQA can lead to quality degradation, and moreover it may not be desirable to train a separate model just for faster inference. We (1) propose a recipe for uptraining existing multi-head language model checkpoints into models with MQA using 5% of original pre-training compute, and (2) introduce grouped-query attention (GQA), a generalization of multi-query attention which uses an intermediate (more than one, less than number of query heads) number of key-value heads. We show that uptrained GQA achieves quality close to multi-head attention with comparable speed to MQA.
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: Qin Liu, Fei Wang, Chaowei Xiao, Muhao Chen
Language models are often at risk of diverse backdoor attacks, especially data poisoning. Thus, it is important to investigate defense solutions for addressing them. Existing backdoor defense methods mainly focus on backdoor attacks with explicit triggers, leaving a universal defense against various backdoor attacks with diverse triggers largely unexplored. In this paper, we propose an end-to-end ensemble-based backdoor defense framework, DPoE (Denoised Product-of-Experts), which is inspired by the shortcut nature of backdoor attacks, to defend various backdoor attacks. DPoE consists of two models: a shallow model that captures the backdoor shortcuts and a main model that is prevented from learning the backdoor shortcuts. To address the label flip caused by backdoor attackers, DPoE incorporates a denoising design. Experiments on SST-2 dataset show that DPoE significantly improves the defense performance against various types of backdoor triggers including word-level, sentence-level, and syntactic triggers. Furthermore, DPoE is also effective under a more challenging but practical setting that mixes multiple types of trigger.
Authors: Guhao Feng, Bohang Zhang, Yuntian Gu, Haotian Ye, Di He, Liwei Wang
Recent studies have discovered that Chain-of-Thought prompting (CoT) can dramatically improve the performance of Large Language Models (LLMs), particularly when dealing with complex tasks involving mathematics or reasoning. Despite the enormous empirical success, the underlying mechanisms behind CoT and how it unlocks the potential of LLMs remain elusive. In this paper, we take a first step towards theoretically answering these questions. Specifically, we examine the expressivity of LLMs with CoT in solving fundamental mathematical and decision-making problems. By using circuit complexity theory, we first give impossibility results showing that bounded-depth Transformers are unable to directly produce correct answers for basic arithmetic/equation tasks unless the model size grows super-polynomially with respect to the input length. In contrast, we then prove by construction that autoregressive Transformers of constant size suffice to solve both tasks by generating CoT derivations using a commonly used math language format. Moreover, we show LLMs with CoT can handle a general class of decision-making problems known as Dynamic Programming, thus justifying its power in tackling complex real-world tasks. Finally, an extensive set of experiments show that, while Transformers always fail to directly predict the answers, they can consistently learn to generate correct solutions step-by-step given sufficient CoT demonstrations.
Authors: Lianmin Zheng, Wei-Lin Chiang, Ying Sheng, Siyuan Zhuang, Zhanghao Wu, Yonghao Zhuang, Zi Lin, Zhuohan Li, Dacheng Li, Eric P. Xing, Hao Zhang, Joseph E. Gonzalez, Ion Stoica
Evaluating large language model (LLM) based chat assistants is challenging due to their broad capabilities and the inadequacy of existing benchmarks in measuring human preferences. To address this, we explore using strong LLMs as judges to evaluate these models on more open-ended questions. We examine the usage and limitations of LLM-as-a-judge, including position, verbosity, and self-enhancement biases, as well as limited reasoning ability, and propose solutions to mitigate some of them. We then verify the agreement between LLM judges and human preferences by introducing two benchmarks: MT-bench, a multi-turn question set; and Chatbot Arena, a crowdsourced battle platform. Our results reveal that strong LLM judges like GPT-4 can match both controlled and crowdsourced human preferences well, achieving over 80% agreement, the same level of agreement between humans. Hence, LLM-as-a-judge is a scalable and explainable way to approximate human preferences, which are otherwise very expensive to obtain. Additionally, we show our benchmark and traditional benchmarks complement each other by evaluating several variants of LLaMA and Vicuna. The MT-bench questions, 3K expert votes, and 30K conversations with human preferences are publicly available at https://github.com/lm-sys/FastChat/tree/main/fastchat/llm_judge.
Authors: Haiyang Sun, Fulin Zhang, Zheng Lian, Yingying Guo, Shilei Zhang
Speech emotion recognition aims to identify and analyze emotional states in target speech similar to humans. Perfect emotion recognition can greatly benefit a wide range of human-machine interaction tasks. Inspired by the human process of understanding emotions, we demonstrate that compared to quantized modeling, understanding speech content from a continuous perspective, akin to human-like comprehension, enables the model to capture more comprehensive emotional information. Additionally, considering that humans adjust their perception of emotional words in textual semantic based on certain cues present in speech, we design a novel search space and search for the optimal fusion strategy for the two types of information. Experimental results further validate the significance of this perception adjustment. Building on these observations, we propose a novel framework called Multiple perspectives Fusion Architecture Search (MFAS). Specifically, we utilize continuous-based knowledge to capture speech semantic and quantization-based knowledge to learn textual semantic. Then, we search for the optimal fusion strategy for them. Experimental results demonstrate that MFAS surpasses existing models in comprehensively capturing speech emotion information and can automatically adjust fusion strategy.
Authors: Lina Bariah, Qiyang Zhao, Hang Zou, Yu Tian, Faouzi Bader, Merouane Debbah
The evolution of generative artificial intelligence (GenAI) constitutes a turning point in reshaping the future of technology in different aspects. Wireless networks in particular, with the blooming of self-evolving networks, represent a rich field for exploiting GenAI and reaping several benefits that can fundamentally change the way how wireless networks are designed and operated nowadays. To be specific, large GenAI models are envisioned to open up a new era of autonomous wireless networks, in which multi-modal GenAI models trained over various Telecom data, can be fine-tuned to perform several downstream tasks, eliminating the need for building and training dedicated AI models for each specific task and paving the way for the realization of artificial general intelligence (AGI)-empowered wireless networks. In this article, we aim to unfold the opportunities that can be reaped from integrating large GenAI models into the Telecom domain. In particular, we first highlight the applications of large GenAI models in future wireless networks, defining potential use-cases and revealing insights on the associated theoretical and practical challenges. Furthermore, we unveil how 6G can open up new opportunities through connecting multiple on-device large GenAI models, and hence, paves the way to the collective intelligence paradigm. Finally, we put a forward-looking vision on how large GenAI models will be the key to realize self-evolving networks.
Authors: Tianchen Yang, Qifan Zhang, Zhaoyang Sun, Yubo Hou
Language models have been increasingly popular for automatic creativity assessment, generating semantic distances to objectively measure the quality of creative ideas. However, there is currently a lack of an automatic assessment system for evaluating creative ideas in the Chinese language. To address this gap, we developed TransDis, a scoring system using transformer-based language models, capable of providing valid originality (quality) and flexibility (variety) scores for Alternative Uses Task (AUT) responses in Chinese. Study 1 demonstrated that the latent model-rated originality factor, comprised of three transformer-based models, strongly predicted human originality ratings, and the model-rated flexibility strongly correlated with human flexibility ratings as well. Criterion validity analyses indicated that model-rated originality and flexibility positively correlated to other creativity measures, demonstrating similar validity to human ratings. Study 2 & 3 showed that TransDis effectively distinguished participants instructed to provide creative vs. common uses (Study 2) and participants instructed to generate ideas in a flexible vs. persistent way (Study 3). Our findings suggest that TransDis can be a reliable and low-cost tool for measuring idea originality and flexibility in Chinese language, potentially paving the way for automatic creativity assessment in other languages. We offer an open platform to compute originality and flexibility for AUT responses in Chinese and over 50 other languages (https://osf.io/59jv2/).
Authors: Likang Wu, Zhaopeng Qiu, Zhi Zheng, Hengshu Zhu, Enhong Chen
Large Language Models (LLMs) have revolutionized natural language processing tasks, demonstrating their exceptional capabilities in various domains. However, their potential for behavior graph understanding in job recommendations remains largely unexplored. This paper focuses on unveiling the capability of large language models in understanding behavior graphs and leveraging this understanding to enhance recommendations in online recruitment, including the promotion of out-of-distribution (OOD) application. We present a novel framework that harnesses the rich contextual information and semantic representations provided by large language models to analyze behavior graphs and uncover underlying patterns and relationships. Specifically, we propose a meta-path prompt constructor that leverages LLM recommender to understand behavior graphs for the first time and design a corresponding path augmentation module to alleviate the prompt bias introduced by path-based sequence input. By leveraging this capability, our framework enables personalized and accurate job recommendations for individual users. We evaluate the effectiveness of our approach on a comprehensive dataset and demonstrate its ability to improve the relevance and quality of recommended quality. This research not only sheds light on the untapped potential of large language models but also provides valuable insights for developing advanced recommendation systems in the recruitment market. The findings contribute to the growing field of natural language processing and offer practical implications for enhancing job search experiences. We release the code at https://github.com/WLiK/GLRec.
Authors: Lana Touma, Mohammad Al Horani, Manar Tailouni, Anas Dahabiah, Khloud Al Jallad
Automatic Deception Detection has been a hot research topic for a long time, using machine learning and deep learning to automatically detect deception, brings new light to this old field. In this paper, we proposed a voting-based method for automatic deception detection from videos using audio, visual and lexical features. Experiments were done on two datasets, the Real-life trial dataset by Michigan University and the Miami University deception detection dataset. Video samples were split into frames of images, audio, and manuscripts. Our Voting-based Multimodal proposed solution consists of three models. The first model is CNN for detecting deception from images, the second model is Support Vector Machine (SVM) on Mel spectrograms for detecting deception from audio and the third model is Word2Vec on Support Vector Machine (SVM) for detecting deception from manuscripts. Our proposed solution outperforms state of the art. Best results achieved on images, audio and text were 97%, 96%, 92% respectively on Real-Life Trial Dataset, and 97%, 82%, 73% on video, audio and text respectively on Miami University Deception Detection.
Authors: Zijie Zeng, Lele Sha, Yuheng Li, Kaixun Yang, Dragan Gašević, Guanliang Chen
The recent large language models (LLMs), e.g., ChatGPT, have been able to generate human-like and fluent responses when provided with specific instructions. While admitting the convenience brought by technological advancement, educators also have concerns that students might leverage LLMs to complete their writing assignments and pass them off as their original work. Although many AI content detection studies have been conducted as a result of such concerns, most of these prior studies modeled AI content detection as a classification problem, assuming that a text is either entirely human-written or entirely AI-generated. In this study, we investigated AI content detection in a rarely explored yet realistic setting where the text to be detected is collaboratively written by human and generative LLMs (i.e., hybrid text). We first formalized the detection task as identifying the transition points between human-written content and AI-generated content from a given hybrid text (boundary detection). Then we proposed a two-step approach where we (1) separated AI-generated content from human-written content during the encoder training process; and (2) calculated the distances between every two adjacent prototypes and assumed that the boundaries exist between the two adjacent prototypes that have the furthest distance from each other. Through extensive experiments, we observed the following main findings: (1) the proposed approach consistently outperformed the baseline methods across different experiment settings; (2) the encoder training process can significantly boost the performance of the proposed approach; (3) when detecting boundaries for single-boundary hybrid essays, the proposed approach could be enhanced by adopting a relatively large prototype size, leading to a 22% improvement in the In-Domain evaluation and an 18% improvement in the Out-of-Domain evaluation.
Authors: Mohammadali Sefidi Esfahani, Mohammad Akbari
Social platforms have emerged as crucial platforms for disseminating information and discussing real-life social events, offering researchers an excellent opportunity to design and implement novel event detection frameworks. However, most existing approaches only exploit keyword burstiness or network structures to detect unspecified events. Thus, they often need help identifying unknown events regarding the challenging nature of events and social data. Social data, e.g., tweets, is characterized by misspellings, incompleteness, word sense ambiguation, irregular language, and variation in aspects of opinions. Moreover, extracting discriminative features and patterns for evolving events by exploiting the limited structural knowledge is almost infeasible. To address these challenges, in this paper, we propose a novel framework, namely EnrichEvent, that leverages the linguistic and contextual representations of streaming social data. In particular, we leverage contextual and linguistic knowledge to detect semantically related tweets and enhance the effectiveness of the event detection approaches. Eventually, our proposed framework produces cluster chains for each event to show the evolving variation of the event through time. We conducted extensive experiments to evaluate our framework, validating its high performance and effectiveness in detecting and distinguishing unspecified social events.
Authors: Xian Shi, Yexin Yang, Zerui Li, Yanni Chen, Zhifu Gao, Shiliang Zhang
Hotword customization is one of the concerned issues remained in ASR field - it is of value to enable users of ASR systems to customize names of entities, persons and other phrases to obtain better experience. The past few years have seen effective modeling strategies for ASR contextualization developed, but they still exhibit space for improvement about training stability and the invisible activation process. In this paper we propose Semantic-Augmented Contextual-Paraformer (SeACo-Paraformer) a novel NAR based ASR system with flexible and effective hotword customization ability. It possesses the advantages of AED-based model's accuracy, NAR model's efficiency, and explicit customization capacity of superior performance. Through extensive experiments with 50,000 hours of industrial big data, our proposed model outperforms strong baselines in customization. Besides, we explore an efficient way to filter large-scale incoming hotwords for further improvement. The industrial models compared, source codes and two hotword test sets are all open source.
Authors: Yonatan Bitton, Hritik Bansal, Jack Hessel, Rulin Shao, Wanrong Zhu, Anas Awadalla, Josh Gardner, Rohan Taori, Ludwig Schmidt
We introduce VisIT-Bench (Visual InsTruction Benchmark), a benchmark for evaluation of instruction-following vision-language models for real-world use. Our starting point is curating 70 'instruction families' that we envision instruction tuned vision-language models should be able to address. Extending beyond evaluations like VQAv2 and COCO, tasks range from basic recognition to game playing and creative generation. Following curation, our dataset comprises 592 test queries, each with a human-authored instruction-conditioned caption. These descriptions surface instruction-specific factors, e.g., for an instruction asking about the accessibility of a storefront for wheelchair users, the instruction-conditioned caption describes ramps/potential obstacles. These descriptions enable 1) collecting human-verified reference outputs for each instance; and 2) automatic evaluation of candidate multimodal generations using a text-only LLM, aligning with human judgment. We quantify quality gaps between models and references using both human and automatic evaluations; e.g., the top-performing instruction-following model wins against the GPT-4 reference in just 27% of the comparison. VisIT-Bench is dynamic to participate, practitioners simply submit their model's response on the project website; Data, code and leaderboard is available at visit-bench.github.io.
Authors: Yu Wang, Nedim Lipka, Ryan A. Rossi, Alexa Siu, Ruiyi Zhang, Tyler Derr
The `pre-train, prompt, predict' paradigm of large language models (LLMs) has achieved remarkable success in open-domain question answering (OD-QA). However, few works explore this paradigm in the scenario of multi-document question answering (MD-QA), a task demanding a thorough understanding of the logical associations among the contents and structures of different documents. To fill this crucial gap, we propose a Knowledge Graph Prompting (KGP) method to formulate the right context in prompting LLMs for MD-QA, which consists of a graph construction module and a graph traversal module. For graph construction, we create a knowledge graph (KG) over multiple documents with nodes symbolizing passages or document structures (e.g., pages/tables), and edges denoting the semantic/lexical similarity between passages or intra-document structural relations. For graph traversal, we design an LLM-based graph traversal agent that navigates across nodes and gathers supporting passages assisting LLMs in MD-QA. The constructed graph serves as the global ruler that regulates the transitional space among passages and reduces retrieval latency. Concurrently, the graph traversal agent acts as a local navigator that gathers pertinent context to progressively approach the question and guarantee retrieval quality. Extensive experiments underscore the efficacy of KGP for MD-QA, signifying the potential of leveraging graphs in enhancing the prompt design for LLMs. Our code: https://github.com/YuWVandy/KG-LLM-MDQA.
Authors: Shuang Li, Jiangjie Chen, Siyu Yuan, Xinyi Wu, Hao Yang, Shimin Tao, Yanghua Xiao
To translate well, machine translation (MT) systems and general-purposed language models (LMs) need a deep understanding of both source and target languages and cultures. Therefore, idioms, with their non-compositional nature, pose particular challenges for Transformer-based systems, as literal translations often miss the intended meaning. Traditional methods, which replace idioms using existing knowledge bases (KBs), often lack scale and context awareness. Addressing these challenges, our approach prioritizes context awareness and scalability, allowing for offline storage of idioms in a manageable KB size. This ensures efficient serving with smaller models and provides a more comprehensive understanding of idiomatic expressions. We introduce a multilingual idiom KB (IdiomKB) developed using large LMs to address this. This KB facilitates better translation by smaller models, such as BLOOMZ (7.1B), Alpaca (7B), and InstructGPT (6.7B), by retrieving idioms' figurative meanings. We present a novel, GPT-4-powered metric for human-aligned evaluation, demonstrating that IdiomKB considerably boosts model performance. Human evaluations further validate our KB's quality.
Authors: Scott L. Fleming, Alejandro Lozano, William J. Haberkorn, Jenelle A. Jindal, Eduardo P. Reis, Rahul Thapa, Louis Blankemeier, Julian Z. Genkins, Ethan Steinberg, Ashwin Nayak, Birju S. Patel, Chia-Chun Chiang, Alison Callahan, Zepeng Huo, Sergios Gatidis, Scott J. Adams, Oluseyi Fayanju, Shreya J. Shah, Thomas Savage, Ethan Goh, Akshay S. Chaudhari, Nima Aghaeepour, Christopher Sharp, Michael A. Pfeffer, Percy Liang, Jonathan H. Chen, Keith E. Morse, Emma P. Brunskill, Jason A. Fries, Nigam H. Shah
The ability of large language models (LLMs) to follow natural language instructions with human-level fluency suggests many opportunities in healthcare to reduce administrative burden and improve quality of care. However, evaluating LLMs on realistic text generation tasks for healthcare remains challenging. Existing question answering datasets for electronic health record (EHR) data fail to capture the complexity of information needs and documentation burdens experienced by clinicians. To address these challenges, we introduce MedAlign, a benchmark dataset of 983 natural language instructions for EHR data. MedAlign is curated by 15 clinicians (7 specialities), includes clinician-written reference responses for 303 instructions, and provides 276 longitudinal EHRs for grounding instruction-response pairs. We used MedAlign to evaluate 6 general domain LLMs, having clinicians rank the accuracy and quality of each LLM response. We found high error rates, ranging from 35% (GPT-4) to 68% (MPT-7B-Instruct), and an 8.3% drop in accuracy moving from 32k to 2k context lengths for GPT-4. Finally, we report correlations between clinician rankings and automated natural language generation metrics as a way to rank LLMs without human review. We make MedAlign available under a research data use agreement to enable LLM evaluations on tasks aligned with clinician needs and preferences.
Authors: Takuma Udagawa, Masayuki Suzuki, Gakuto Kurata, Masayasu Muraoka, George Saon
Transferring the knowledge of large language models (LLMs) is a promising technique to incorporate linguistic knowledge into end-to-end automatic speech recognition (ASR) systems. However, existing works only transfer a single representation of LLM (e.g. the last layer of pretrained BERT), while the representation of a text is inherently non-unique and can be obtained variously from different layers, contexts and models. In this work, we explore a wide range of techniques to obtain and transfer multiple representations of LLMs into a transducer-based ASR system. While being conceptually simple, we show that transferring multiple representations of LLMs can be an effective alternative to transferring only a single representation.
Authors: Adel Khorramrouz, Sujan Dutta, Arka Dutta, Ashiqur R. KhudaBukhsh
This paper conducts a robustness audit of the safety feedback of PaLM 2 through a novel toxicity rabbit hole framework introduced here. Starting with a stereotype, the framework instructs PaLM 2 to generate more toxic content than the stereotype. Every subsequent iteration it continues instructing PaLM 2 to generate more toxic content than the previous iteration until PaLM 2 safety guardrails throw a safety violation. Our experiments uncover highly disturbing antisemitic, Islamophobic, racist, homophobic, and misogynistic (to list a few) generated content that PaLM 2 safety guardrails do not evaluate as highly unsafe. We briefly discuss the generalizability of this framework across eight other large language models.
Authors: Zeyuan Allen-Zhu, Yuanzhi Li
Large language models (LLMs) can store a vast amount of world knowledge, often extractable via question-answering (e.g., "What is Abraham Lincoln's birthday?"). However, do they answer such questions based on exposure to similar questions during training (i.e., cheating), or by genuinely learning to extract knowledge from sources like Wikipedia?
In this paper, we investigate this issue using a controlled biography dataset. We find a strong correlation between the model's ability to extract knowledge and various diversity measures of the training data. $\textbf{Essentially}$, for knowledge to be reliably extracted, it must be sufficiently augmented (e.g., through paraphrasing, sentence shuffling) $\textit{during pretraining}$. Without such augmentation, knowledge may be memorized but not extractable, leading to 0% accuracy, regardless of subsequent instruction fine-tuning.
To understand why this occurs, we employ (nearly) linear probing to demonstrate a strong connection between the observed correlation and how the model internally encodes knowledge -- whether it is linearly encoded in the hidden embeddings of entity names or distributed across other token embeddings in the training text.
This paper provides $\textbf{several key recommendations for LLM pretraining in the industry}$: (1) rewrite the pretraining data -- using small, auxiliary models -- to provide knowledge augmentation, and (2) incorporate more instruction-finetuning data into the pretraining stage before it becomes too late.
Authors: Haoyi Xiong, Jiang Bian, Sijia Yang, Xiaofei Zhang, Linghe Kong, Daqing Zhang
Large language models (LLMs) have become phenomenally surging, since 2018--two decades after introducing context-awareness into computing systems. Through taking into account the situations of ubiquitous devices, users and the societies, context-aware computing has enabled a wide spectrum of innovative applications, such as assisted living, location-based social network services and so on. To recognize contexts and make decisions for actions accordingly, various artificial intelligence technologies, such as Ontology and OWL, have been adopted as representations for context modeling and reasoning. Recently, with the rise of LLMs and their improved natural language understanding and reasoning capabilities, it has become feasible to model contexts using natural language and perform context reasoning by interacting with LLMs such as ChatGPT and GPT-4. In this tutorial, we demonstrate the use of texts, prompts, and autonomous agents (AutoAgents) that enable LLMs to perform context modeling and reasoning without requiring fine-tuning of the model. We organize and introduce works in the related field, and name this computing paradigm as the LLM-driven Context-aware Computing (LCaC). In the LCaC paradigm, users' requests, sensors reading data, and the command to actuators are supposed to be represented as texts. Given the text of users' request and sensor data, the AutoAgent models the context by prompting and sends to the LLM for context reasoning. LLM generates a plan of actions and responds to the AutoAgent, which later follows the action plan to foster context-awareness. To prove the concepts, we use two showcases--(1) operating a mobile z-arm in an apartment for assisted living, and (2) planning a trip and scheduling the itinerary in a context-aware and personalized manner.
Authors: Siyuan Cheng, Bozhong Tian, Qingbin Liu, Xi Chen, Yongheng Wang, Huajun Chen, Ningyu Zhang
In this paper, we focus on editing Multimodal Large Language Models (MLLMs). Compared to editing single-modal LLMs, multimodal model editing is more challenging, which demands a higher level of scrutiny and careful consideration in the editing process. To facilitate research in this area, we construct a new benchmark, dubbed MMEdit, for editing multimodal LLMs and establishing a suite of innovative metrics for evaluation. We conduct comprehensive experiments involving various model editing baselines and analyze the impact of editing different components for multimodal LLMs. Empirically, we notice that previous baselines can implement editing multimodal LLMs to some extent, but the effect is still barely satisfactory, indicating the potential difficulty of this task. We hope that our work can provide the NLP community with insights. Code and dataset are available in https://github.com/zjunlp/EasyEdit.
Authors: Junjie Ye, Jie Zhou, Junfeng Tian, Rui Wang, Qi Zhang, Tao Gui, Xuanjing Huang
Recently, Target-oriented Multimodal Sentiment Classification (TMSC) has gained significant attention among scholars. However, current multimodal models have reached a performance bottleneck. To investigate the causes of this problem, we perform extensive empirical evaluation and in-depth analysis of the datasets to answer the following questions: Q1: Are the modalities equally important for TMSC? Q2: Which multimodal fusion modules are more effective? Q3: Do existing datasets adequately support the research? Our experiments and analyses reveal that the current TMSC systems primarily rely on the textual modality, as most of targets' sentiments can be determined solely by text. Consequently, we point out several directions to work on for the TMSC task in terms of model design and dataset construction. The code and data can be found in https://github.com/Junjie-Ye/RethinkingTMSC.
Authors: Ehsan Latif, Xiaoming Zhai
This study highlights the potential of fine-tuned ChatGPT (GPT-3.5) for automatically scoring student written constructed responses using example assessment tasks in science education. Recent studies on OpenAI's generative model GPT-3.5 proved its superiority in predicting the natural language with high accuracy and human-like responses. GPT-3.5 has been trained over enormous online language materials such as journals and Wikipedia; therefore, more than direct usage of pre-trained GPT-3.5 is required for automatic scoring as students utilize a different language than trained material. These imply that a domain-specific model, fine-tuned over data for specific tasks, can enhance model performance. In this study, we fine-tuned GPT-3.5 on six assessment tasks with a diverse dataset of middle-school and high-school student responses and expert scoring. The six tasks comprise two multi-label and four multi-class assessment tasks. We compare the performance of fine-tuned GPT-3.5 with the fine-tuned state-of-the-art Google's generated language model, BERT. The results show that in-domain training corpora constructed from science questions and responses for BERT achieved average accuracy = 0.838, SD = 0.069. GPT-3.5 shows a remarkable average increase (9.1%) in automatic scoring accuracy (mean = 9.15, SD = 0.042) for the six tasks, p =0.001 < 0.05. Specifically, for multi-label tasks (item 1 with 5 labels; item 2 with 10 labels), GPT-3.5 achieved significantly higher scoring accuracy than BERT across all the labels, with the second item achieving a 7.1% increase. The average scoring increase for the four multi-class items for GPT-3.5 was 10.6% compared to BERT. Our study confirmed the effectiveness of fine-tuned GPT-3.5 for automatic scoring of student responses on domain-specific data in education with high accuracy. We have released fine-tuned models for public use and community engagement.
Authors: Yuchen Xiao, Yanchao Sun, Mengda Xu, Udari Madhushani, Jared Vann, Deepeka Garg, Sumitra Ganesh
Recent advancements in large language models (LLMs) have exhibited promising performance in solving sequential decision-making problems. By imitating few-shot examples provided in the prompts (i.e., in-context learning), an LLM agent can interact with an external environment and complete given tasks without additional training. However, such few-shot examples are often insufficient to generate high-quality solutions for complex and long-horizon tasks, while the limited context length cannot consume larger-scale demonstrations. To this end, we propose an offline learning framework that utilizes offline data at scale (e.g, logs of human interactions) to facilitate the in-context learning performance of LLM agents. We formally define LLM-powered policies with both text-based approaches and code-based approaches. We then introduce an Offline Data-driven Discovery and Distillation (O3D) framework to improve LLM-powered policies without finetuning. O3D automatically discovers reusable skills and distills generalizable knowledge across multiple tasks based on offline interaction data, advancing the capability of solving downstream tasks. Empirical results under two interactive decision-making benchmarks (ALFWorld and WebShop) demonstrate that O3D can notably enhance the decision-making capabilities of LLMs through the offline discovery and distillation process, and consistently outperform baselines across various LLMs with both text-based-policy and code-based-policy.
Authors: Weijie Xu, Wenxiang Hu, Fanyou Wu, Srinivasan Sengamedu
In the burgeoning field of natural language processing (NLP), Neural Topic Models (NTMs) , Large Language Models (LLMs) and Diffusion model have emerged as areas of significant research interest. Despite this, NTMs primarily utilize contextual embeddings from LLMs, which are not optimal for clustering or capable for topic based text generation. NTMs have never been combined with diffusion model for text generation. Our study addresses these gaps by introducing a novel framework named Diffusion-Enhanced Topic Modeling using Encoder-Decoder-based LLMs (DeTiME). DeTiME leverages Encoder-Decoder-based LLMs to produce highly clusterable embeddings that could generate topics that exhibit both superior clusterability and enhanced semantic coherence compared to existing methods. Additionally, by exploiting the power of diffusion model, our framework also provides the capability to do topic based text generation. This dual functionality allows users to efficiently produce highly clustered topics and topic based text generation simultaneously. DeTiME's potential extends to generating clustered embeddings as well. Notably, our proposed framework(both encoder-decoder based LLM and diffusion model) proves to be efficient to train and exhibits high adaptability to other LLMs and diffusion model, demonstrating its potential for a wide array of applications.
Authors: Seongsu Bae, Daeun Kyung, Jaehee Ryu, Eunbyeol Cho, Gyubok Lee, Sunjun Kweon, Jungwoo Oh, Lei Ji, Eric I-Chao Chang, Tackeun Kim, Edward Choi
Electronic Health Records (EHRs), which contain patients' medical histories in various multi-modal formats, often overlook the potential for joint reasoning across imaging and table modalities underexplored in current EHR Question Answering (QA) systems. In this paper, we introduce EHRXQA, a novel multi-modal question answering dataset combining structured EHRs and chest X-ray images. To develop our dataset, we first construct two uni-modal resources: 1) The MIMIC-CXR-VQA dataset, our newly created medical visual question answering (VQA) benchmark, specifically designed to augment the imaging modality in EHR QA, and 2) EHRSQL (MIMIC-IV), a refashioned version of a previously established table-based EHR QA dataset. By integrating these two uni-modal resources, we successfully construct a multi-modal EHR QA dataset that necessitates both uni-modal and cross-modal reasoning. To address the unique challenges of multi-modal questions within EHRs, we propose a NeuralSQL-based strategy equipped with an external VQA API. This pioneering endeavor enhances engagement with multi-modal EHR sources and we believe that our dataset can catalyze advances in real-world medical scenarios such as clinical decision-making and research. EHRXQA is available at https://github.com/baeseongsu/ehrxqa.
Authors: Ziyi Ye, Qingyao Ai, Yiqun Liu, Min Zhang, Christina Lioma, Tuukka Ruotsalo
Generating human language through non-invasive brain-computer interfaces (BCIs) has the potential to unlock many applications, such as serving disabled patients and improving communication. Currently, however, generating language via BCIs has been previously successful only within a classification setup for selecting pre-generated sentence continuation candidates with the most likely cortical semantic representation. Inspired by recent research that revealed associations between the brain and the large computational language models, we propose a generative language BCI that utilizes the capacity of a large language model (LLM) jointly with a semantic brain decoder to directly generate language from functional magnetic resonance imaging (fMRI) input. The proposed model can generate coherent language sequences aligned with the semantic content of visual or auditory language stimuli perceived, without prior knowledge of any pre-generated candidates. We compare the language generated from the presented model with a random control, pre-generated language selection approach, and a standard LLM, which generates common coherent text solely based on the next word likelihood according to statistical language training data. The proposed model is found to generate language that is more aligned with semantic stimulus in response to which brain input is sampled. Our findings demonstrate the potential and feasibility of employing BCIs in direct language generation.
Authors: Panfeng Li, Mohamed Abouelenien, Rada Mihalcea
Deception detection is gaining increasing interest due to ethical and security concerns. This paper explores the application of convolutional neural networks for the purpose of multimodal deception detection. We use a dataset built by interviewing 104 subjects about two topics, with one truthful and one falsified response from each subject about each topic. In particular, we make three main contributions. First, we extract linguistic and physiological features from this data to train and construct the neural network models. Second, we propose a fused convolutional neural network model using both modalities in order to achieve an improved overall performance. Third, we compare our new approach with earlier methods designed for multimodal deception detection. We find that our system outperforms regular classification methods; our results indicate the feasibility of using neural networks for deception detection even in the presence of limited amounts of data.
Authors: Bingkang Shi, Xiaodan Zhang, Dehan Kong, Yulei Wu, Zongzhen Liu, Honglei Lyu, Longtao Huang
The social biases and unwelcome stereotypes revealed by pretrained language models are becoming obstacles to their application. Compared to numerous debiasing methods targeting word level, there has been relatively less attention on biases present at phrase level, limiting the performance of debiasing in discipline domains. In this paper, we propose an automatic multi-token debiasing pipeline called \textbf{General Phrase Debiaser}, which is capable of mitigating phrase-level biases in masked language models. Specifically, our method consists of a \textit{phrase filter stage} that generates stereotypical phrases from Wikipedia pages as well as a \textit{model debias stage} that can debias models at the multi-token level to tackle bias challenges on phrases. The latter searches for prompts that trigger model's bias, and then uses them for debiasing. State-of-the-art results on standard datasets and metrics show that our approach can significantly reduce gender biases on both career and multiple disciplines, across models with varying parameter sizes.
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: Yingdong Hu, Fanqi Lin, Tong Zhang, Li Yi, Yang Gao
In this study, we are interested in imbuing robots with the capability of physically-grounded task planning. Recent advancements have shown that large language models (LLMs) possess extensive knowledge useful in robotic tasks, especially in reasoning and planning. However, LLMs are constrained by their lack of world grounding and dependence on external affordance models to perceive environmental information, which cannot jointly reason with LLMs. We argue that a task planner should be an inherently grounded, unified multimodal system. To this end, we introduce Robotic Vision-Language Planning (ViLa), a novel approach for long-horizon robotic planning that leverages vision-language models (VLMs) to generate a sequence of actionable steps. ViLa directly integrates perceptual data into its reasoning and planning process, enabling a profound understanding of commonsense knowledge in the visual world, including spatial layouts and object attributes. It also supports flexible multimodal goal specification and naturally incorporates visual feedback. Our extensive evaluation, conducted in both real-robot and simulated environments, demonstrates ViLa's superiority over existing LLM-based planners, highlighting its effectiveness in a wide array of open-world manipulation tasks.
Authors: Kamyar Zeinalipour, Mohamed Zaky Saad, Marco Maggini, Marco Gori
This paper presents the first Arabic crossword puzzle generator driven by advanced AI technology. Leveraging cutting-edge large language models including GPT4, GPT3-Davinci, GPT3-Curie, GPT3-Babbage, GPT3-Ada, and BERT, the system generates distinctive and challenging clues. Based on a dataset comprising over 50,000 clue-answer pairs, the generator employs fine-tuning, few/zero-shot learning strategies, and rigorous quality-checking protocols to enforce the generation of high-quality clue-answer pairs. Importantly, educational crosswords contribute to enhancing memory, expanding vocabulary, and promoting problem-solving skills, thereby augmenting the learning experience through a fun and engaging approach, reshaping the landscape of traditional learning methods. The overall system can be exploited as a powerful educational tool that amalgamates AI and innovative learning techniques, heralding a transformative era for Arabic crossword puzzles and the intersection of technology and education.
Authors: Haochen Zhang, Yuyang Dong, Chuan Xiao, Masafumi Oyamada
In this paper, we present Jellyfish, an open-source LLM as a universal task solver for DP. Built on the Llama 2 13B model, Jellyfish is instruction-tuned with the datasets of several typical DP tasks including error detection, data imputation, schema matching, and entity matching, and delivers generalizability to other tasks. Remarkably, Jellyfish can operate on a local, single, and low-priced GPU with its 13 billion parameters, ensuring data security and enabling further tuning. Its proficiency in understanding natural language allows users to manually craft instructions for DP tasks. Unlike many existing methods that heavily rely on prior knowledge, Jellyfish acquires domain knowledge during its tuning process and integrates optional knowledge injection during inference. A distinctive feature of Jellyfish is its interpreter, which elucidates its output decisions. To construct Jellyfish, we develop a series of pre-tuning and DP-tuning techniques. Jellyfish is equipped with an instance serializer, which automatically translates raw data into model prompts, and a knowledge injector, which optionally introduces task- and dataset-specific knowledge to enhance DP performance. Our evaluation of Jellyfish, using a range of real datasets, shows its competitiveness compared to state-of-the-art methods and its strong generalizability to unseen tasks. Jellyfish's performance rivals that of GPT series models, and its interpreter offers enhanced reasoning capabilities compared to GPT-3.5. Furthermore, our evaluation highlights the effectiveness of the techniques employed in constructing Jellyfish. Our model is available at Hugging Face: https://huggingface.co/NECOUDBFM/Jellyfish .
Authors: Zige Wang, Wanjun Zhong, Yufei Wang, Qi Zhu, Fei Mi, Baojun Wang, Lifeng Shang, Xin Jiang, Qun Liu
Data plays a fundamental role in the training of Large Language Models (LLMs). Effective data management, particularly in the formulation of a well-suited training dataset, holds significance for enhancing model performance and improving training efficiency during pretraining and supervised fine-tuning phases. Despite the considerable importance of data management, the current research community still falls short in providing a systematic analysis of the rationale behind management strategy selection, its consequential effects, methodologies for evaluating curated datasets, and the ongoing pursuit of improved strategies. Consequently, the exploration of data management has attracted more and more attention among the research community. This survey provides a comprehensive overview of current research in data management within both the pretraining and supervised fine-tuning stages of LLMs, covering various noteworthy aspects of data management strategy design: data quantity, data quality, domain/task composition, etc. Looking toward the future, we extrapolate existing challenges and outline promising directions for development in this field. Therefore, this survey serves as a guiding resource for practitioners aspiring to construct powerful LLMs through effective data management practices. The collection of the latest papers is available at https://github.com/ZigeW/data_management_LLM.
Authors: Ruijie Wang, Luca Rossetto, Michael Cochez, Abraham Bernstein
Most current methods for multi-hop question answering (QA) over knowledge graphs (KGs) only provide final conclusive answers without explanations, such as a set of KG entities that is difficult for normal users to review and comprehend. This issue severely limits the application of KG-based QA in real-world scenarios. However, it is non-trivial to solve due to two challenges: First, annotations of reasoning chains of multi-hop questions, which could serve as supervision for explanation generation, are usually lacking. Second, it is difficult to maintain high efficiency when explicit KG triples need to be retrieved to generate explanations. In this paper, we propose a novel Graph Neural Network-based Two-Step Reasoning model (GNN2R) to solve this issue. GNN2R can provide both final answers and reasoning subgraphs as a rationale behind final answers efficiently with only weak supervision that is available through question-final answer pairs. We extensively evaluated GNN2R with detailed analyses in experiments. The results demonstrate that, in terms of effectiveness, efficiency, and quality of generated explanations, GNN2R outperforms existing state-of-the-art methods that are applicable to this task. Our code and pre-trained models are available at https://github.com/ruijie-wang-uzh/GNN2R.
Authors: Ehsan Latif, Xiaoming Zhai
This study explores the efficacy of a multi-perspective hybrid neural network (HNN) for scoring student responses in science education with an analytic rubric. We compared the accuracy of the HNN model with four ML approaches (BERT, AACR, Naive Bayes, and Logistic Regression). The results have shown that HHN achieved 8%, 3%, 1%, and 0.12% higher accuracy than Naive Bayes, Logistic Regression, AACR, and BERT, respectively, for five scoring aspects (p<0.001). The overall HNN's perceived accuracy (M = 96.23%, SD = 1.45%) is comparable to the (training and inference) expensive BERT model's accuracy (M = 96.12%, SD = 1.52%). We also have observed that HNN is x2 more efficient in training and inferencing than BERT and has comparable efficiency to the lightweight but less accurate Naive Bayes model. Our study confirmed the accuracy and efficiency of using HNN to score students' science writing automatically.
Authors: Zhongwei Wan, Xin Wang, Che Liu, Samiul Alam, Yu Zheng, Jiachen Liu, Zhongnan Qu, Shen Yan, Yi Zhu, Quanlu Zhang, Mosharaf Chowdhury, Mi Zhang
Large Language Models (LLMs) have demonstrated remarkable capabilities in important tasks such as natural language understanding, language generation, and complex reasoning and have the potential to make a substantial impact on our society. Such capabilities, however, come with the considerable resources they demand, highlighting the strong need to develop effective techniques for addressing their efficiency challenges. In this survey, we provide a systematic and comprehensive review of efficient LLMs research. We organize the literature in a taxonomy consisting of three main categories, covering distinct yet interconnected efficient LLMs topics from model-centric, data-centric, and framework-centric perspective, respectively. We have also created a GitHub repository where we compile the papers featured in this survey at https://github.com/AIoT-MLSys-Lab/EfficientLLMs, and will actively maintain this repository and incorporate new research as it emerges. We hope our survey can serve as a valuable resource to help researchers and practitioners gain a systematic understanding of the research developments in efficient LLMs and inspire them to contribute to this important and exciting field.
Authors: Songlin Yang, Bailin Wang, Yikang Shen, Rameswar Panda, Yoon Kim
Transformers with linear attention allow for efficient parallel training but can simultaneously be formulated as an RNN with 2D (matrix-valued) hidden states, thus enjoying linear (with respect to output length) inference complexity. Recent works such as RetNet (Sun et al., 2023) and TransNormerLLM (Qin et al., 2023a) observe that adding a global decay term to the additive RNN update rule greatly improves performance, sometimes outperforming standard Transformers with softmax attention when trained at scale. In this work we show that adding a data-dependent gating mechanism further improves performance. We derive a parallel form of this gated linear attention layer that enables efficient training. However, a straightforward, numerically stable implementation of this parallel form requires generalized matrix multiplications in log-space for numerical stability, and thus cannot take advantage of tensor cores on modern GPUs which are optimized for standard matrix multiplications. We develop a hardware-efficient version of the parallel form that can still make use of tensor cores through block-parallel computations over sequence chunks. Experiments on moderate-scale language modeling (340M-parameter models trained on 15B tokens, 1.3B-parameter models trained on 100B tokens) show that gated linear attention (GLA) Transformers perform competitively against a strong LLaMA-architecture Transformer baseline (Touvron et al., 2023) as well as Mamba (Gu & Dao, 2023), a recently introduced state-space model with a data-dependent state transition mechanism. For training speed, our Triton-based implementation performs comparably to CUDA-optimized FlashAttention-2 (Dao, 2023) under the regular 2048 training length setting, while outperforming FlashAttention-2 when training on longer sequences beyond 4096.
Authors: Jingxuan Wei, Linzhuang Sun, Xu Tan, Bihui Yu, Ruifeng Guo
Knowledge distillation, a technique for model compression and performance enhancement, has gained significant traction in Neural Machine Translation (NMT). However, existing research primarily focuses on empirical applications, and there is a lack of comprehensive understanding of how student model capacity, data complexity, and decoding strategies collectively influence distillation effectiveness. Addressing this gap, our study conducts an in-depth investigation into these factors, particularly focusing on their interplay in word-level and sequence-level distillation within NMT. Through extensive experimentation across datasets like IWSLT13 En$\rightarrow$Fr, IWSLT14 En$\rightarrow$De, and others, we empirically validate hypotheses related to the impact of these factors on knowledge distillation. Our research not only elucidates the significant influence of model capacity, data complexity, and decoding strategies on distillation effectiveness but also introduces a novel, optimized distillation approach. This approach, when applied to the IWSLT14 de$\rightarrow$en translation task, achieves state-of-the-art performance, demonstrating its practical efficacy in advancing the field of NMT.
Authors: Yu Ji, Wen Wu, Yi Hu, Hong Zheng, Liang He
Few-shot prompting elicits the remarkable abilities of large language models by equipping them with a few demonstration examples in the input. However, the traditional method of providing large language models with all demonstration input-output pairs at once may not effectively guide large language models to learn the specific input-output mapping relationship. In this paper, inspired by the regulatory and supportive role of metacognition in students' learning, we propose a novel metacognition-enhanced few-shot prompting, which guides large language models to reflect on their thought processes to comprehensively learn the given demonstration examples. Furthermore, considering that positive reinforcement can improve students' learning motivation, we introduce positive reinforcement into our metacognition-enhanced few-shot prompting to promote the few-shot learning of large language models by providing response-based positive feedback. The experimental results on two real-world datasets show that our metacognition-enhanced few-shot prompting with positive reinforcement surpasses traditional few-shot prompting in classification accuracy and macro F1.
Authors: Xijie Huang, Li Lyna Zhang, Kwang-Ting Cheng, Mao Yang
Large language models (LLMs) have shown impressive capabilities in various tasks, yet they still struggle with math reasoning. Despite efforts to optimize Chain-of-Thoughts (CoT) prompts and fine-tune LLMs, the potential of few-shot learning remains unexplored. In this work, we propose CoT-Influx, a novel approach pushing the boundaries of few-shot CoT learning to improve LLM math reasoning capabilities. CoT-Influx addresses the challenges of the selection of useful examples and limited number of examples due to restricted context window length. Inspired by our observation that natural language inputs contain many redundancy, we propose a coarse-to-fine pruner as a plug-and-play module for LLMs, which first identifies as many crucial CoT examples as possible and then further prunes unimportant tokens within the context window. To train the pruner, we collect a math reasoning dataset with diverse difficulty and steps, introduce a reward to measure both the input's effectiveness for math reasoning and token length constraints, and propose a novel training approach with reinforcement learning. As a result, CoT-Influx significantly outperforms CoT and few-shot prompting baselines across various LLMs (LLaMA2-7B, 13B, 70B) and 5 mathematical datasets, achieving up to 4.55% absolute improvements. Remarkably, without any fine-tuning, LLaMA2-70B with CoT-Influx surpasses GPT-3.5 and a wide range of larger LLMs (PaLM, Minerva, etc.) on the GSM8K.
Authors: Min-Han Shih, Ho-Lam Chung, Yu-Chi Pai, Ming-Hao Hsu, Guan-Ting Lin, Shang-Wen Li, Hung-yi Lee
In recent advancements in spoken question answering (QA), end-to-end models have made significant strides. However, previous research has primarily focused on extractive span selection. While this extractive-based approach is effective when answers are present directly within the input, it falls short in addressing abstractive questions, where answers are not directly extracted but inferred from the given information. To bridge this gap, we introduce the first end-to-end Generative Spoken Question Answering (GSQA) model that empowers the system to engage in abstractive reasoning. The challenge in training our GSQA model lies in the absence of a spoken abstractive QA dataset. We propose using text models for initialization and leveraging the extractive QA dataset to transfer knowledge from the text generative model to the spoken generative model. Experimental results indicate that our model surpasses the previous extractive model by 3% on extractive QA datasets. Furthermore, the GSQA model has only been fine-tuned on the spoken extractive QA dataset. Despite not having seen any spoken abstractive QA data, it can still closely match the performance of the cascade model. In conclusion, our GSQA model shows the potential to generalize to a broad spectrum of questions, thus further expanding the spoken question answering capabilities of abstractive QA. Our code is available at https://voidful.github.io/GSQA
Authors: Yingzhe Peng, Xu Yang, Haoxuan Ma, Shuo Xu, Chi Zhang, Yucheng Han, Hanwang Zhang
This paper studies how to configure powerful In-Context Demonstration (ICD) sequences for a Large Vision-Language Model (LVLM) to solve Vision-Language tasks through In-Context Learning (ICL). After observing that configuring an ICD sequence is a mirror process of composing a sentence, i.e., just as a sentence can be composed word by word via a Language Model, an ICD sequence can also be configured one by one. Consequently, we introduce an ICD Language Model (ICD-LM) specifically designed to generate effective ICD sequences. This involves creating a dataset of hand-crafted ICD sequences for various query samples and using it to train the ICD-LM. Our approach, diverging from traditional methods in NLP that select and order ICDs separately, enables to simultaneously learn how to select and order ICDs, enhancing the effect of the sequences. Moreover, during data construction, we use the LVLM intended for ICL implementation to validate the strength of each ICD sequence, resulting in a model-specific dataset and the ICD-LM trained by this dataset is also model-specific. We validate our methodology through experiments in Visual Question Answering and Image Captioning, confirming the viability of using a Language Model for ICD configuration. Our comprehensive ablation studies further explore the impact of various dataset construction and ICD-LM development settings on the outcomes. The code is given in https://github.com/ForJadeForest/ICD-LM.
Authors: Congchi Yin, Qian Yu, Zhiwei Fang, Jie He, Changping Peng, Zhangang Lin, Jingping Shao, Piji Li
Decoding non-invasive cognitive signals to natural language has long been the goal of building practical brain-computer interfaces (BCIs). Recent major milestones have successfully decoded cognitive signals like functional Magnetic Resonance Imaging (fMRI) and electroencephalogram (EEG) into text under open vocabulary setting. However, how to split the datasets for training, validating, and testing in cognitive signal decoding task still remains controversial. In this paper, we conduct systematic analysis on current dataset splitting methods and find the existence of data contamination largely exaggerates model performance. Specifically, first we find the leakage of test subjects' cognitive signals corrupts the training of a robust encoder. Second, we prove the leakage of text stimuli causes the auto-regressive decoder to memorize information in test set. The decoder generates highly accurate text not because it truly understands cognitive signals. To eliminate the influence of data contamination and fairly evaluate different models' generalization ability, we propose a new splitting method for different types of cognitive datasets (e.g. fMRI, EEG). We also test the performance of SOTA Brain-to-Text decoding models under the proposed dataset splitting paradigm as baselines for further research.
Authors: Kun Peng, Lei Jiang, Hao Peng, Rui Liu, Zhengtao Yu, Jiaqian Ren, Zhifeng Hao, Philip S.Yu
Aspect Sentiment Triplet Extraction (ASTE) is an emerging task to extract a given sentence's triplets, which consist of aspects, opinions, and sentiments. Recent studies tend to address this task with a table-filling paradigm, wherein word relations are encoded in a two-dimensional table, and the process involves clarifying all the individual cells to extract triples. However, these studies ignore the deep interaction between neighbor cells, which we find quite helpful for accurate extraction. To this end, we propose a novel model for the ASTE task, called Prompt-based Tri-Channel Graph Convolution Neural Network (PT-GCN), which converts the relation table into a graph to explore more comprehensive relational information. Specifically, we treat the original table cells as nodes and utilize a prompt attention score computation module to determine the edges' weights. This enables us to construct a target-aware grid-like graph to enhance the overall extraction process. After that, a triple-channel convolution module is conducted to extract precise sentiment knowledge. Extensive experiments on the benchmark datasets show that our model achieves state-of-the-art performance. The code is available at https://github.com/KunPunCN/PT-GCN.
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: Bing Wang, Changyu Ren, Jian Yang, Xinnian Liang, Jiaqi Bai, Qian-Wen Zhang, Zhao Yan, Zhoujun Li
Recent advancements in Text-to-SQL methods employing Large Language Models (LLMs) have demonstrated remarkable performance. Nonetheless, these approaches continue to encounter difficulties when handling extensive databases, intricate user queries, and erroneous SQL results. To tackle these challenges, we present \textsc{MAC-SQL}, a novel LLM-based multi-agent collaborative framework designed for the Text-to-SQL task. Our framework comprises three agents: the \textit{Selector}, accountable for condensing voluminous databases and preserving relevant table schemas for user questions; the \textit{Decomposer}, which disassembles complex user questions into more straightforward sub-problems and resolves them progressively; and the \textit{Refiner}, tasked with validating and refining defective SQL queries. We perform comprehensive experiments on two Text-to-SQL datasets, BIRD and Spider, achieving a state-of-the-art execution accuracy of 59.59\% on the BIRD test set. Moreover, we have open-sourced an instruction fine-tuning model, SQL-Llama, based on Code Llama 7B, in addition to an agent instruction dataset derived from training data based on BIRD and Spider. The SQL-Llama model has demonstrated encouraging results on the development sets of both BIRD and Spider. However, when compared to GPT-4, there remains a notable potential for enhancement. Our code and data are publicly available at https://github.com/wbbeyourself/MAC-SQL.
Authors: Syeda Nahida Akter, Zichun Yu, Aashiq Muhamed, Tianyue Ou, Alex Bäuerle, Ángel Alexander Cabrera, Krish Dholakia, Chenyan Xiong, Graham Neubig
The recently released Google Gemini class of models are the first to comprehensively report results that rival the OpenAI GPT series across a wide variety of tasks. In this paper, we do an in-depth exploration of Gemini's language abilities, making two contributions. First, we provide a third-party, objective comparison of the abilities of the OpenAI GPT and Google Gemini models with reproducible code and fully transparent results. Second, we take a closer look at the results, identifying areas where one of the two model classes excels. We perform this analysis over 10 datasets testing a variety of language abilities, including reasoning, answering knowledge-based questions, solving math problems, translating between languages, generating code, and acting as instruction-following agents. From this analysis, we find that Gemini Pro achieves accuracy that is close but slightly inferior to the corresponding GPT 3.5 Turbo on all tasks that we benchmarked. We further provide explanations for some of this under-performance, including failures in mathematical reasoning with many digits, sensitivity to multiple-choice answer ordering, aggressive content filtering, and others. We also identify areas where Gemini demonstrates comparably high performance, including generation into non-English languages, and handling longer and more complex reasoning chains. Code and data for reproduction can be found at https://github.com/neulab/gemini-benchmark
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: 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: Da Luo, 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: Michael Dalvean
In English place name analysis, meanings are often derived from the resemblance of roots in place names to topographical features, proper names and/or habitation terms in one of the languages that have had an influence on English place names. The problem here is that it is sometimes difficult to determine the base language to use to interpret the roots. The purpose of this paper is to stochastically determine the resemblance between 18799 English place names and 84685 place names from Ireland, Scotland, Wales, Denmark, Norway, Sweden, France, Germany, the Netherlands and Ancient Rome. Each English place name is ranked according to the extent to which it resembles place names from the other countries, and this provides a basis for determining the likely language to use to interpret the place name. A number of observations can be made using the ranking provided. In particular, it is found that `Didlington' is the most archetypically English place name in the English sample, and `Anna' is the least. Furthermore, it is found that the place names in the non-English datasets are most similar to Norwegian place names and least similar to Welsh place names.
Authors: Viktor Schlegel, Abhinav Ramesh Kashyap, Thanh-Tung Nguyen, Tsung-Han Yang, Vijay Prakash Dwivedi, Wei-Hsian Yin, Jeng Wei, Stefan Winkler
Computerised clinical coding approaches aim to automate the process of assigning a set of codes to medical records. While there is active research pushing the state of the art on clinical coding for hospitalized patients, the outpatient setting -- where doctors tend to non-hospitalised patients -- is overlooked. Although both settings can be formalised as a multi-label classification task, they present unique and distinct challenges, which raises the question of whether the success of inpatient clinical coding approaches translates to the outpatient setting. This paper is the first to investigate how well state-of-the-art deep learning-based clinical coding approaches work in the outpatient setting at hospital scale. To this end, we collect a large outpatient dataset comprising over 7 million notes documenting over half a million patients. We adapt four state-of-the-art clinical coding approaches to this setting and evaluate their potential to assist coders. We find evidence that clinical coding in outpatient settings can benefit from more innovations in popular inpatient coding benchmarks. A deeper analysis of the factors contributing to the success -- amount and form of data and choice of document representation -- reveals the presence of easy-to-solve examples, the coding of which can be completely automated with a low error rate.
Authors: Chengzu Li, Han Zhou, Goran Glavaš, Anna Korhonen, Ivan Vulić
Following the standard supervised fine-tuning (SFT) paradigm, in-context learning (ICL) has become an efficient approach propelled by the recent advancements in large language models (LLMs), yielding promising performance across various tasks in few-shot data setups. However, both paradigms are prone to suffer from the critical problem of overconfidence (i.e., miscalibration), especially in such limited data setups. In this work, we deliver an in-depth analysis of the behavior across different choices of learning methods from the perspective of both performance and calibration, as well as their interplay. Through extensive controlled experiments, we find that simultaneous gains for both task performance and calibration are difficult to achieve, and the problem of miscalibration exists across all learning methods in low-resource scenarios. To address this challenging trade-off between performance and calibration, we then investigate the potential of self-ensembling techniques applied at different modeling stages (e.g., variations of in-context examples or variations in prompts or different ensembling strategies). We justify the feasibility of self-ensembling on SFT in addition to ICL, to make the predictions more calibrated and have comparable or even better performance. Our work sheds light on which learning paradigm to choose and how to enhance both task performance and calibration of LLMs.
Authors: Dou Hu, Lingwei Wei, Yaxin Liu, Wei Zhou, Songlin Hu
This paper presents a new supervised representation learning framework, namely structured probabilistic coding (SPC), to learn compact and informative representations from input related to the target task. SPC is an encoder-only probabilistic coding technology with a structured regularization from the target label space. It can enhance the generalization ability of pre-trained language models for better language understanding. Specifically, our probabilistic coding technology simultaneously performs information encoding and task prediction in one module to more fully utilize the effective information from input data. It uses variational inference in the output space to reduce randomness and uncertainty. Besides, to better control the probability distribution in the latent space, a structured regularization is proposed to promote class-level uniformity in the latent space. With the regularization term, SPC can preserve the Gaussian distribution structure of latent code as well as better cover the hidden space with class uniformly. Experimental results on 12 natural language understanding tasks demonstrate that our SPC effectively improves the performance of pre-trained language models for classification and regression. Extensive experiments show that SPC can enhance the generalization capability, robustness to label noise, and clustering quality of output representations.
Authors: Zhaojian Yu, Xin Zhang, Ning Shang, Yangyu Huang, Can Xu, Yishujie Zhao, Wenxiang Hu, Qiufeng Yin
Recent work demonstrates that, after being fine-tuned on a high-quality instruction dataset, the resulting model can obtain impressive capabilities to address a wide range of tasks. However, existing methods for instruction data generation often produce duplicate data and are not controllable enough on data quality. In this paper, we extend the generalization of instruction tuning by classifying the instruction data to 4 code-related tasks and propose a LLM-based Generator-Discriminator data process framework to generate diverse, high-quality instruction data from open source code. Hence, we introduce CodeOcean, a dataset comprising 20,000 instruction instances across 4 universal code-related tasks,which is aimed at augmenting the effectiveness of instruction tuning and improving the generalization ability of fine-tuned model. Subsequently, we present WaveCoder, a fine-tuned Code LLM with Widespread And Versatile Enhanced instruction tuning. This model is specifically designed for enhancing instruction tuning of Code Language Models (LLMs). Our experiments demonstrate that Wavecoder models outperform other open-source models in terms of generalization ability across different code-related tasks at the same level of fine-tuning scale. Moreover, Wavecoder exhibits high efficiency in previous code generation tasks. This paper thus offers a significant contribution to the field of instruction data generation and fine-tuning models, providing new insights and tools for enhancing performance in code-related tasks.
Authors: Abiodun Finbarrs Oketunji, Muhammad Anas, Deepthi Saina
The Large Language Model Bias Index (LLMBI) is a pioneering approach designed to quantify and address biases inherent in large language models (LLMs), such as GPT-4. We recognise the increasing prevalence and impact of LLMs across diverse sectors. This research introduces a novel metric, LLMBI, to systematically measure and mitigate biases potentially skewing model responses. We formulated LLMBI using a composite scoring system incorporating multiple dimensions of bias, including but not limited to age, gender, and racial biases.
To operationalise this metric, we engaged in a multi-step process involving collecting and annotating LLM responses, applying sophisticated Natural Language Processing (NLP) techniques for bias detection, and computing the LLMBI score through a specially crafted mathematical formula. The formula integrates weighted averages of various bias dimensions, a penalty for dataset diversity deficiencies, and a correction for sentiment biases. Our empirical analysis, conducted using responses from OpenAI's API, employs advanced sentiment analysis as a representative method for bias detection.
The research reveals LLMs, whilst demonstrating impressive capabilities in text generation, exhibit varying degrees of bias across different dimensions. LLMBI provides a quantifiable measure to compare biases across models and over time, offering a vital tool for systems engineers, researchers and regulators in enhancing the fairness and reliability of LLMs. It highlights the potential of LLMs in mimicking unbiased human-like responses. Additionally, it underscores the necessity of continuously monitoring and recalibrating such models to align with evolving societal norms and ethical standards.
Authors: Lizhou Fan, Wenyue Hua, Lingyao Li, Haoyang Ling, Yongfeng Zhang, Libby Hemphill
Complex reasoning ability is one of the most important features of current LLMs, which has also been leveraged to play an integral role in complex decision-making tasks. Therefore, the investigation into the reasoning capabilities of Large Language Models (LLMs) is critical: numerous benchmarks have been established to assess the reasoning abilities of LLMs. However, current benchmarks are inadequate in offering a rigorous evaluation of the full extent of reasoning abilities that LLMs are capable of achieving. They are also prone to the risk of overfitting, as these benchmarks, being publicly accessible and static, allow models to potentially tailor their responses to specific benchmark metrics, thereby inflating their performance. Addressing these limitations, our research introduces a new benchmark, named NPHardEval. This benchmark is designed to evaluate the reasoning abilities of LLMs across a broad spectrum of 900 algorithmic questions, extending up to the NP-Hard complexity class. These questions are meticulously chosen to represent a wide range of complexity class below the NP-hard complexity class, offering a rigorous measure of the reasoning ability of LLMs. Through this study, we shed light on the current state of reasoning in LLMs, providing an objective and rigorous perspective through the comparison of LLMs' performance across complex classes. Moreover, this benchmark is designed with a dynamic update mechanism, where the datapoints are refreshed on a monthly basis. Such regular updates play a crucial role in mitigating the risk of LLMs overfitting to the benchmark, promoting a more accurate and reliable assessment of their reasoning capabilities. The benchmark dataset and code of NPHardEval are available at https://github.com/casmlab/NPHardEval.
Authors: Timothy R. McIntosh, Teo Susnjak, Tong Liu, Paul Watters, Malka N. Halgamuge
This comprehensive survey explored the evolving landscape of generative Artificial Intelligence (AI), with a specific focus on the transformative impacts of Mixture of Experts (MoE), multimodal learning, and the speculated advancements towards Artificial General Intelligence (AGI). It critically examined the current state and future trajectory of generative Artificial Intelligence (AI), exploring how innovations like Google's Gemini and the anticipated OpenAI Q* project are reshaping research priorities and applications across various domains, including an impact analysis on the generative AI research taxonomy. It assessed the computational challenges, scalability, and real-world implications of these technologies while highlighting their potential in driving significant progress in fields like healthcare, finance, and education. It also addressed the emerging academic challenges posed by the proliferation of both AI-themed and AI-generated preprints, examining their impact on the peer-review process and scholarly communication. The study highlighted the importance of incorporating ethical and human-centric methods in AI development, ensuring alignment with societal norms and welfare, and outlined a strategy for future AI research that focuses on a balanced and conscientious use of MoE, multimodality, and AGI in generative AI.