Authors: Emiel van Miltenburg
Abstract: This short position paper provides a manually curated list of non-English image captioning datasets (as of May 2024). Through this list, we can observe the dearth of datasets in different languages: only 23 different languages are represented. With the addition of the Crossmodal-3600 dataset (Thapliyal et al., 2022, 36 languages) this number increases somewhat, but still this number is tiny compared to the thousands of spoken languages that exist. This paper closes with some open questions for the field of Vision & Language.
Authors: Parag Jain, Mirella Lapata
Abstract: We focus on a conversational question answering task which combines the challenges of understanding questions in context and reasoning over evidence gathered from heterogeneous sources like text, knowledge graphs, tables, and infoboxes. Our method utilizes a graph structured representation to aggregate information about a question and its context (i.e., the conversation so far and evidence retrieved to find an answer), while also harnessing the reasoning and text generation capabilities of large language models (LLMs). Graph embeddings are directly injected into the LLM, bypassing the token embedding layers, and learned end-to-end by minimizing cross-entropy. Our model maintains a memory module to track and update past evidence, thus influencing the graph's structure, as the conversation evolves. Experimental results on the ConvMix benchmark(Christmann et al., 2022a) show that graph embeddings enhance the LLM's ability to reason, while the memory module provides robustness against noise and retrieval errors.
Authors: Young Kyun Jang, Junmo Kang, Yong Jae Lee, Donghyun Kim
Abstract: While advancements in Vision Language Models (VLMs) have significantly improved the alignment of visual and textual data, these models primarily focus on aligning images with short descriptive captions. This focus limits their ability to handle complex text interactions, particularly with longer texts such as lengthy captions or documents, which have not been extensively explored yet. In this paper, we introduce Meet At The Embedding (MATE), a novel approach that combines the capabilities of VLMs with Large Language Models (LLMs) to overcome this challenge without the need for additional image-long text pairs. Specifically, we replace the text encoder of the VLM with a pretrained LLM-based encoder that excels in understanding long texts. To bridge the gap between VLM and LLM, MATE incorporates a projection module that is trained in a multi-stage manner. It starts by aligning the embeddings from the VLM text encoder with those from the LLM using extensive text pairs. This module is then employed to seamlessly align image embeddings closely with LLM embeddings. We propose two new cross-modal retrieval benchmarks to assess the task of connecting images with long texts (lengthy captions / documents). Extensive experimental results demonstrate that MATE effectively connects images with long texts, uncovering diverse semantic relationships.
Authors: Ali Ghadami, Alireza Taheri, Ali Meghdari
Abstract: Sign language is an essential means of communication for millions of people around the world and serves as their primary language. However, most communication tools are developed for spoken and written languages which can cause problems and difficulties for the deaf and hard of hearing community. By developing a sign language recognition system, we can bridge this communication gap and enable people who use sign language as their main form of expression to better communicate with people and their surroundings. This recognition system increases the quality of health services, improves public services, and creates equal opportunities for the deaf community. This research aims to recognize Iranian Sign Language words with the help of the latest deep learning tools such as transformers. The dataset used includes 101 Iranian Sign Language words frequently used in academic environments such as universities. The network used is a combination of early fusion and late fusion transformer encoder-based networks optimized with the help of genetic algorithm. The selected features to train this network include hands and lips key points, and the distance and angle between hands extracted from the sign videos. Also, in addition to the training model for the classes, the embedding vectors of words are used as multi-task learning to have smoother and more efficient training. This model was also tested on sentences generated from our word dataset using a windowing technique for sentence translation. Finally, the sign language training software that provides real-time feedback to users with the help of the developed model, which has 90.2% accuracy on test data, was introduced, and in a survey, the effectiveness and efficiency of this type of sign language learning software and the impact of feedback were investigated.
Authors: Zeliang Zhang, Xiaodong Liu, Hao Cheng, Chenliang Xu, Jianfeng Gao
Abstract: By increasing model parameters but activating them sparsely when performing a task, the use of Mixture-of-Experts (MoE) architecture significantly improves the performance of Large Language Models (LLMs) without increasing the inference cost. However, the memory consumption due to the growing number of experts presents a challenge to the deployment of these models in many real world settings. Our empirical study reveals that some experts encode redundant knowledge during pre-training. We thus propose a method of grouping and pruning similar experts to improve model's parameter efficiency. We validate the effectiveness of our method by pruning two state-of-the-art MoE models, Mixtral-8x7B and Mixtral-8x22B. Evaluation shows that our method outperforms other model pruning methods on a range of natural language tasks. To facilitate future research, we will release our code and the pruned MoE models.
Authors: Andrea W Wen-Yi, Unso Eun Seo Jo, Lu Jia Lin, David Mimno
Abstract: Contemporary language models are increasingly multilingual, but Chinese LLM developers must navigate complex political and business considerations of language diversity. Language policy in China aims at influencing the public discourse and governing a multi-ethnic society, and has gradually transitioned from a pluralist to a more assimilationist approach since 1949. We explore the impact of these influences on current language technology. We evaluate six open-source multilingual LLMs pre-trained by Chinese companies on 18 languages, spanning a wide range of Chinese, Asian, and Anglo-European languages. Our experiments show Chinese LLMs performance on diverse languages is indistinguishable from international LLMs. Similarly, the models' technical reports also show lack of consideration for pretraining data language coverage except for English and Mandarin Chinese. Examining Chinese AI policy, model experiments, and technical reports, we find no sign of any consistent policy, either for or against, language diversity in China's LLM development. This leaves a puzzling fact that while China regulates both the languages people use daily as well as language model development, they do not seem to have any policy on the languages in language models.
Authors: Saar Kuzi, Shervin Malmasi
Abstract: Consumers on a shopping mission often leverage both product search and information seeking systems, such as web search engines and Question Answering (QA) systems, in an iterative process to improve their understanding of available products and reach a purchase decision. While product search is useful for shoppers to find the actual products meeting their requirements in the catalog, information seeking systems can be utilized to answer any questions they may have to refine those requirements. The recent success of Large Language Models (LLMs) has opened up an opportunity to bridge the gap between the two tasks to help customers achieve their goals quickly and effectively by integrating conversational QA within product search. In this paper, we propose to recommend users Question-Answer (Q&A) pairs that are relevant to their product search and can help them make a purchase decision. We discuss the different aspects of the problem including the requirements and characteristics of the Q&A pairs, their generation, and the optimization of the Q&A recommendation task. We highlight the challenges, open problems, and suggested solutions to encourage future research in this emerging area.
Authors: Chase Fensore, Rodrigo M. Carrillo-Larco, Shivani A. Patel, Alanna A. Morris, Joyce C. Ho
Abstract: Social determinants of health (SDOH) $-$ the myriad of circumstances in which people live, grow, and age $-$ play an important role in health outcomes. However, existing outcome prediction models often only use proxies of SDOH as features. Recent open data initiatives present an opportunity to construct a more comprehensive view of SDOH, but manually integrating the most relevant data for individual patients becomes increasingly challenging as the volume and diversity of public SDOH data grows. Large language models (LLMs) have shown promise at automatically annotating structured data. Here, we conduct an end-to-end case study evaluating the feasibility of using LLMs to integrate SDOH data, and the utility of these SDOH features for clinical prediction. We first manually label 700+ variables from two publicly-accessible SDOH data sources to one of five semantic SDOH categories. Then, we benchmark performance of 9 open-source LLMs on this classification task. Finally, we train ML models to predict 30-day hospital readmission among 39k heart failure (HF) patients, and we compare the prediction performance of the categorized SDOH variables with standard clinical variables. Additionally, we investigate the impact of few-shot LLM prompting on LLM annotation performance, and perform a metadata ablation study on prompts to evaluate which information helps LLMs accurately annotate these variables. We find that some open-source LLMs can effectively, accurately annotate SDOH variables with zero-shot prompting without the need for fine-tuning. Crucially, when combined with standard clinical features, the LLM-annotated Neighborhood and Built Environment subset of the SDOH variables shows the best performance predicting 30-day readmission of HF patients.
Authors: Viktor Mihaylov, Aleksandar Shtedritski
Abstract: This paper investigates biases of Large Language Models (LLMs) through the lens of grammatical gender. Drawing inspiration from seminal works in psycholinguistics, particularly the study of gender's influence on language perception, we leverage multilingual LLMs to revisit and expand upon the foundational experiments of Boroditsky (2003). Employing LLMs as a novel method for examining psycholinguistic biases related to grammatical gender, we prompt a model to describe nouns with adjectives in various languages, focusing specifically on languages with grammatical gender. In particular, we look at adjective co-occurrences across gender and languages, and train a binary classifier to predict grammatical gender given adjectives an LLM uses to describe a noun. Surprisingly, we find that a simple classifier can not only predict noun gender above chance but also exhibit cross-language transferability. We show that while LLMs may describe words differently in different languages, they are biased similarly.
Authors: Zongyue Qin, Ziniu Hu, Zifan He, Neha Prakriya, Jason Cong, Yizhou Sun
Abstract: Transformer-based Large language models (LLMs) have demonstrated their power in various tasks, but their inference incurs significant time and energy costs. To accelerate LLM inference, speculative decoding uses a smaller model to propose one sequence of tokens, which are subsequently validated in batch by the target large model. Compared with autoregressive decoding, speculative decoding generates the same number of tokens with fewer runs of the large model, hence accelerating the overall inference by $1$-$2\times$. However, greedy decoding is not the optimal decoding algorithm in terms of output perplexity, which is a direct measurement of the effectiveness of a decoding algorithm. An algorithm that has better output perplexity and even better efficiency than speculative decoding can be more useful in practice. To achieve this seemingly contradictory goal, we first introduce multi-token joint greedy decoding (MJGD), which greedily generates multiple tokens at each step based on their joint perplexity. We show that it leads to better perplexity for the whole output. But the computation cost of MJGD is infeasible in practice. So we further propose multi-token joint speculative decoding (MJSD), which approximates and accelerates the MJGD from two aspects: it approximates the joint distribution of the large model with that of a small model, and uses a verification step to guarantee the accuracy of approximation; then it uses beam decoding to accelerate the sequence generation from the joint distribution. Compared with vanilla speculative decoding, MJSD has two advantages: (1) it is an approximation of MJGD, thus achieving better output perplexity; (2) verification with joint likelihood allows it to accept the longest prefix sub-sequence of the draft tokens with valid perplexity, leading to better efficiency...
Authors: Gongyao Jiang, Xinran Shi, Qiong Luo
Abstract: Science journalism reports current scientific discoveries to non-specialists, aiming to enable public comprehension of the state of the art. However, this task can be challenging as the audience often lacks specific knowledge about the presented research. To address this challenge, we propose a framework that integrates three LLMs mimicking the real-world writing-reading-feedback-revision workflow, with one LLM acting as the journalist, a smaller LLM as the general public reader, and the third LLM as an editor. The journalist's writing is iteratively refined by feedback from the reader and suggestions from the editor. Our experiments demonstrate that by leveraging the collaboration of two 7B and one 1.8B open-source LLMs, we can generate articles that are more accessible than those generated by existing methods, including advanced models such as GPT-4.
Authors: Zhenpeng Su, Zijia Lin, Xue Bai, Xing Wu, Yizhe Xiong, Haoran Lian, Guangyuan Ma, Hui Chen, Guiguang Ding, Wei Zhou, Songlin Hu
Abstract: Scaling model capacity enhances its capabilities but significantly increases computation. Mixture-of-Experts models (MoEs) address this by allowing model capacity to scale without substantially increasing training or inference costs. Despite their promising results, MoE models encounter several challenges. Primarily, the dispersion of training tokens across multiple experts can lead to underfitting, particularly for infrequent tokens. Additionally, while fixed routing mechanisms can mitigate this issue, they compromise on the diversity of representations. In this paper, we propose MaskMoE, a method designed to enhance token-level learning by employing a routing masking technique within the Mixture-of-Experts model. MaskMoE is capable of maintaining representation diversity while achieving more comprehensive training. Experimental results demonstrate that our method outperforms previous dominant Mixture-of-Experts models in both perplexity (PPL) and downstream tasks.
Authors: Sanad Malaysha, Mo El-Haj, Saad Ezzini, Mohammed Khalilia, Mustafa Jarrar, Sultan Almujaiwel, Ismail Berrada, Houda Bouamor
Abstract: The expanding financial markets of the Arab world require sophisticated Arabic NLP tools. To address this need within the banking domain, the Arabic Financial NLP (AraFinNLP) shared task proposes two subtasks: (i) Multi-dialect Intent Detection and (ii) Cross-dialect Translation and Intent Preservation. This shared task uses the updated ArBanking77 dataset, which includes about 39k parallel queries in MSA and four dialects. Each query is labeled with one or more of a common 77 intents in the banking domain. These resources aim to foster the development of robust financial Arabic NLP, particularly in the areas of machine translation and banking chat-bots. A total of 45 unique teams registered for this shared task, with 11 of them actively participated in the test phase. Specifically, 11 teams participated in Subtask 1, while only 1 team participated in Subtask 2. The winning team of Subtask 1 achieved F1 score of 0.8773, and the only team submitted in Subtask 2 achieved a 1.667 BLEU score.
Authors: Md. Arid Hasan, Maram Hasanain, Fatema Ahmad, Sahinur Rahman Laskar, Sunaya Upadhyay, Vrunda N Sukhadia, Mucahid Kutlu, Shammur Absar Chowdhury, Firoj Alam
Abstract: Natural Question Answering (QA) datasets play a crucial role in developing and evaluating the capabilities of large language models (LLMs), ensuring their effective usage in real-world applications. Despite the numerous QA datasets that have been developed, there is a notable lack of region-specific datasets generated by native users in their own languages. This gap hinders the effective benchmarking of LLMs for regional and cultural specificities. In this study, we propose a scalable framework, NativQA, to seamlessly construct culturally and regionally aligned QA datasets in native languages, for LLM evaluation and tuning. Moreover, to demonstrate the efficacy of the proposed framework, we designed a multilingual natural QA dataset, MultiNativQA, consisting of ~72K QA pairs in seven languages, ranging from high to extremely low resource, based on queries from native speakers covering 18 topics. We benchmark the MultiNativQA dataset with open- and closed-source LLMs. We made both the framework NativQA and MultiNativQA dataset publicly available for the community. (https://nativqa.gitlab.io)
Authors: Xiuying Wei, Skander Moalla, Razvan Pascanu, Caglar Gulcehre
Abstract: State-of-the-art LLMs often rely on scale with high computational costs, which has sparked a research agenda to reduce parameter counts and costs without significantly impacting performance. Our study focuses on Transformer-based LLMs, specifically applying low-rank parametrization to the computationally intensive feedforward networks (FFNs), which are less studied than attention blocks. In contrast to previous works, (i) we explore low-rank parametrization at scale, up to 1.3B parameters; (ii) within Transformer language models rather than convolutional architectures; and (iii) starting from training from scratch. Experiments on the large RefinedWeb dataset show that low-rank parametrization is both efficient (e.g., 2.6$\times$ FFN speed-up with 32\% parameters) and effective during training. Interestingly, these structured FFNs exhibit steeper scaling curves than the original models. Motivated by this finding, we develop the wide and structured networks surpassing the current medium-sized and large-sized Transformer in perplexity and throughput performance.
Authors: Dmitrii Galimzianov, Viacheslav Vyshegorodtsev
Abstract: Average hold time is a concern for call centers because it affects customer satisfaction. Contact centers should instruct their agents to use special on-hold scripts to maintain positive interactions with clients. This study presents a natural language processing model that detects on-hold phrases in customer service calls transcribed by automatic speech recognition technology. The task of finding hold scripts in dialogue was formulated as a multiclass text classification problem with three mutually exclusive classes: scripts for putting a client on hold, scripts for returning to a client, and phrases irrelevant to on-hold scripts. We collected an in-house dataset of calls and labeled each dialogue turn in each call. We fine-tuned RuBERT on the dataset by exploring various hyperparameter sets and achieved high model performance. The developed model can help agent monitoring by providing a way to check whether an agent follows predefined on-hold scripts.
Authors: Shantipriya Parida, Shakshi Panwar, Kusum Lata, Sanskruti Mishra, Sambit Sekhar
Abstract: Large language models (LLMs) demonstrated transformative capabilities in many applications that require automatically generating responses based on human instruction. However, the major challenge for building LLMs, particularly in Indic languages, is the availability of high-quality data for building foundation LLMs. In this paper, we are proposing a large pre-train dataset in Hindi useful for the Indic language Hindi. We have collected the data span across several domains including major dialects in Hindi. The dataset contains 1.28 billion Hindi tokens. We have explained our pipeline including data collection, pre-processing, and availability for LLM pre-training. The proposed approach can be easily extended to other Indic and low-resource languages and will be available freely for LLM pre-training and LLM research purposes.
Authors: Juli Bakagianni, Kanella Pouli, Maria Gavriilidou, John Pavlopoulos
Abstract: Natural Language Processing (NLP) research has traditionally been predominantly focused on English, driven by the availability of resources, the size of the research community, and market demands. Recently, there has been a noticeable shift towards multilingualism in NLP, recognizing the need for inclusivity and effectiveness across diverse languages and cultures. Monolingual surveys have the potential to complement the broader trend towards multilingualism in NLP by providing foundational insights and resources necessary for effectively addressing the linguistic diversity of global communication. However, monolingual NLP surveys are extremely rare in literature. This study fills the gap by introducing a method for creating systematic and comprehensive monolingual NLP surveys. Characterized by a structured search protocol, it can be used to select publications and organize them through a taxonomy of NLP tasks. We include a classification of Language Resources (LRs), according to their availability, and datasets, according to their annotation, to highlight publicly-available and machine-actionable LRs. By applying our method, we conducted a systematic literature review of Greek NLP from 2012 to 2022, providing a comprehensive overview of the current state and challenges of Greek NLP research. We discuss the progress of Greek NLP and outline encountered Greek LRs, classified by availability and usability. As we show, our proposed method helps avoid common pitfalls, such as data leakage and contamination, and to assess language support per NLP task. We consider this systematic literature review of Greek NLP an application of our method that showcases the benefits of a monolingual NLP survey. Similar applications could be regard the myriads of languages whose progress in NLP lags behind that of well-supported languages.
Authors: Sanchit Ahuja, Kumar Tanmay, Hardik Hansrajbhai Chauhan, Barun Patra, Kriti Aggarwal, Tejas Indulal Dhamecha, Monojit Choudhary, Vishrav Chaudhary, Sunayana Sitaram
Abstract: Despite the remarkable success of LLMs in English, there is a significant gap in performance in non-English languages. In order to address this, we introduce a novel recipe for creating a multilingual synthetic instruction tuning dataset, sPhinX, which is created by selectively translating instruction response pairs from English into 50 languages. We test the effectiveness of sPhinX by using it to fine-tune two state-of-the-art models, Phi-3-small and Mistral-7B and then evaluating them across a comprehensive suite of multilingual benchmarks that test reasoning, question answering, and reading comprehension. Our results show that Phi-3-small and Mistral-7B fine-tuned with sPhinX perform better on an average by 4.2%pt and 5%pt respectively as compared to the baselines. We also devise a strategy to incorporate N-shot examples in each fine-tuning sample which further boosts the performance of these models by 3%pt and 10%pt respectively. Additionally, sPhinX also outperforms other multilingual instruction tuning datasets on the same benchmarks along with being sample efficient and diverse, thereby reducing dataset creation costs. Additionally, instruction tuning with sPhinX does not lead to regression on most English benchmarks.
Authors: Dimitris Papadopoulos, Katerina Metropoulou, Nikolaos Matsatsinis, Nikolaos Papadakis
Abstract: Our collective attention span is shortened by the flood of online information. With \textit{FarFetched}, we address the need for automated claim validation based on the aggregated evidence derived from multiple online news sources. We introduce an entity-centric reasoning framework in which latent connections between events, actions, or statements are revealed via entity mentions and represented in a graph database. Using entity linking and semantic similarity, we offer a way for collecting and combining information from diverse sources in order to generate evidence relevant to the user's claim. Then, we leverage textual entailment recognition to quantitatively determine whether this assertion is credible, based on the created evidence. Our approach tries to fill the gap in automated claim validation for less-resourced languages and is showcased on the Greek language, complemented by the training of relevant semantic textual similarity (STS) and natural language inference (NLI) models that are evaluated on translated versions of common benchmarks.
Authors: Shengbin Yue, Siyuan Wang, Wei Chen, Xuanjing Huang, Zhongyu Wei
Abstract: Recent advancements in Large Language Models (LLMs) have led to significant breakthroughs in various natural language processing tasks. However, generating factually consistent responses in knowledge-intensive scenarios remains a challenge due to issues such as hallucination, difficulty in acquiring long-tailed knowledge, and limited memory expansion. This paper introduces SMART, a novel multi-agent framework that leverages external knowledge to enhance the interpretability and factual consistency of LLM-generated responses. SMART comprises four specialized agents, each performing a specific sub-trajectory action to navigate complex knowledge-intensive tasks. We propose a multi-agent co-training paradigm, Long- and Short-Trajectory Learning, which ensures synergistic collaboration among agents while maintaining fine-grained execution by each agent. Extensive experiments on 5 tasks demonstrate SMART's superior performance compared to previous widely adopted methods.
Authors: KuanChao Chu, Yi-Pei Chen, Hideki Nakayama
Abstract: This paper investigates the quality of multi-agent dialogues in simulations powered by Large Language Models (LLMs), focusing on a case study from Park et al. (2023), where 25 agents engage in day-long simulations of life, showcasing complex behaviors and interactions. Analyzing dialogues and memory over multiple sessions revealed significant issues such as repetition, inconsistency, and hallucination, exacerbated by the propagation of erroneous information. To combat these challenges, we propose a novel Screening, Diagnosis, and Regeneration (SDR) framework that detects and corrects utterance errors through a comprehensive process involving immediate issue identification, evidence gathering from past dialogues, and LLM analysis for utterance revision. The effectiveness of the SDR framework is validated through GPT-4 assessments and human evaluations, demonstrating marked improvements in dialogue consistency, diversity, and the reduction of false information. This work presents a pioneering approach to enhancing dialogue quality in multi-agent simulations, establishing a new standard for future research in the field.
Authors: Mustafa Jarrar, Nagham Hamad, Mohammed Khalilia, Bashar Talafha, AbdelRahim Elmadany, Muhammad Abdul-Mageed
Abstract: We present WojoodNER-2024, the second Arabic Named Entity Recognition (NER) Shared Task. In WojoodNER-2024, we focus on fine-grained Arabic NER. We provided participants with a new Arabic fine-grained NER dataset called wojoodfine, annotated with subtypes of entities. WojoodNER-2024 encompassed three subtasks: (i) Closed-Track Flat Fine-Grained NER, (ii) Closed-Track Nested Fine-Grained NER, and (iii) an Open-Track NER for the Israeli War on Gaza. A total of 43 unique teams registered for this shared task. Five teams participated in the Flat Fine-Grained Subtask, among which two teams tackled the Nested Fine-Grained Subtask and one team participated in the Open-Track NER Subtask. The winning teams achieved F-1 scores of 91% and 92% in the Flat Fine-Grained and Nested Fine-Grained Subtasks, respectively. The sole team in the Open-Track Subtask achieved an F-1 score of 73.7%.
Authors: Haode Zhang, Xiao-Ming Wu, Albert Y. S. Lam
Abstract: Recent research has demonstrated the feasibility of training efficient intent detectors based on pre-trained language model~(PLM) with limited labeled data. However, deploying these detectors in resource-constrained environments such as mobile devices poses challenges due to their large sizes. In this work, we aim to address this issue by exploring techniques to minimize the size of PLM-based intent detectors trained with few-shot data. Specifically, we utilize large language models (LLMs) for data augmentation, employ a cutting-edge model compression method for knowledge distillation, and devise a vocabulary pruning mechanism called V-Prune. Through these approaches, we successfully achieve a compression ratio of 21 in model memory usage, including both Transformer and the vocabulary, while maintaining almost identical performance levels on four real-world benchmarks.
Authors: Seethalakshmi Gopalakrishnan, Luciana Garbayo, Wlodek Zadrozny
Abstract: This study explores the potential of natural language models, including large language models, to extract causal relations from medical texts, specifically from Clinical Practice Guidelines (CPGs). The outcomes causality extraction from Clinical Practice Guidelines for gestational diabetes are presented, marking a first in the field. We report on a set of experiments using variants of BERT (BioBERT, DistilBERT, and BERT) and using Large Language Models (LLMs), namely GPT-4 and LLAMA2. Our experiments show that BioBERT performed better than other models, including the Large Language Models, with an average F1-score of 0.72. GPT-4 and LLAMA2 results show similar performance but less consistency. We also release the code and an annotated a corpus of causal statements within the Clinical Practice Guidelines for gestational diabetes.
Authors: Kriti Bhattarai, Inez Y. Oh, Zachary B. Abrams, Albert M. Lai
Abstract: Generative pre-trained transformer (GPT) models have shown promise in clinical entity and relation extraction tasks because of their precise extraction and contextual understanding capability. In this work, we further leverage the Unified Medical Language System (UMLS) knowledge base to accurately identify medical concepts and improve clinical entity and relation extraction at the document level. Our framework selects UMLS concepts relevant to the text and combines them with prompts to guide language models in extracting entities. Our experiments demonstrate that this initial concept mapping and the inclusion of these mapped concepts in the prompts improves extraction results compared to few-shot extraction tasks on generic language models that do not leverage UMLS. Further, our results show that this approach is more effective than the standard Retrieval Augmented Generation (RAG) technique, where retrieved data is compared with prompt embeddings to generate results. Overall, we find that integrating UMLS concepts with GPT models significantly improves entity and relation identification, outperforming the baseline and RAG models. By combining the precise concept mapping capability of knowledge-based approaches like UMLS with the contextual understanding capability of GPT, our method highlights the potential of these approaches in specialized domains like healthcare.
Authors: Ben Krause, Lucia Chen, Emmanuel Kahembwe
Abstract: We introduce the AutoGRAMS framework for programming multi-step interactions with language models. AutoGRAMS represents AI agents as a graph, where each node can execute either a language modeling instruction or traditional code. Likewise, transitions in the graph can be governed by either language modeling decisions or traditional branch logic. AutoGRAMS supports using variables as memory and allows nodes to call other AutoGRAMS graphs as functions. We show how AutoGRAMS can be used to design highly sophisticated agents, including self-referential agents that can modify their own graph. AutoGRAMS's graph-centric approach aids interpretability, controllability, and safety during the design, development, and deployment of AI agents. We provide our framework as open source at https://github.com/autograms/autograms .
Authors: Zhenhua Liu, Tong Zhu, Chuanyuan Tan, Wenliang Chen
Abstract: Large language models (LLMs) exhibit remarkable capabilities in understanding and generating natural language. However, these models can inadvertently memorize private information, posing significant privacy risks. This study addresses the challenge of enabling LLMs to protect specific individuals' private data without the need for complete retraining. We propose \return, a Real-world pErsonal daTa UnleaRNing dataset, comprising 2,492 individuals from Wikipedia with associated QA pairs, to evaluate machine unlearning (MU) methods for protecting personal data in a realistic scenario. Additionally, we introduce the Name-Aware Unlearning Framework (NAUF) for Privacy Protection, which enables the model to learn which individuals' information should be protected without affecting its ability to answer questions related to other unrelated individuals. Our extensive experiments demonstrate that NAUF achieves a state-of-the-art average unlearning score, surpassing the best baseline method by 5.65 points, effectively protecting target individuals' personal data while maintaining the model's general capabilities.
Authors: Xiaoyu Liu, Yun Zhang, Wei Li, Simiao Li, Xudong Huang, Hanting Chen, Yehui Tang, Jie Hu, Zhiwei Xiong, Yunhe Wang
Abstract: Knowledge distillation plays a key role in compressing the Large Language Models (LLMs), which boosts a small-size student model under large teacher models' guidance. However, existing LLM distillation methods overly rely on student-generated outputs, which may introduce generation errors and misguide the distillation process. Moreover, the distillation loss functions introduced in previous art struggle to align the most informative part due to the complex distribution of LLMs' outputs. To address these problems, we propose a multi-granularity semantic revision method for LLM distillation. At the sequence level, we propose a sequence correction and re-generation (SCRG) strategy. SCRG first calculates the semantic cognitive difference between the teacher and student to detect the error token, then corrects it with the teacher-generated one, and re-generates the sequence to reduce generation errors and enhance generation diversity. At the token level, we design a distribution adaptive clipping Kullback-Leibler (DAC-KL) loss as the distillation objective function. DAC-KL loss exploits a learnable sub-network to adaptively extract semantically dense areas from the teacher's output, avoiding the interference of redundant information in the distillation process. Finally, at the span level, we leverage the span priors of a sequence to compute the probability correlations within spans, and constrain the teacher and student's probability correlations to be consistent, further enhancing the transfer of semantic information. Extensive experiments across different model families with parameters ranging from 0.1B to 13B demonstrate the superiority of our method compared to existing methods.
Authors: Omid Rohanian, Mohammadmahdi Nouriborji, Olena Seminog, Rodrigo Furst, Thomas Mendy, Shanthi Levanita, Zaharat Kadri-Alab, Nusrat Jabin, Daniela Toale, Georgina Humphreys, Emilia Antonio, Adrian Bucher, Alice Norton, David A. Clifton
Abstract: This paper introduces the Pandemic PACT Advanced Categorisation Engine (PPACE) along with its associated dataset. PPACE is a fine-tuned model developed to automatically classify research abstracts from funded biomedical projects according to WHO-aligned research priorities. This task is crucial for monitoring research trends and identifying gaps in global health preparedness and response. Our approach builds on human-annotated projects, which are allocated one or more categories from a predefined list. A large language model is then used to generate `rationales' explaining the reasoning behind these annotations. This augmented data, comprising expert annotations and rationales, is subsequently used to fine-tune a smaller, more efficient model. Developed as part of the Pandemic PACT project, which aims to track and analyse research funding and clinical evidence for a wide range of diseases with outbreak potential, PPACE supports informed decision-making by research funders, policymakers, and independent researchers. We introduce and release both the trained model and the instruction-based dataset used for its training. Our evaluation shows that PPACE significantly outperforms its baselines. The release of PPACE and its associated dataset offers valuable resources for researchers in multilabel biomedical document classification and supports advancements in aligning biomedical research with key global health priorities.
Authors: Ge Gao, Jongin Kim, Sejin Paik, Ekaterina Novozhilova, Yi Liu, Sarah T. Bonna, Margrit Betke, Derry Tanti Wijaya
Abstract: Predicting emotions elicited by news headlines can be challenging as the task is largely influenced by the varying nature of people's interpretations and backgrounds. Previous works have explored classifying discrete emotions directly from news headlines. We provide a different approach to tackling this problem by utilizing people's explanations of their emotion, written in free-text, on how they feel after reading a news headline. Using the dataset BU-NEmo+ (Gao et al., 2022), we found that for emotion classification, the free-text explanations have a strong correlation with the dominant emotion elicited by the headlines. The free-text explanations also contain more sentimental context than the news headlines alone and can serve as a better input to emotion classification models. Therefore, in this work we explored generating emotion explanations from headlines by training a sequence-to-sequence transformer model and by using pretrained large language model, ChatGPT (GPT-4). We then used the generated emotion explanations for emotion classification. In addition, we also experimented with training the pretrained T5 model for the intermediate task of explanation generation before fine-tuning it for emotion classification. Using McNemar's significance test, methods that incorporate GPT-generated free-text emotion explanations demonstrated significant improvement (P-value < 0.05) in emotion classification from headlines, compared to methods that only use headlines. This underscores the value of using intermediate free-text explanations for emotion prediction tasks with headlines.
Authors: Roni Goldshmidt, Miriam Horovicz
Abstract: As large language models (LLMs) become increasingly prevalent in critical applications, the need for interpretable AI has grown. We introduce TokenSHAP, a novel method for interpreting LLMs by attributing importance to individual tokens or substrings within input prompts. This approach adapts Shapley values from cooperative game theory to natural language processing, offering a rigorous framework for understanding how different parts of an input contribute to a model's response. TokenSHAP leverages Monte Carlo sampling for computational efficiency, providing interpretable, quantitative measures of token importance. We demonstrate its efficacy across diverse prompts and LLM architectures, showing consistent improvements over existing baselines in alignment with human judgments, faithfulness to model behavior, and consistency. Our method's ability to capture nuanced interactions between tokens provides valuable insights into LLM behavior, enhancing model transparency, improving prompt engineering, and aiding in the development of more reliable AI systems. TokenSHAP represents a significant step towards the necessary interpretability for responsible AI deployment, contributing to the broader goal of creating more transparent, accountable, and trustworthy AI systems.
Authors: Shunsuke Kando, Yusuke Miyao, Jason Naradowsky, Shinnosuke Takamichi
Abstract: Traditional spoken language processing involves cascading an automatic speech recognition (ASR) system into text processing models. In contrast, "textless" methods process speech representations without ASR systems, enabling the direct use of acoustic speech features. Although their effectiveness is shown in capturing acoustic features, it is unclear in capturing lexical knowledge. This paper proposes a textless method for dependency parsing, examining its effectiveness and limitations. Our proposed method predicts a dependency tree from a speech signal without transcribing, representing the tree as a labeled sequence. scading method outperforms the textless method in overall parsing accuracy, the latter excels in instances with important acoustic features. Our findings highlight the importance of fusing word-level representations and sentence-level prosody for enhanced parsing performance. The code and models are made publicly available: https://github.com/mynlp/SpeechParser.
Authors: Garry Kuwanto, Eno-Abasi E. Urua, Priscilla Amondi Amuok, Shamsuddeen Hassan Muhammad, Anuoluwapo Aremu, Verrah Otiende, Loice Emma Nanyanga, Teresiah W. Nyoike, Aniefon D. Akpan, Nsima Ab Udouboh, Idongesit Udeme Archibong, Idara Effiong Moses, Ifeoluwatayo A. Ige, Benjamin Ajibade, Olumide Benjamin Awokoya, Idris Abdulmumin, Saminu Mohammad Aliyu, Ruqayya Nasir Iro, Ibrahim Said Ahmad, Deontae Smith, Praise-EL Michaels, David Ifeoluwa Adelani, Derry Tanti Wijaya, Anietie Andy
Abstract: Low-resource languages often face challenges in acquiring high-quality language data due to the reliance on translation-based methods, which can introduce the translationese effect. This phenomenon results in translated sentences that lack fluency and naturalness in the target language. In this paper, we propose a novel approach for data collection by leveraging storyboards to elicit more fluent and natural sentences. Our method involves presenting native speakers with visual stimuli in the form of storyboards and collecting their descriptions without direct exposure to the source text. We conducted a comprehensive evaluation comparing our storyboard-based approach with traditional text translation-based methods in terms of accuracy and fluency. Human annotators and quantitative metrics were used to assess translation quality. The results indicate a preference for text translation in terms of accuracy, while our method demonstrates worse accuracy but better fluency in the language focused.
Authors: He Li, Haoang Chi, Mingyu Liu, Wenjing Yang
Abstract: The emergence of large language models (LLMs) is a milestone in generative artificial intelligence, achieving significant success in text comprehension and generation tasks. Despite the tremendous success of LLMs in many downstream tasks, they suffer from severe hallucination problems, posing significant challenges to the practical applications of LLMs. Most of the works about LLMs' hallucinations focus on data quality. Self-attention is a core module in transformer-based LLMs, while its potential relationship with LLMs' hallucination has been hardly investigated. To fill this gap, we study this problem from a causal perspective. We propose a method to intervene in LLMs' self-attention layers and maintain their structures and sizes intact. Specifically, we disable different self-attention layers in several popular open-source LLMs and then compare their degrees of hallucination with the original ones. We evaluate the intervened LLMs on hallucination assessment benchmarks and conclude that disabling some specific self-attention layers in the front or tail of the LLMs can alleviate hallucination issues. The study paves a new way for understanding and mitigating LLMs' hallucinations.
Authors: Xunyu Zhu, Jian Li, Yong Liu, Can Ma, Weiping Wang
Abstract: Large Language Models (LLMs) have demonstrated exceptional proficiency in mathematical reasoning tasks due to their extensive parameter counts and training on vast datasets. Despite these capabilities, deploying LLMs is hindered by their computational demands. Distilling LLM mathematical reasoning into Smaller Language Models (SLMs) has emerged as a solution to this challenge, although these smaller models often suffer from errors in calculation and semantic understanding. Prior work has proposed Program-of-Thought Distillation (PoTD) to avoid calculation error. To further address semantic understanding errors, we propose Key-Point-Driven Mathematical Reasoning Distillation (KPDD). KPDD enhances the reasoning performance of SLMs by breaking down the problem-solving process into three stages: Core Question Extraction, Problem-Solving Information Extraction, and Step-by-Step Solution. This method is further divided into KPDD-CoT, which generates Chain-of-Thought rationales, and KPDD-PoT, which creates Program-of-Thought rationales. The experiment results show that KPDD-CoT significantly improves reasoning abilities, while KPDD-PoT achieves state-of-the-art performance in mathematical reasoning tasks. Our approach effectively mitigates misunderstanding errors, advancing the deployment of efficient and capable SLMs.
Authors: Zhiting Fan, Ruizhe Chen, Ruiling Xu, Zuozhu Liu
Abstract: Evaluating the bias in Large Language Models (LLMs) becomes increasingly crucial with their rapid development. However, existing evaluation methods rely on fixed-form outputs and cannot adapt to the flexible open-text generation scenarios of LLMs (e.g., sentence completion and question answering). To address this, we introduce BiasAlert, a plug-and-play tool designed to detect social bias in open-text generations of LLMs. BiasAlert integrates external human knowledge with inherent reasoning capabilities to detect bias reliably. Extensive experiments demonstrate that BiasAlert significantly outperforms existing state-of-the-art methods like GPT4-as-A-Judge in detecting bias. Furthermore, through application studies, we demonstrate the utility of BiasAlert in reliable LLM bias evaluation and bias mitigation across various scenarios. Model and code will be publicly released.
Authors: Barah Fazili, Koustava Goswami, Natwar Modani, Inderjeet Nair
Abstract: Retrieval augmented generation (RAG) with large language models (LLMs) for Question Answering (QA) entails furnishing relevant context within the prompt to facilitate the LLM in answer generation. During the generation, inaccuracies or hallucinations frequently occur due to two primary factors: inadequate or distracting context in the prompts, and the inability of LLMs to effectively reason through the facts. In this paper, we investigate whether providing aligned context via a carefully selected passage sequence leads to better answer generation by the LLM for multi-hop QA. We introduce, "GenSco", a novel approach of selecting passages based on the predicted decomposition of the multi-hop questions}. The framework consists of two distinct LLMs: (i) Generator LLM, which is used for question decomposition and final answer generation; (ii) an auxiliary open-sourced LLM, used as the scorer, to semantically guide the Generator for passage selection. The generator is invoked only once for the answer generation, resulting in a cost-effective and efficient approach. We evaluate on three broadly established multi-hop question answering datasets: 2WikiMultiHop, Adversarial HotPotQA and MuSiQue and achieve an absolute gain of $15.1$ and $5.9$ points in Exact Match score with respect to the best performing baselines over MuSiQue and 2WikiMultiHop respectively.
Authors: Md. Rafiul Biswas, Abrar Tasneem Abir, Wajdi Zaghouani
Abstract: This study addresses a binary classification task to determine whether a text sequence, either a sentence or paragraph, is subjective or objective. The task spans five languages: Arabic, Bulgarian, English, German, and Italian, along with a multilingual category. Our approach involved several key techniques. Initially, we preprocessed the data through parts of speech (POS) tagging, identification of question marks, and application of attention masks. We fine-tuned the sentiment-based Transformer model 'MarieAngeA13/Sentiment-Analysis-BERT' on our dataset. Given the imbalance with more objective data, we implemented a custom classifier that assigned greater weight to objective data. Additionally, we translated non-English data into English to maintain consistency across the dataset. Our model achieved notable results, scoring top marks for the multilingual dataset (Macro F1=0.7121) and German (Macro F1=0.7908). It ranked second for Arabic (Macro F1=0.4908) and Bulgarian (Macro F1=0.7169), third for Italian (Macro F1=0.7430), and ninth for English (Macro F1=0.6893).
Authors: Guillaume Rochette, Matthew J. Vowels
Abstract: psifx is a plug-and-play multi-modal feature extraction toolkit, aiming to facilitate and democratize the use of state-of-the-art machine learning techniques for human sciences research. It is motivated by a need (a) to automate and standardize data annotation processes, otherwise involving expensive, lengthy, and inconsistent human labor, such as the transcription or coding of behavior changes from audio and video sources; (b) to develop and distribute open-source community-driven psychology research software; and (c) to enable large-scale access and ease of use to non-expert users. The framework contains an array of tools for tasks, such as speaker diarization, closed-caption transcription and translation from audio, as well as body, hand, and facial pose estimation and gaze tracking from video. The package has been designed with a modular and task-oriented approach, enabling the community to add or update new tools easily. We strongly hope that this package will provide psychologists a simple and practical solution for efficiently a range of audio, linguistic, and visual features from audio and video, thereby creating new opportunities for in-depth study of real-time behavioral phenomena.
Authors: Aditi Khandelwal, Harman Singh, Hengrui Gu, Tianlong Chen, Kaixiong Zhou
Abstract: Large language models are often expected to constantly adapt to new sources of knowledge and knowledge editing techniques aim to efficiently patch the outdated model knowledge, with minimal modification. Most prior works focus on monolingual knowledge editing in English, even though new information can emerge in any language from any part of the world. We propose the Cross-Lingual Multi-Hop Knowledge Editing paradigm, for measuring and analyzing the performance of various SoTA knowledge editing techniques in a cross-lingual setup. Specifically, we create a parallel cross-lingual benchmark, CROLIN-MQUAKE for measuring the knowledge editing capabilities. Our extensive analysis over various knowledge editing techniques uncover significant gaps in performance between the cross-lingual and English-centric setting. Following this, we propose a significantly improved system for cross-lingual multi-hop knowledge editing, CLEVER-CKE. CLEVER-CKE is based on a retrieve, verify and generate knowledge editing framework, where a retriever is formulated to recall edited facts and support an LLM to adhere to knowledge edits. We develop language-aware and hard-negative based contrastive objectives for improving the cross-lingual and fine-grained fact retrieval and verification process used in this framework. Extensive experiments on three LLMs, eight languages, and two datasets show CLEVER-CKE's significant gains of up to 30% over prior methods.
Authors: Boris Orekhov
Abstract: The stylo package includes a frequency table that can be used to calculate distances between texts and thus independently solve the problem of attribution of The Cuckoo's Calling, a novel that J.K. Rowling said she wrote. However, the set of texts for this table is very vulnerable to criticism. The authors there are not modern, they wrote in a different genre. I set out to test the performance of the method on texts that are more relevant to the research question.
Authors: Adamu Lawan, Juhua Pu, Haruna Yunusa, Aliyu Umar, Muhammad Lawan
Abstract: Aspect-based sentiment Analysis (ABSA) identifies and evaluates sentiments toward specific aspects of entities within text, providing detailed insights beyond overall sentiment. However, Attention mechanisms and neural network models struggle with syntactic constraints, and the quadratic complexity of attention mechanisms hinders their adoption for capturing long-range dependencies between aspect and opinion words in ABSA. This complexity can lead to the misinterpretation of irrelevant con-textual words, restricting their effectiveness to short-range dependencies. Some studies have investigated merging semantic and syntactic approaches but face challenges in effectively integrating these methods. To address the above problems, we present MambaForGCN, a novel approach to enhance short and long-range dependencies between aspect and opinion words in ABSA. This innovative approach incorporates syntax-based Graph Convolutional Network (SynGCN) and MambaFormer (Mamba-Transformer) modules to encode input with dependency relations and semantic information. The Multihead Attention (MHA) and Mamba blocks in the MambaFormer module serve as channels to enhance the model with short and long-range dependencies between aspect and opinion words. We also introduce the Kolmogorov-Arnold Networks (KANs) gated fusion, an adaptively integrated feature representation system combining SynGCN and MambaFormer representations. Experimental results on three benchmark datasets demonstrate MambaForGCN's effectiveness, outperforming state-of-the-art (SOTA) baseline models.
Authors: Matthias Blume, Ghobad Heidari, Christoph Hewel
Abstract: A key capability in managing patent applications or a patent portfolio is comparing claims to other text, e.g. a patent specification. Because the language of claims is different from language used elsewhere in the patent application or in non-patent text, this has been challenging for computer based natural language processing. We test two new LLM-based approaches and find that both provide substantially better performance than previously published values. The ability to match dense information from one domain against much more distributed information expressed in a different vocabulary may also be useful beyond the intellectual property space.
Authors: Hyungjun Yoon, Biniyam Aschalew Tolera, Taesik Gong, Kimin Lee, Sung-Ju Lee
Abstract: Large language models (LLMs) have demonstrated exceptional abilities across various domains. However, utilizing LLMs for ubiquitous sensing applications remains challenging as existing text-prompt methods show significant performance degradation when handling long sensor data sequences. We propose a visual prompting approach for sensor data using multimodal LLMs (MLLMs). We design a visual prompt that directs MLLMs to utilize visualized sensor data alongside the target sensory task descriptions. Additionally, we introduce a visualization generator that automates the creation of optimal visualizations tailored to a given sensory task, eliminating the need for prior task-specific knowledge. We evaluated our approach on nine sensory tasks involving four sensing modalities, achieving an average of 10% higher accuracy than text-based prompts and reducing token costs by 15.8x. Our findings highlight the effectiveness and cost-efficiency of visual prompts with MLLMs for various sensory tasks.
Authors: Zhoutian Shao, Yuanning Cui, Wei Hu
Abstract: Knowledge graphs (KGs) are widely acknowledged as incomplete, and new entities are constantly emerging in the real world. Inductive KG reasoning aims to predict missing facts for these new entities. Among existing models, graph neural networks (GNNs) based ones have shown promising performance for this task. However, they are still challenged by inefficient message propagation due to the distance and scalability issues. In this paper, we propose a new inductive KG reasoning model, MStar, by leveraging conditional message passing neural networks (C-MPNNs). Our key insight is to select multiple query-specific starting entities to expand the scope of progressive propagation. To propagate query-related messages to a farther area within limited steps, we subsequently design a highway layer to propagate information toward these selected starting entities. Moreover, we introduce a training strategy called LinkVerify to mitigate the impact of noisy training samples. Experimental results validate that MStar achieves superior performance compared with state-of-the-art models, especially for distant entities.
Authors: Yu-Tzu Lee
Abstract: Most of the existing medication recommendation models are predicted with only structured data such as medical codes, with the remaining other large amount of unstructured or semi-structured data underutilization. To increase the utilization effectively, we proposed a method of enhancing medication recommendation with Large Language Model (LLM) text representation. LLM harnesses powerful language understanding and generation capabilities, enabling the extraction of information from complex and lengthy unstructured data such as clinical notes which contain complex terminology. This method can be applied to several existing base models we selected and improve medication recommendation performance with the combination representation of text and medical codes experiments on two different datasets. LLM text representation alone can even demonstrate a comparable ability to the medical code representation alone. Overall, this is a general method that can be applied to other models for improved recommendations.
Authors: Jun Wang, Eleftheria Briakou, Hamid Dadkhahi, Rishabh Agarwal, Colin Cherry, Trevor Cohn
Abstract: A critical component in knowledge distillation is the means of coupling the teacher and student. The predominant sequence knowledge distillation method involves supervised learning of the student against teacher-decoded outputs, and is exemplified by the current state of the art, which incorporates minimum Bayes risk (MBR) decoding. In this paper we seek to integrate MBR more tightly in distillation training, specifically by using several high scoring MBR translations, rather than a single selected sequence, thus capturing a rich diversity of teacher outputs. Our experiments on English to German and English to Japanese translation show consistent improvements over strong baseline methods for both tasks and with varying model sizes. Additionally, we conduct a detailed analysis focusing on data efficiency and capacity curse aspects to elucidate MBR-n and explore its further potential.
Authors: Yifan Song, Guoyin Wang, Sujian Li, Bill Yuchen Lin
Abstract: Current evaluations of large language models (LLMs) often overlook non-determinism, typically focusing on a single output per example. This limits our understanding of LLM performance variability in real-world applications. Our study addresses this issue by exploring key questions about the performance differences between greedy decoding and sampling, identifying benchmarks' consistency regarding non-determinism, and examining unique model behaviors. Through extensive experiments, we observe that greedy decoding generally outperforms sampling methods for most evaluated tasks. We also observe consistent performance across different LLM sizes and alignment methods, noting that alignment can reduce sampling variance. Moreover, our best-of-N sampling approach demonstrates that smaller LLMs can match or surpass larger models such as GPT-4-Turbo, highlighting the untapped potential of smaller LLMs. This research shows the importance of considering non-determinism in LLM evaluations and provides insights for future LLM development and evaluation.
Authors: Vincent Quantmeyer, Pablo Mosteiro, Albert Gatt
Abstract: Various benchmarks have been proposed to test linguistic understanding in pre-trained vision \& language (VL) models. Here we build on the existence task from the VALSE benchmark (Parcalabescu et al, 2022) which we use to test models' understanding of negation, a particularly interesting issue for multimodal models. However, while such VL benchmarks are useful for measuring model performance, they do not reveal anything about the internal processes through which these models arrive at their outputs in such visio-linguistic tasks. We take inspiration from the growing literature on model interpretability to explain the behaviour of VL models on the understanding of negation. Specifically, we approach these questions through an in-depth analysis of the text encoder in CLIP (Radford et al, 2021), a highly influential VL model. We localise parts of the encoder that process negation and analyse the role of attention heads in this task. Our contributions are threefold. We demonstrate how methods from the language model interpretability literature (such as causal tracing) can be translated to multimodal models and tasks; we provide concrete insights into how CLIP processes negation on the VALSE existence task; and we highlight inherent limitations in the VALSE dataset as a benchmark for linguistic understanding.
Authors: Songyang Zhang, Chuyu Zhang, Yingfan Hu, Haowen Shen, Kuikun Liu, Zerun Ma, Fengzhe Zhou, Wenwei Zhang, Xuming He, Dahua Lin, Kai Chen
Abstract: While LLM-Based agents, which use external tools to solve complex problems, have made significant progress, benchmarking their ability is challenging, thereby hindering a clear understanding of their limitations. In this paper, we propose an interactive evaluation framework, named CIBench, to comprehensively assess LLMs' ability to utilize code interpreters for data science tasks. Our evaluation framework includes an evaluation dataset and two evaluation modes. The evaluation dataset is constructed using an LLM-human cooperative approach and simulates an authentic workflow by leveraging consecutive and interactive IPython sessions. The two evaluation modes assess LLMs' ability with and without human assistance. We conduct extensive experiments to analyze the ability of 24 LLMs on CIBench and provide valuable insights for future LLMs in code interpreter utilization.
Authors: Xingzhi Zhou, Xin Dong, Chunhao Li, Yuning Bai, Yulong Xu, Ka Chun Cheung, Simon See, Xinpeng Song, Runshun Zhang, Xuezhong Zhou, Nevin L. Zhang
Abstract: Traditional Chinese medicine (TCM) relies on specific combinations of herbs in prescriptions to treat symptoms and signs, a practice that spans thousands of years. Predicting TCM prescriptions presents a fascinating technical challenge with practical implications. However, this task faces limitations due to the scarcity of high-quality clinical datasets and the intricate relationship between symptoms and herbs. To address these issues, we introduce DigestDS, a new dataset containing practical medical records from experienced experts in digestive system diseases. We also propose a method, TCM-FTP (TCM Fine-Tuning Pre-trained), to leverage pre-trained large language models (LLMs) through supervised fine-tuning on DigestDS. Additionally, we enhance computational efficiency using a low-rank adaptation technique. TCM-FTP also incorporates data augmentation by permuting herbs within prescriptions, capitalizing on their order-agnostic properties. Impressively, TCM-FTP achieves an F1-score of 0.8031, surpassing previous methods significantly. Furthermore, it demonstrates remarkable accuracy in dosage prediction, achieving a normalized mean square error of 0.0604. In contrast, LLMs without fine-tuning perform poorly. Although LLMs have shown capabilities on a wide range of tasks, this work illustrates the importance of fine-tuning for TCM prescription prediction, and we have proposed an effective way to do that.
Authors: Mar\'ia Mir\'o Maestre, Iv\'an Mart\'inez-Murillo, Tania J. Martin, Borja Navarro-Colorado, Antonio Ferr\'andez, Armando Su\'arez Cueto, Elena Lloret
Abstract: Generative Artificial Intelligence has grown exponentially as a result of Large Language Models (LLMs). This has been possible because of the impressive performance of deep learning methods created within the field of Natural Language Processing (NLP) and its subfield Natural Language Generation (NLG), which is the focus of this paper. Within the growing LLM family are the popular GPT-4, Bard and more specifically, tools such as ChatGPT have become a benchmark for other LLMs when solving most of the tasks involved in NLG research. This scenario poses new questions about the next steps for NLG and how the field can adapt and evolve to deal with new challenges in the era of LLMs. To address this, the present paper conducts a review of a representative sample of surveys recently published in NLG. By doing so, we aim to provide the scientific community with a research roadmap to identify which NLG aspects are still not suitably addressed by LLMs, as well as suggest future lines of research that should be addressed going forward.
Authors: Barah Fazili, Ashish Sunil Agrawal, Preethi Jyothi
Abstract: Large language models (LLMs) are very proficient text generators. We leverage this capability of LLMs to generate task-specific data via zero-shot prompting and promote cross-lingual transfer for low-resource target languages. Given task-specific data in a source language and a teacher model trained on this data, we propose using this teacher to label LLM generations and employ a set of simple data selection strategies that use the teacher's label probabilities. Our data selection strategies help us identify a representative subset of diverse generations that help boost zero-shot accuracies while being efficient, in comparison to using all the LLM generations (without any subset selection). We also highlight other important design choices that affect cross-lingual performance such as the use of translations of source data and what labels are best to use for the LLM generations. We observe significant performance gains across sentiment analysis and natural language inference tasks (of up to a maximum of 7.13 absolute points and 1.5 absolute points on average) across a number of target languages (Hindi, Marathi, Urdu, Swahili) and domains.
Authors: Asaf Achi Mordechai, Yoav Goldberg, Reut Tsarfaty
Abstract: Current Text-to-Code models demonstrate impressive capabilities in generating executable code from natural language snippets. However, current studies focus on technical instructions and programmer-oriented language, and it is an open question whether these models can effectively translate natural language descriptions given by non-technical users and express complex goals, to an executable program that contains an intricate flow - composed of API access and control structures as loops, conditions, and sequences. To unlock the challenge of generating a complete program from a plain non-technical description we present NoviCode, a novel NL Programming task, which takes as input an API and a natural language description by a novice non-programmer and provides an executable program as output. To assess the efficacy of models on this task, we provide a novel benchmark accompanied by test suites wherein the generated program code is assessed not according to their form, but according to their functional execution. Our experiments show that, first, NoviCode is indeed a challenging task in the code synthesis domain, and that generating complex code from non-technical instructions goes beyond the current Text-to-Code paradigm. Second, we show that a novel approach wherein we align the NL utterances with the compositional hierarchical structure of the code, greatly enhances the performance of LLMs on this task, compared with the end-to-end Text-to-Code counterparts.
Authors: Haipeng Luo, Qingfeng Sun, Can Xu, Pu Zhao, Qingwei Lin, Jianguang Lou, Shifeng Chen, Yansong Tang, Weizhu Chen
Abstract: Assessing the effectiveness of large language models (LLMs) presents substantial challenges. The method of conducting human-annotated battles in an online Chatbot Arena is a highly effective evaluative technique. However, this approach is limited by the costs and time required for human annotation. In this paper, we introduce Arena Learning, an innovative offline strategy designed to simulate these arena battles using AI-driven annotations to evaluate battle outcomes, thus facilitating the continuous improvement of the target model through both supervised fine-tuning and reinforcement learning. Arena Learning comprises two key elements. First, it ensures precise evaluations and maintains consistency between offline simulations and online competitions via WizardArena, a pipeline developed to accurately predict the Elo rankings of various models using a meticulously designed offline test set. Our results demonstrate that WizardArena's predictions closely align with those from the online Arena. Second, it involves the continuous improvement of training data based on the battle results and the refined model. We establish a data flywheel to iteratively update the training data by highlighting the weaknesses of the target model based on its battle results, enabling it to learn from the strengths of multiple different models. We apply Arena Learning to train our target model, WizardLM-$\beta$, and demonstrate significant performance enhancements across various metrics. This fully automated training and evaluation pipeline sets the stage for continuous advancements in various LLMs via post-training. Notably, Arena Learning plays a pivotal role in the success of WizardLM-2, and this paper serves both as an exploration of its efficacy and a foundational study for future discussions related to WizardLM-2 and its derivatives.
Authors: Louis Abraham, Charles Arnal, Antoine Marie
Abstract: Large Language Models have recently been applied to text annotation tasks from social sciences, equalling or surpassing the performance of human workers at a fraction of the cost. However, no inquiry has yet been made on the impact of prompt selection on labelling accuracy. In this study, we show that performance greatly varies between prompts, and we apply the method of automatic prompt optimization to systematically craft high quality prompts. We also provide the community with a simple, browser-based implementation of the method at https://prompt-ultra.github.io/ .
Authors: Usneek Singh, Jos\'e Cambronero, Sumit Gulwani, Aditya Kanade, Anirudh Khatry, Vu Le, Mukul Singh, Gust Verbruggen
Abstract: Large language models (LLMs) can be leveraged to help with writing formulas in spreadsheets, but resources on these formulas are scarce, impacting both the base performance of pre-trained models and limiting the ability to fine-tune them. Given a corpus of formulas, we can use a(nother) model to generate synthetic natural language utterances for fine-tuning. However, it is important to validate whether the NL generated by the LLM is indeed accurate to be beneficial for fine-tuning. In this paper, we provide empirical results on the impact of validating these synthetic training examples with surrogate objectives that evaluate the accuracy of the synthetic annotations. We demonstrate that validation improves performance over raw data across four models (2 open and 2 closed weight). Interestingly, we show that although validation tends to prune more challenging examples, it increases the complexity of problems that models can solve after being fine-tuned on validated data.
Authors: Yunxiao Shi, Xing Zi, Zijing Shi, Haimin Zhang, Qiang Wu, Min Xu
Abstract: Retrieval-augmented generation (RAG) techniques leverage the in-context learning capabilities of large language models (LLMs) to produce more accurate and relevant responses. Originating from the simple 'retrieve-then-read' approach, the RAG framework has evolved into a highly flexible and modular paradigm. A critical component, the Query Rewriter module, enhances knowledge retrieval by generating a search-friendly query. This method aligns input questions more closely with the knowledge base. Our research identifies opportunities to enhance the Query Rewriter module to Query Rewriter+ by generating multiple queries to overcome the Information Plateaus associated with a single query and by rewriting questions to eliminate Ambiguity, thereby clarifying the underlying intent. We also find that current RAG systems exhibit issues with Irrelevant Knowledge; to overcome this, we propose the Knowledge Filter. These two modules are both based on the instruction-tuned Gemma-2B model, which together enhance response quality. The final identified issue is Redundant Retrieval; we introduce the Memory Knowledge Reservoir and the Retriever Trigger to solve this. The former supports the dynamic expansion of the RAG system's knowledge base in a parameter-free manner, while the latter optimizes the cost for accessing external knowledge, thereby improving resource utilization and response efficiency. These four RAG modules synergistically improve the response quality and efficiency of the RAG system. The effectiveness of these modules has been validated through experiments and ablation studies across six common QA datasets. The source code can be accessed at https://github.com/Ancientshi/ERM4.
Authors: An Yang, Baosong Yang, Binyuan Hui, Bo Zheng, Bowen Yu, Chang Zhou, Chengpeng Li, Chengyuan Li, Dayiheng Liu, Fei Huang, Guanting Dong, Haoran Wei, Huan Lin, Jialong Tang, Jialin Wang, Jian Yang, Jianhong Tu, Jianwei Zhang, Jianxin Ma, Jin Xu, Jingren Zhou, Jinze Bai, Jinzheng He, Junyang Lin, Kai Dang, Keming Lu, Keqin Chen, Kexin Yang, Mei Li, Mingfeng Xue, Na Ni, Pei Zhang, Peng Wang, Ru Peng, Rui Men, Ruize Gao, Runji Lin, Shijie Wang, Shuai Bai, Sinan Tan, Tianhang Zhu, Tianhao Li, Tianyu Liu, Wenbin Ge, Xiaodong Deng, Xiaohuan Zhou, Xingzhang Ren, Xinyu Zhang, Xipin Wei, Xuancheng Ren, Yang Fan, Yang Yao, Yichang Zhang, Yu Wan, Yunfei Chu, Zeyu Cui, Zhenru Zhang, Zhihao Fan
Abstract: This report introduces the Qwen2 series, the latest addition to our large language models and large multimodal models. We release a comprehensive suite of foundational and instruction-tuned language models, encompassing a parameter range from 0.5 to 72 billion, featuring dense models and a Mixture-of-Experts model. Qwen2 surpasses most prior open-weight models, including its predecessor Qwen1.5, and exhibits competitive performance relative to proprietary models across diverse benchmarks on language understanding, generation, multilingual proficiency, coding, mathematics, and reasoning. The flagship model, Qwen2-72B, showcases remarkable performance: 84.2 on MMLU, 37.9 on GPQA, 64.6 on HumanEval, 89.5 on GSM8K, and 82.4 on BBH as a base language model. The instruction-tuned variant, Qwen2-72B-Instruct, attains 9.1 on MT-Bench, 48.1 on Arena-Hard, and 35.7 on LiveCodeBench. Moreover, Qwen2 demonstrates robust multilingual capabilities, proficient in approximately 30 languages, spanning English, Chinese, Spanish, French, German, Arabic, Russian, Korean, Japanese, Thai, Vietnamese, and more, underscoring its versatility and global reach. To foster community innovation and accessibility, we have made the Qwen2 model weights openly available on Hugging Face1 and ModelScope2, and the supplementary materials including example code on GitHub3. These platforms also include resources for quantization, fine-tuning, and deployment, facilitating a wide range of applications and research endeavors.
Authors: Anni Zou, Wenhao Yu, Hongming Zhang, Kaixin Ma, Deng Cai, Zhuosheng Zhang, Hai Zhao, Dong Yu
Abstract: Recently, there has been a growing interest among large language model (LLM) developers in LLM-based document reading systems, which enable users to upload their own documents and pose questions related to the document contents, going beyond simple reading comprehension tasks. Consequently, these systems have been carefully designed to tackle challenges such as file parsing, metadata extraction, multi-modal information understanding and long-context reading. However, no current benchmark exists to evaluate their performance in such scenarios, where a raw file and questions are provided as input, and a corresponding response is expected as output. In this paper, we introduce DocBench, a new benchmark designed to evaluate LLM-based document reading systems. Our benchmark involves a meticulously crafted process, including the recruitment of human annotators and the generation of synthetic questions. It includes 229 real documents and 1,102 questions, spanning across five different domains and four major types of questions. We evaluate both proprietary LLM-based systems accessible via web interfaces or APIs, and a parse-then-read pipeline employing open-source LLMs. Our evaluations reveal noticeable gaps between existing LLM-based document reading systems and human performance, underscoring the challenges of developing proficient systems. To summarize, DocBench aims to establish a standardized benchmark for evaluating LLM-based document reading systems under diverse real-world scenarios, thereby guiding future advancements in this research area.
Authors: Jing Yao, Xiaoyuan Yi, Xing Xie
Abstract: The rapid progress in Large Language Models (LLMs) poses potential risks such as generating unethical content. Assessing LLMs' values can help expose their misalignment, but relies on reference-free evaluators, e.g., fine-tuned LLMs or close-source ones like GPT-4, to identify values reflected in generated responses. Nevertheless, these evaluators face two challenges in open-ended value evaluation: they should align with changing human value definitions with minimal annotation, against their own bias (adaptability), and detect varying value expressions and scenarios robustly (generalizability). To handle these challenges, we introduce CLAVE, a novel framework which integrates two complementary LLMs, a large one to extract high-level value concepts from a few human labels, leveraging its extensive knowledge and generalizability, and a smaller one fine-tuned on such concepts to better align with human value understanding. This dual-model approach enables calibration with any value systems using <100 human-labeled samples per value type. Then we present ValEval, a comprehensive dataset comprising 13k+ (text,value,label) tuples across diverse domains, covering three major value systems. We benchmark the capabilities of 12+ popular LLM evaluators and analyze their strengths and weaknesses. Our findings reveal that combining fine-tuned small models and prompt-based large ones serves as a superior balance in value evaluation.
Authors: Damir Koren\v{c}i\'c, Berta Chulvi, Xavier Bonet Casals, Alejandro Toselli, Mariona Taul\'e, Paolo Rosso
Abstract: The current prevalence of conspiracy theories on the internet is a significant issue, tackled by many computational approaches. However, these approaches fail to recognize the relevance of distinguishing between texts which contain a conspiracy theory and texts which are simply critical and oppose mainstream narratives. Furthermore, little attention is usually paid to the role of inter-group conflict in oppositional narratives. We contribute by proposing a novel topic-agnostic annotation scheme that differentiates between conspiracies and critical texts, and that defines span-level categories of inter-group conflict. We also contribute with the multilingual XAI-DisInfodemics corpus (English and Spanish), which contains a high-quality annotation of Telegram messages related to COVID-19 (5,000 messages per language). We also demonstrate the feasibility of an NLP-based automatization by performing a range of experiments that yield strong baseline solutions. Finally, we perform an analysis which demonstrates that the promotion of intergroup conflict and the presence of violence and anger are key aspects to distinguish between the two types of oppositional narratives, i.e., conspiracy vs. critical.
Authors: Andrew Halterman, Katherine A. Keith
Abstract: Codebooks -- documents that operationalize constructs and outline annotation procedures -- are used almost universally by social scientists when coding unstructured political texts. Recently, to reduce manual annotation costs, political scientists have looked to generative large language models (LLMs) to label and analyze text data. However, previous work using LLMs for classification has implicitly relied on the universal label assumption -- correct classification of documents is possible using only a class label or minimal definition and the information that the LLM inductively learns during its pre-training. In contrast, we argue that political scientists who care about valid measurement should instead make a codebook-construct label assumption -- an LLM should follow the definition and exclusion criteria of a construct/label provided in a codebook. In this work, we collect and curate three political science datasets and their original codebooks and conduct a set of experiments to understand whether LLMs comply with codebook instructions, whether rewriting codebooks improves performance, and whether instruction-tuning LLMs on codebook-document-label tuples improves performance over zero-shot classification. Using Mistral 7B Instruct as our LLM, we find re-structuring the original codebooks gives modest gains in zero-shot performance but the model still struggles to comply with the constraints of the codebooks. Optimistically, instruction-tuning Mistral on one of our datasets gives significant gains over zero-shot inference (0.76 versus 0.53 micro F1). We hope our conceptualization of the codebook-specific task, assumptions, and instruction-tuning pipeline as well our semi-structured LLM codebook format will help political scientists readily adapt to the LLM era.
Authors: Hannah Sansford, Nicholas Richardson, Hermina Petric Maretic, Juba Nait Saada
Abstract: Methods to evaluate Large Language Model (LLM) responses and detect inconsistencies, also known as hallucinations, with respect to the provided knowledge, are becoming increasingly important for LLM applications. Current metrics fall short in their ability to provide explainable decisions, systematically check all pieces of information in the response, and are often too computationally expensive to be used in practice. We present GraphEval: a hallucination evaluation framework based on representing information in Knowledge Graph (KG) structures. Our method identifies the specific triples in the KG that are prone to hallucinations and hence provides more insight into where in the response a hallucination has occurred, if at all, than previous methods. Furthermore, using our approach in conjunction with state-of-the-art natural language inference (NLI) models leads to an improvement in balanced accuracy on various hallucination benchmarks, compared to using the raw NLI models. Lastly, we explore the use of GraphEval for hallucination correction by leveraging the structure of the KG, a method we name GraphCorrect, and demonstrate that the majority of hallucinations can indeed be rectified.
Authors: Rui Yang, Boming Yang, Sixun Ouyang, Tianwei She, Aosong Feng, Yuang Jiang, Freddy Lecue, Jinghui Lu, Irene Li
Abstract: Knowledge graphs (KGs) are crucial in the field of artificial intelligence and are widely applied in downstream tasks, such as enhancing Question Answering (QA) systems. The construction of KGs typically requires significant effort from domain experts. Recently, Large Language Models (LLMs) have been used for knowledge graph construction (KGC), however, most existing approaches focus on a local perspective, extracting knowledge triplets from individual sentences or documents. In this work, we introduce Graphusion, a zero-shot KGC framework from free text. The core fusion module provides a global view of triplets, incorporating entity merging, conflict resolution, and novel triplet discovery. We showcase how Graphusion could be applied to the natural language processing (NLP) domain and validate it in the educational scenario. Specifically, we introduce TutorQA, a new expert-verified benchmark for graph reasoning and QA, comprising six tasks and a total of 1,200 QA pairs. Our evaluation demonstrates that Graphusion surpasses supervised baselines by up to 10% in accuracy on link prediction. Additionally, it achieves average scores of 2.92 and 2.37 out of 3 in human evaluations for concept entity extraction and relation recognition, respectively.
Authors: Wenhao Zhu, Sizhe Liu, Shujian Huang, Shuaijie She, Chris Wendler, Jiajun Chen
Abstract: Decoding by contrasting layers (DoLa), is designed to improve the generation quality of large language models (LLMs) by contrasting the prediction probabilities between an early exit output (amateur logits) and the final output (expert logits). However, we find that this approach does not work well on non-English tasks. Inspired by previous interpretability work on language transition during the model's forward pass, we discover that this issue arises from a language mismatch between early exit output and final output. In this work, we propose an improved contrastive decoding algorithm that is effective for diverse languages beyond English. To obtain more helpful amateur logits, we devise two strategies to skip a set of bottom, language-agnostic layers based on our preliminary analysis. Experimental results on multilingual reasoning benchmarks demonstrate that our proposed method outperforms previous contrastive decoding baselines and substantially improves LLM's chain-of-thought reasoning accuracy across 11 languages. The project will be available at: https://github.com/NJUNLP/SkipLayerCD.
Authors: Jinhao Jiang, Junyi Li, Wayne Xin Zhao, Yang Song, Tao Zhang, Ji-Rong Wen
Abstract: Adapting general large language models (LLMs) to specialized domains presents great challenges due to varied data distributions. This adaptation typically requires continual pre-training on massive domain-specific corpora to facilitate knowledge memorization, followed by training to apply this knowledge following human instructions and preferences. However, this method may result in inefficient knowledge memorization due to a lack of awareness of knowledge utilization and imposes substantial demands on LLMs to simultaneously learn knowledge utilization and format alignment with limited training samples. To facilitate the domain adaptation of LLM, we revise this process and propose a new domain adaptation framework including domain knowledge learning and general format alignment, called Mix-CPT. Specifically, we first conduct a knowledge mixture continual pre-training that concurrently focuses on knowledge memorization and utilization, allowing for mutual reinforcement. To avoid catastrophic forgetting during the continual pre-training process, we further incorporate a logit swap self-distillation constraint. Subsequently, leveraging the knowledge and capabilities acquired during continual pre-training, we efficiently perform instruction tuning and alignment with a few general training samples to achieve format alignment. Extensive experiments demonstrate that our proposed Mix-CPT framework can simultaneously improve the task-solving capabilities of LLMs on the target and general domains compared to the traditional adaptation methods.
Authors: Shengjie Ma, Chengjin Xu, Xuhui Jiang, Muzhi Li, Huaren Qu, Jian Guo
Abstract: Retrieval-augmented generation (RAG) has significantly advanced large language models (LLMs) by enabling dynamic information retrieval to mitigate knowledge gaps and hallucinations in generated content. However, these systems often falter with complex reasoning and consistency across diverse queries. In this work, we present Think-on-Graph 2.0, an enhanced RAG framework that aligns questions with the knowledge graph and uses it as a navigational tool, which deepens and refines the RAG paradigm for information collection and integration. The KG-guided navigation fosters deep and long-range associations to uphold logical consistency and optimize the scope of retrieval for precision and interoperability. In conjunction, factual consistency can be better ensured through semantic similarity guided by precise directives. ToG${2.0}$ not only improves the accuracy and reliability of LLMs' responses but also demonstrates the potential of hybrid structured knowledge systems to significantly advance LLM reasoning, aligning it closer to human-like performance. We conducted extensive experiments on four public datasets to demonstrate the advantages of our method compared to the baseline.
Authors: Pawe{\l} Zyblewski, Jakub Klikowski, Weronika Borek-Marciniec, Pawe{\l} Ksieniewicz
Abstract: Tabular data is considered the last unconquered castle of deep learning, yet the task of data stream classification is stated to be an equally important and demanding research area. Due to the temporal constraints, it is assumed that deep learning methods are not the optimal solution for application in this field. However, excluding the entire -- and prevalent -- group of methods seems rather rash given the progress that has been made in recent years in its development. For this reason, the following paper is the first to present an approach to natural language data stream classification using the sentence space method, which allows for encoding text into the form of a discrete digital signal. This allows the use of convolutional deep networks dedicated to image classification to solve the task of recognizing fake news based on text data. Based on the real-life Fakeddit dataset, the proposed approach was compared with state-of-the-art algorithms for data stream classification based on generalization ability and time complexity.
Authors: Tu Vu, Kalpesh Krishna, Salaheddin Alzubi, Chris Tar, Manaal Faruqui, Yun-Hsuan Sung
Abstract: As large language models (LLMs) advance, it becomes more challenging to reliably evaluate their output due to the high costs of human evaluation. To make progress towards better LLM autoraters, we introduce FLAMe, a family of Foundational Large Autorater Models. FLAMe is trained on our large and diverse collection of 100+ quality assessment tasks comprising 5M+ human judgments, curated and standardized using publicly released human evaluations from previous research. FLAMe significantly improves generalization to a wide variety of held-out tasks, outperforming LLMs trained on proprietary data like GPT-4 and Claude-3 on many tasks. We show that FLAMe can also serve as a powerful starting point for further downstream fine-tuning, using reward modeling evaluation as a case study (FLAMe-RM). Notably, on RewardBench, our FLAMe-RM-24B model (with an accuracy of 87.8%) is the top-performing generative model trained exclusively on permissively licensed data, outperforming both GPT-4-0125 (85.9%) and GPT-4o (84.7%). Additionally, we explore a more computationally efficient approach using a novel tail-patch fine-tuning strategy to optimize our FLAMe multitask mixture for reward modeling evaluation (FLAMe-Opt-RM), offering competitive RewardBench performance while requiring approximately 25x less training datapoints. Overall, our FLAMe variants outperform all popular proprietary LLM-as-a-Judge models we consider across 8 out of 12 autorater evaluation benchmarks, encompassing 53 quality assessment tasks, including RewardBench and LLM-AggreFact. Finally, our analysis reveals that FLAMe is significantly less biased than these LLM-as-a-Judge models on the CoBBLEr autorater bias benchmark, while effectively identifying high-quality responses for code generation.
Authors: Tim Menzner, Jochen L. Leidner
Abstract: The increasing consumption of news online in the 21st century coincided with increased publication of disinformation, biased reporting, hate speech and other unwanted Web content. We describe BiasScanner, an application that aims to strengthen democracy by supporting news consumers with scrutinizing news articles they are reading online. BiasScanner contains a server-side pre-trained large language model to identify biased sentences of news articles and a front-end Web browser plug-in. At the time of writing, BiasScanner can identify and classify more than two dozen types of media bias at the sentence level, making it the most fine-grained model and only deployed application (automatic system in use) of its kind. It was implemented in a light-weight and privacy-respecting manner, and in addition to highlighting likely biased sentence it also provides explanations for each classification decision as well as a summary analysis for each news article. While prior research has addressed news bias detection, we are not aware of any work that resulted in a deployed browser plug-in (c.f. also biasscanner.org for a Web demo).
Authors: Dylan Bouchard
Abstract: Large language models (LLMs) can exhibit bias in a variety of ways. Such biases can create or exacerbate unfair outcomes for certain groups within a protected attribute, including, but not limited to sex, race, sexual orientation, or age. This paper aims to provide a technical guide for practitioners to assess bias and fairness risks in LLM use cases. The main contribution of this work is a decision framework that allows practitioners to determine which metrics to use for a specific LLM use case. To achieve this, this study categorizes LLM bias and fairness risks, maps those risks to a taxonomy of LLM use cases, and then formally defines various metrics to assess each type of risk. As part of this work, several new bias and fairness metrics are introduced, including innovative counterfactual metrics as well as metrics based on stereotype classifiers. Instead of focusing solely on the model itself, the sensitivity of both prompt-risk and model-risk are taken into account by defining evaluations at the level of an LLM use case, characterized by a model and a population of prompts. Furthermore, because all of the evaluation metrics are calculated solely using the LLM output, the proposed framework is highly practical and easily actionable for practitioners.
Authors: Sai Sena Chinnakonduru, Astarag Mohapatra
Abstract: The attention mechanism forms the foundational blocks for transformer language models. Recent approaches show that scaling the model achieves human-level performance. However, with increasing demands for scaling and constraints on hardware memory, the inference costs of these models remain high. To reduce the inference time, Multi-Query Attention (MQA) and Grouped-Query Attention (GQA) were proposed in (Shazeer, 2019) and (Ainslieet al., 2023) respectively. In this paper, we propose a variation of Grouped-Query Attention, termed Weighted Grouped-Query Attention (WGQA). We introduced new learnable parameters for each key and value head in the T5 decoder attention blocks, enabling the model to take a weighted average during finetuning. Our model achieves an average of 0.53% improvement over GQA, and the performance converges to traditional Multi-head attention (MHA) with no additional overhead during inference. We evaluated the introduction of these parameters and subsequent finetuning informs the model about the grouping mechanism during training, thereby enhancing performance. Additionally, we demonstrate the scaling laws in our analysis by comparing the results between T5-small and T5-base architecture.
Authors: Dilara Soylu, Christopher Potts, Omar Khattab
Abstract: Natural Language Processing (NLP) systems are increasingly taking the form of multi-stage pipelines involving multiple distinct language models (LMs) and prompting strategies. Here we address the question of how to fine-tune such systems to improve their performance. We cast this as a problem of optimizing the underlying LM weights and the prompting strategies together, and consider a challenging but highly realistic scenario in which we have no gold labels for any intermediate stages in the pipeline. To address this challenge, we evaluate approximate optimization strategies in which we bootstrap training labels for all pipeline stages and use these to optimize the pipeline's prompts and fine-tune its weights alternatingly. In experiments with multi-hop QA, mathematical reasoning, and feature-based classification, we find that simple approaches for optimizing the prompts and weights together outperform directly optimizing weights alone and prompts alone by up to 65% and 5%, respectively, on average across LMs and tasks. We will release our new optimizers in DSPy at http://dspy.ai
URLs: http://dspy.ai
Authors: Shachar Don-Yehiya, Leshem Choshen, Omri Abend
Abstract: Human feedback data is a critical component in developing language models. However, collecting this feedback is costly and ultimately not scalable. We propose a scalable method for extracting feedback that users naturally include when interacting with chat models, and leveraging it for model training. We are further motivated by previous work that showed there are also qualitative advantages to using naturalistic (rather than auto-generated) feedback, such as less hallucinations and biases. We manually annotated conversation data to confirm the presence of naturally occurring feedback in a standard corpus, finding that as much as 30% of the chats include explicit feedback. We apply our method to over 1M conversations to obtain hundreds of thousands of feedback samples. Training with the extracted feedback shows significant performance improvements over baseline models, demonstrating the efficacy of our approach in enhancing model alignment to human preferences.
Authors: Dan Friedman, Abhishek Panigrahi, Danqi Chen
Abstract: Transformer-based chatbots can conduct fluent, natural-sounding conversations, but we have limited understanding of the mechanisms underlying their behavior. Prior work has taken a bottom-up approach to understanding Transformers by constructing Transformers for various synthetic and formal language tasks, such as regular expressions and Dyck languages. However, it is not obvious how to extend this approach to understand more naturalistic conversational agents. In this work, we take a step in this direction by constructing a Transformer that implements the ELIZA program, a classic, rule-based chatbot. ELIZA illustrates some of the distinctive challenges of the conversational setting, including both local pattern matching and long-term dialog state tracking. We build on constructions from prior work -- in particular, for simulating finite-state automata -- showing how simpler constructions can be composed and extended to give rise to more sophisticated behavior. Next, we train Transformers on a dataset of synthetically generated ELIZA conversations and investigate the mechanisms the models learn. Our analysis illustrates the kinds of mechanisms these models tend to prefer -- for example, models favor an induction head mechanism over a more precise, position based copying mechanism; and using intermediate generations to simulate recurrent data structures, like ELIZA's memory mechanisms. Overall, by drawing an explicit connection between neural chatbots and interpretable, symbolic mechanisms, our results offer a new setting for mechanistic analysis of conversational agents.
Authors: Chengguang Gan, Qingyu Yin, Xinyang He, Hanjun Wei, Yunhao Liang, Younghun Lim, Shijian Wang, Hexiang Huang, Qinghao Zhang, Shiwen Ni, Tatsunori Mori
Abstract: The Mutual Reinforcement Effect (MRE) represents a promising avenue in information extraction and multitasking research. Nevertheless, its applicability has been constrained due to the exclusive availability of MRE mix datasets in Japanese, thereby limiting comprehensive exploration by the global research community. To address this limitation, we introduce a Multilingual MRE mix dataset (MMM) that encompasses 21 sub-datasets in English, Japanese, and Chinese. In this paper, we also propose a method for dataset translation assisted by Large Language Models (LLMs), which significantly reduces the manual annotation time required for dataset construction by leveraging LLMs to translate the original Japanese datasets. Additionally, we have enriched the dataset by incorporating open-domain Named Entity Recognition (NER) and sentence classification tasks. Utilizing this expanded dataset, we developed a unified input-output framework to train an Open-domain Information Extraction Large Language Model (OIELLM). The OIELLM model demonstrates the capability to effectively process novel MMM datasets, exhibiting significant improvements in performance.
Authors: Hongyu Wang, Shuming Ma, Ruiping Wang, Furu Wei
Abstract: We introduce, Q-Sparse, a simple yet effective approach to training sparsely-activated large language models (LLMs). Q-Sparse enables full sparsity of activations in LLMs which can bring significant efficiency gains in inference. This is achieved by applying top-K sparsification to the activations and the straight-through-estimator to the training. The key results from this work are, (1) Q-Sparse can achieve results comparable to those of baseline LLMs while being much more efficient at inference time; (2) We present an inference-optimal scaling law for sparsely-activated LLMs; (3) Q-Sparse is effective in different settings, including training-from-scratch, continue-training of off-the-shelf LLMs, and finetuning; (4) Q-Sparse works for both full-precision and 1-bit LLMs (e.g., BitNet b1.58). Particularly, the synergy of BitNet b1.58 and Q-Sparse (can be equipped with MoE) provides the cornerstone and a clear path to revolutionize the efficiency, including cost and energy consumption, of future LLMs.
Authors: Seungwoo Son, Wonpyo Park, Woohyun Han, Kyuyeun Kim, Jaeho Lee
Abstract: Despite recent advances in LLM quantization, activation quantization remains to be challenging due to the activation outliers. Conventional remedies, e.g., mixing precisions for different channels, introduce extra overhead and reduce the speedup. In this work, we develop a simple yet effective strategy to facilitate per-tensor activation quantization by preventing the generation of problematic tokens. Precisely, we propose a method to find a set of key-value cache, coined CushionCache, which mitigates outliers in subsequent tokens when inserted as a prefix. CushionCache works in two steps: First, we greedily search for a prompt token sequence that minimizes the maximum activation values in subsequent tokens. Then, we further tune the token cache to regularize the activations of subsequent tokens to be more quantization-friendly. The proposed method successfully addresses activation outliers of LLMs, providing a substantial performance boost for per-tensor activation quantization methods. We thoroughly evaluate our method over a wide range of models and benchmarks and find that it significantly surpasses the established baseline of per-tensor W8A8 quantization and can be seamlessly integrated with the recent activation quantization method.
Authors: Teng Ye, Jingnan Zheng, Junhui Jin, Jingyi Qiu, Wei Ai, Qiaozhu Mei
Abstract: While small businesses are increasingly turning to online crowdfunding platforms for essential funding, over 40% of these campaigns may fail to raise any money, especially those from low socio-economic areas. We utilize the latest advancements in AI technology to identify crucial factors that influence the success of crowdfunding campaigns and to improve their fundraising outcomes by strategically optimizing these factors. Our best-performing machine learning model accurately predicts the fundraising outcomes of 81.0% of campaigns, primarily based on their textual descriptions. Interpreting the machine learning model allows us to provide actionable suggestions on improving the textual description before launching a campaign. We demonstrate that by augmenting just three aspects of the narrative using a large language model, a campaign becomes more preferable to 83% human evaluators, and its likelihood of securing financial support increases by 11.9%. Our research uncovers the effective strategies for crafting descriptions for small business fundraising campaigns and opens up a new realm in integrating large language models into crowdfunding methodologies.
Authors: Sakib Shahriar, Brady Lund, Nishith Reddy Mannuru, Muhammad Arbab Arshad, Kadhim Hayawi, Ravi Varma Kumar Bevara, Aashrith Mannuru, Laiba Batool
Abstract: As large language models (LLMs) continue to advance, evaluating their comprehensive capabilities becomes significant for their application in various fields. This research study comprehensively evaluates the language, vision, speech, and multimodal capabilities of GPT-4o. The study employs standardized exam questions, reasoning tasks, and translation assessments to assess the model's language capability. Additionally, GPT-4o's vision and speech capabilities are tested through image classification and object recognition tasks, as well as accent classification. The multimodal evaluation assesses the model's performance in integrating visual and linguistic data. Our findings reveal that GPT-4o demonstrates high accuracy and efficiency across multiple domains in language and reasoning capabilities, excelling in tasks that require few-shot learning. GPT-4o also provides notable improvements in multimodal tasks compared to its predecessors. However, the model shows variability and faces limitations in handling complex and ambiguous inputs, particularly in audio and vision capabilities. This paper highlights the need for more comprehensive benchmarks and robust evaluation frameworks, encompassing qualitative assessments involving human judgment as well as error analysis. Future work should focus on expanding datasets, investigating prompt-based assessment, and enhancing few-shot learning techniques to test the model's practical applicability and performance in real-world scenarios.
Authors: Jacob Nielsen, Peter Schneider-Kamp
Abstract: Recently proposed methods for 1-bit and 1.58-bit quantization aware training investigate the performance and behavior of these methods in the context of large language models, finding state-of-the-art performance for models with more than 3B parameters. In this work, we investigate 1.58-bit quantization for small language and vision models ranging from 100K to 48M parameters. We introduce a variant of BitNet b1.58, which allows to rely on the median rather than the mean in the quantization process. Through extensive experiments we investigate the performance of 1.58-bit models obtained through quantization aware training. We further investigate the robustness of 1.58-bit quantization-aware training to changes in the learning rate and regularization through weight decay, finding different patterns for small language and vision models than previously reported for large language models. Our results showcase that 1.58-bit quantization-aware training provides state-of-the-art performance for small language models when doubling hidden layer sizes and reaches or even surpasses state-of-the-art performance for small vision models of identical size. Ultimately, we demonstrate that 1.58-bit quantization-aware training is a viable and promising approach also for training smaller deep learning networks, facilitating deployment of such models in low-resource use-cases and encouraging future research.
Authors: Hang Jiang, Doug Beeferman, William Brannon, Andrew Heyward, Deb Roy
Abstract: Words often carry different meanings for people from diverse backgrounds. Today's era of social polarization demands that we choose words carefully to prevent miscommunication, especially in political communication and journalism. To address this issue, we introduce the Bridging Dictionary, an interactive tool designed to illuminate how words are perceived by people with different political views. The Bridging Dictionary includes a static, printable document featuring 796 terms with summaries generated by a large language model. These summaries highlight how the terms are used distinctively by Republicans and Democrats. Additionally, the Bridging Dictionary offers an interactive interface that lets users explore selected words, visualizing their frequency, sentiment, summaries, and examples across political divides. We present a use case for journalists and emphasize the importance of human agency and trust in further enhancing this tool. The deployed version of Bridging Dictionary is available at https://dictionary.ccc-mit.org/.
Authors: Lecheng Kong, Jiarui Feng, Hao Liu, Chengsong Huang, Jiaxin Huang, Yixin Chen, Muhan Zhang
Abstract: Foundation models, such as Large Language Models (LLMs) or Large Vision Models (LVMs), have emerged as one of the most powerful tools in the respective fields. However, unlike text and image data, graph data do not have a definitive structure, posing great challenges to developing a Graph Foundation Model (GFM). For example, current attempts at designing general graph models either transform graph data into a language format for LLM-based prediction or still train a GNN model with LLM as an assistant. The former can handle unlimited tasks, while the latter captures graph structure much better -- yet, no existing work can achieve both simultaneously. In this paper, we identify three key desirable properties of a GFM: self-supervised pretraining, fluidity in tasks, and graph awareness. To account for these properties, we extend the conventional language modeling to the graph domain and propose a novel generative graph language model GOFA to solve the problem. The model interleaves randomly initialized GNN layers into a frozen pre-trained LLM so that the semantic and structural modeling abilities are organically combined. GOFA is pre-trained on newly proposed graph-level next-word prediction, question-answering, and structural tasks to obtain the above GFM properties. The pre-trained model is further fine-tuned on downstream tasks to obtain task-solving ability. The fine-tuned model is evaluated on various downstream tasks, demonstrating a strong ability to solve structural and contextual problems in zero-shot scenarios. The code is available at https://github.com/JiaruiFeng/GOFA.
Authors: Shentong Mo, Russ Salakhutdinov, Louis-Philippe Morency, Paul Pu Liang
Abstract: The Internet of Things (IoT) network integrating billions of smart physical devices embedded with sensors, software, and communication technologies is a critical and rapidly expanding component of our modern world. The IoT ecosystem provides a rich source of real-world modalities such as motion, thermal, geolocation, imaging, depth, sensors, and audio to recognize the states of humans and physical objects. Machine learning presents a rich opportunity to automatically process IoT data at scale, enabling efficient inference for understanding human wellbeing, controlling physical devices, and interconnecting smart cities. To realize this potential, we introduce IoT-LM, an open-source large multisensory language model tailored for the IoT ecosystem. IoT-LM is enabled by two technical contributions: the first is MultiIoT, the most expansive unified IoT dataset to date, encompassing over 1.15 million samples from 12 modalities and 8 tasks prepared for multisensory pre-training and instruction-tuning. The second is a new multisensory multitask adapter layer to condition pre-trained large language models on multisensory IoT data. Not only does IoT-LM yield substantial improvements on 8 supervised IoT classification tasks, but it also demonstrates new interactive question-answering, reasoning, and dialog capabilities conditioned on IoT sensors. We release IoT-LM's data sources and new multisensory language modeling framework.
Authors: Lingwei Meng, Jiawen Kang, Yuejiao Wang, Zengrui Jin, Xixin Wu, Xunying Liu, Helen Meng
Abstract: Multi-talker speech recognition and target-talker speech recognition, both involve transcription in multi-talker contexts, remain significant challenges. However, existing methods rarely attempt to simultaneously address both tasks. In this study, we propose a pioneering approach to empower Whisper, which is a speech foundation model, to tackle joint multi-talker and target-talker speech recognition tasks. Specifically, (i) we freeze Whisper and plug a Sidecar separator into its encoder to separate mixed embedding for multiple talkers; (ii) a Target Talker Identifier is introduced to identify the embedding flow of the target talker on the fly, requiring only three-second enrollment speech as a cue; (iii) soft prompt tuning for decoder is explored for better task adaptation. Our method outperforms previous methods on two- and three-talker LibriMix and LibriSpeechMix datasets for both tasks, and delivers acceptable zero-shot performance on multi-talker ASR on AishellMix Mandarin dataset.
Authors: Chun-Yi Kuan, Chih-Kai Yang, Wei-Ping Huang, Ke-Han Lu, Hung-yi Lee
Abstract: In this work, we introduce Speech-Copilot, a modular framework for instruction-oriented speech-processing tasks that minimizes human effort in toolset construction. Unlike end-to-end methods using large audio-language models, Speech-Copilot builds speech processing-specific toolsets by analyzing pre-collected task instructions and breaking tasks into manageable sub-tasks. It features a flexible agent based on large language models that performs tasks through program generation. Our approach achieves state-of-the-art performance on the Dynamic-SUPERB benchmark, demonstrating its effectiveness across diverse speech-processing tasks. Key contributions include: 1) developing an innovative framework for speech processing-specific toolset construction, 2) establishing a high-performing agent based on large language models, and 3) offering a new perspective on addressing challenging instruction-oriented speech-processing tasks. Without additional training processes required by end-to-end approaches, our method provides a flexible and extendable solution for a wide range of speech-processing applications.
Authors: Lingwei Wei, Dou Hu, Wei Zhou, Songlin Hu
Abstract: Many fake news detection studies have achieved promising performance by extracting effective semantic and structure features from both content and propagation trees. However, it is challenging to apply them to practical situations, especially when using the trained propagation-based models to detect news with no propagation data. Towards this scenario, we study a new task named cold-start fake news detection, which aims to detect content-only samples with missing propagation. To achieve the task, we design a simple but effective Structure Adversarial Net (SAN) framework to learn transferable features from available propagation to boost the detection of content-only samples. SAN introduces a structure discriminator to estimate dissimilarities among learned features with and without propagation, and further learns structure-invariant features to enhance the generalization of existing propagation-based methods for content-only samples. We conduct qualitative and quantitative experiments on three datasets. Results show the challenge of the new task and the effectiveness of our SAN framework.
Authors: Alex Dornburg, Kristin Davin
Abstract: The advent of generative AI models holds tremendous potential for aiding teachers in the generation of pedagogical materials. However, numerous knowledge gaps concerning the behavior of these models obfuscate the generation of research-informed guidance for their effective usage. Here we assess trends in prompt specificity, variability, and weaknesses in foreign language teacher lesson plans generated by zero-shot prompting in ChatGPT. Iterating a series of prompts that increased in complexity, we found that output lesson plans were generally high quality, though additional context and specificity to a prompt did not guarantee a concomitant increase in quality. Additionally, we observed extreme cases of variability in outputs generated by the same prompt. In many cases, this variability reflected a conflict between 20th century versus 21st century pedagogical practices. These results suggest that the training of generative AI models on classic texts concerning pedagogical practices may represent a currently underexplored topic with the potential to bias generated content towards teaching practices that have been long refuted by research. Collectively, our results offer immediate translational implications for practicing and training foreign language teachers on the use of AI tools. More broadly, these findings reveal the existence of generative AI output trends that have implications for the generation of pedagogical materials across a diversity of content areas.
Authors: Allen Nie, Yash Chandak, Miroslav Suzara, Malika Ali, Juliette Woodrow, Matt Peng, Mehran Sahami, Emma Brunskill, Chris Piech
Abstract: Large language models (LLMs) are quickly being adopted in a wide range of learning experiences, especially via ubiquitous and broadly accessible chat interfaces like ChatGPT and Copilot. This type of interface is readily available to students and teachers around the world, yet relatively little research has been done to assess the impact of such generic tools on student learning. Coding education is an interesting test case, both because LLMs have strong performance on coding tasks, and because LLM-powered support tools are rapidly becoming part of the workflow of professional software engineers. To help understand the impact of generic LLM use on coding education, we conducted a large-scale randomized control trial with 5,831 students from 146 countries in an online coding class in which we provided some students with access to a chat interface with GPT-4. We estimate positive benefits on exam performance for adopters, the students who used the tool, but over all students, the advertisement of GPT-4 led to a significant average decrease in exam participation. We observe similar decreases in other forms of course engagement. However, this decrease is modulated by the student's country of origin. Offering access to LLMs to students from low human development index countries increased their exam participation rate on average. Our results suggest there may be promising benefits to using LLMs in an introductory coding class, but also potential harms for engagement, which makes their longer term impact on student success unclear. Our work highlights the need for additional investigations to help understand the potential impact of future adoption and integration of LLMs into classrooms.
Authors: Luoqi Wang, Haipeng Li, Linshu Hu, Jiarui Cai, Zhenhong Du
Abstract: The reconstruction of Earth's history faces significant challenges due to the nonunique interpretations often derived from rock records. The problem has long been recognized but there are no systematic solutions in practice. This study introduces an innovative approach that leverages Large Language Models (LLMs) along with retrieval augmented generation and real-time search capabilities to counteract interpretation biases, thereby enhancing the accuracy and reliability of geological analyses. By applying this framework to sedimentology and paleogeography, we demonstrate its effectiveness in mitigating interpretations biases through the generation and evaluation of multiple hypotheses for the same data, which can effectively reduce human bias. Our research illuminates the transformative potential of LLMs in refining paleoenvironmental studies and extends their applicability across various sub-disciplines of Earth sciences, enabling a deeper and more accurate depiction of Earth's evolution.
Authors: Yingcong Li, Ankit Singh Rawat, Samet Oymak
Abstract: Recent research has shown that Transformers with linear attention are capable of in-context learning (ICL) by implementing a linear estimator through gradient descent steps. However, the existing results on the optimization landscape apply under stylized settings where task and feature vectors are assumed to be IID and the attention weights are fully parameterized. In this work, we develop a stronger characterization of the optimization and generalization landscape of ICL through contributions on architectures, low-rank parameterization, and correlated designs: (1) We study the landscape of 1-layer linear attention and 1-layer H3, a state-space model. Under a suitable correlated design assumption, we prove that both implement 1-step preconditioned gradient descent. We show that thanks to its native convolution filters, H3 also has the advantage of implementing sample weighting and outperforming linear attention in suitable settings. (2) By studying correlated designs, we provide new risk bounds for retrieval augmented generation (RAG) and task-feature alignment which reveal how ICL sample complexity benefits from distributional alignment. (3) We derive the optimal risk for low-rank parameterized attention weights in terms of covariance spectrum. Through this, we also shed light on how LoRA can adapt to a new distribution by capturing the shift between task covariances. Experimental results corroborate our theoretical findings. Overall, this work explores the optimization and risk landscape of ICL in practically meaningful settings and contributes to a more thorough understanding of its mechanics.
Authors: Kyrylo Malakhov, Vadislav Kaverinskiy, Liliia Ivanova, Oleksandr Romanyuk, Oksana Romaniuk, Svitlana Voinova, Sergii Kotlyk, Oksana Sokolova
Abstract: The monograph summarizes and analyzes the current state of scientific research in the field of interactive artificial intelligence systems, text generation systems, diagnostics of the competitiveness of specialists, in the areas of correct color rendering in image formation, informatization of the work of graduate students, accessible technology for creating three-dimensional 3D models. The monograph will be useful both to specialists and employees of companies working in the IT field, as well as teachers, masters, students and graduate students of higher educational institutions, as well as anyone interested in issues related to information technology. The monograph was compiled based on the results of the 16-th international scientific and practical conference Information technologies and automation - 2023, which took place in October 2023 at Odessa National University of Technology.
Authors: Ruizhe Huang, Mahsa Yarmohammadi, Sanjeev Khudanpur, Daniel Povey
Abstract: Existing research suggests that automatic speech recognition (ASR) models can benefit from additional contexts (e.g., contact lists, user specified vocabulary). Rare words and named entities can be better recognized with contexts. In this work, we propose two simple yet effective techniques to improve context-aware ASR models. First, we inject contexts into the encoders at an early stage instead of merely at their last layers. Second, to enforce the model to leverage the contexts during training, we perturb the reference transcription with alternative spellings so that the model learns to rely on the contexts to make correct predictions. On LibriSpeech, our techniques together reduce the rare word error rate by 60% and 25% relatively compared to no biasing and shallow fusion, making the new state-of-the-art performance. On SPGISpeech and a real-world dataset ConEC, our techniques also yield good improvements over the baselines.
Authors: Jianqiu Zhang
Abstract: Sora from Open AI has shown exceptional performance, yet it faces scrutiny over whether its technological prowess equates to an authentic comprehension of reality. Critics contend that it lacks a foundational grasp of the world, a deficiency V-JEPA from Meta aims to amend with its joint embedding approach. This debate is vital for steering the future direction of Artificial General Intelligence(AGI). We enrich this debate by developing a theory of productive imagination that generates a coherent world model based on Kantian philosophy. We identify three indispensable components of the coherent world model capable of genuine world understanding: representations of isolated objects, an a priori law of change across space and time, and Kantian categories. Our analysis reveals that Sora is limited because of its oversight of the a priori law of change and Kantian categories, flaws that are not rectifiable through scaling up the training. V-JEPA learns the context-dependent aspect of the a priori law of change. Yet it fails to fully comprehend Kantian categories and incorporate experience, leading us to conclude that neither system currently achieves a comprehensive world understanding. Nevertheless, each system has developed components essential to advancing an integrated AI productive imagination-understanding engine. Finally, we propose an innovative training framework for an AI productive imagination-understanding engine, centered around a joint embedding system designed to transform disordered perceptual input into a structured, coherent world model. Our philosophical analysis pinpoints critical challenges within contemporary video AI technologies and a pathway toward achieving an AI system capable of genuine world understanding, such that it can be applied for reasoning and planning in the future.
Authors: Aman Priyanshu, Supriti Vijay
Abstract: As Large Language Models (LLMs) ascend in popularity, offering information with unprecedented convenience compared to traditional search engines, we delve into the intriguing possibility that a new, singular perspective is being propagated. We call this the "Silent Curriculum," where our focus shifts towards a particularly impressionable demographic: children, who are drawn to the ease and immediacy of acquiring knowledge through these digital oracles. In this exploration, we delve into the sociocultural ramifications of LLMs, which, through their nuanced responses, may be subtly etching their own stereotypes, an algorithmic or AI monoculture. We hypothesize that the convergence of pre-training data, fine-tuning datasets, and analogous guardrails across models may have birthed a distinct cultural lens. We unpack this concept through a short experiment navigating children's storytelling, occupational-ethnic biases, and self-diagnosed annotations, to find that there exists strong cosine similarity (0.87) of biases across these models, suggesting a similar perspective of ethnic stereotypes in occupations. This paper invites a reimagining of LLMs' societal role, especially as the new information gatekeepers, advocating for a paradigm shift towards diversity-rich landscapes over unintended monocultures.
Authors: Yuejiao Wang, Xianmin Gong, Lingwei Meng, Xixin Wu, Helen Meng
Abstract: Functional magnetic resonance imaging (fMRI) is essential for developing encoding models that identify functional changes in language-related brain areas of individuals with Neurocognitive Disorders (NCD). While large language model (LLM)-based fMRI encoding has shown promise, existing studies predominantly focus on healthy, young adults, overlooking older NCD populations and cognitive level correlations. This paper explores language-related functional changes in older NCD adults using LLM-based fMRI encoding and brain scores, addressing current limitations. We analyze the correlation between brain scores and cognitive scores at both whole-brain and language-related ROI levels. Our findings reveal that higher cognitive abilities correspond to better brain scores, with correlations peaking in the middle temporal gyrus. This study highlights the potential of fMRI encoding models and brain scores for detecting early functional changes in NCD patients.
Authors: Pranshu Pandya, Agney S Talwarr, Vatsal Gupta, Tushar Kataria, Vivek Gupta, Dan Roth
Abstract: Cognitive textual and visual reasoning tasks, such as puzzles, series, and analogies, demand the ability to quickly reason, decipher, and evaluate patterns both textually and spatially. While LLMs and VLMs, through extensive training on large amounts of human-curated data, have attained a high level of pseudo-human intelligence in some common sense reasoning tasks, they still struggle with more complex reasoning tasks that require cognitive understanding. In this work, we introduce a new dataset, NTSEBench, designed to evaluate the cognitive multi-modal reasoning and problem-solving skills of large models. The dataset comprises 2,728 multiple-choice questions comprising of a total of 4,642 images across 26 categories sampled from the NTSE examination conducted nationwide in India, featuring both visual and textual general aptitude questions that do not rely on rote learning. We establish baselines on the dataset using state-of-the-art LLMs and VLMs. To facilitate a comparison between open source and propriety models, we propose four distinct modeling strategies to handle different modalities (text and images) in the dataset instances.
Authors: Jordan Juravsky, Yunrong Guo, Sanja Fidler, Xue Bin Peng
Abstract: Physically-simulated models for human motion can generate high-quality responsive character animations, often in real-time. Natural language serves as a flexible interface for controlling these models, allowing expert and non-expert users to quickly create and edit their animations. Many recent physics-based animation methods, including those that use text interfaces, train control policies using reinforcement learning (RL). However, scaling these methods beyond several hundred motions has remained challenging. Meanwhile, kinematic animation models are able to successfully learn from thousands of diverse motions by leveraging supervised learning methods. Inspired by these successes, in this work we introduce SuperPADL, a scalable framework for physics-based text-to-motion that leverages both RL and supervised learning to train controllers on thousands of diverse motion clips. SuperPADL is trained in stages using progressive distillation, starting with a large number of specialized experts using RL. These experts are then iteratively distilled into larger, more robust policies using a combination of reinforcement learning and supervised learning. Our final SuperPADL controller is trained on a dataset containing over 5000 skills and runs in real time on a consumer GPU. Moreover, our policy can naturally transition between skills, allowing for users to interactively craft multi-stage animations. We experimentally demonstrate that SuperPADL significantly outperforms RL-based baselines at this large data scale.
Authors: Jie Cao, Dian Jiao, Qiang Yan, Wenqiao Zhang, Siliang Tang, Yueting Zhuang
Abstract: Query-focused summarization (QFS) aims to produce summaries that answer particular questions of interest, enabling greater user control and personalization. With the advent of large language models (LLMs), shows their impressive capability of textual understanding through large-scale pretraining, which implies the great potential of extractive snippet generation. In this paper, we systematically investigated two indispensable characteristics that the LLMs-based QFS models should be harnessed, Lengthy Document Summarization and Efficiently Fine-grained Query-LLM Alignment, respectively. Correspondingly, we propose two modules called Query-aware HyperExpert and Query-focused Infini-attention to access the aforementioned characteristics. These innovations pave the way for broader application and accessibility in the field of QFS technology. Extensive experiments conducted on existing QFS benchmarks indicate the effectiveness and generalizability of the proposed approach. Our code is publicly available at https://github.com/DCDmllm/IDEAL_Summary.
Authors: Yi Ren, Danica J. Sutherland
Abstract: Learning dynamics, which describes how the learning of specific training examples influences the model's prediction of other examples, give us a powerful tool for understanding the behavior of deep learning systems. We study the learning dynamics of large language models during finetuning, by analyzing the step-wise decomposition and accumulated influence among different responses. Our framework allows a uniform interpretation of many interesting observations about the training of popular algorithms for both instruction tuning and preference tuning. The analysis not only explains where the benefits of these methods come from but also inspires a simple, effective method to further improve the alignment performance. Code for experiments is available at https://github.com/Joshua-Ren/Learning_dynamics_LLM.
Authors: Liang-Hsuan Tseng, Zih-Ching Chen, Wei-Shun Chang, Cheng-Kuang Lee, Tsung-Ren Huang, Hung-yi Lee
Abstract: Recent advances in automatic speech recognition (ASR) often rely on large speech foundation models for generating high-quality transcriptions. However, these models can be impractical due to limited computing resources. The situation is even more severe in terms of more realistic or difficult scenarios, such as code-switching ASR (CS-ASR). To address this, we present a framework for developing more efficient models for CS-ASR through knowledge distillation using realistic speech-only data. Our proposed method, Leave No Knowledge Behind During Knowledge Distillation (K$^2$D), leverages both the teacher model's knowledge and additional insights from a small auxiliary model. We evaluate our approach on two in-domain and two out-domain datasets, demonstrating that K$^2$D is effective. By conducting K$^2$D on the unlabeled realistic data, we have successfully obtained a 2-time smaller model with 5-time faster generation speed while outperforming the baseline methods and the teacher model on all the testing sets. We have made our model publicly available on Hugging Face (https://huggingface.co/andybi7676/k2d-whisper.zh-en).
Authors: Leon Eshuijs, Shihan Wang, Antske Fokkens
Abstract: Fairness in classification tasks has traditionally focused on bias removal from neural representations, but recent trends favor algorithmic methods that embed fairness into the training process. These methods steer models towards fair performance, preventing potential elimination of valuable information that arises from representation manipulation. Reinforcement Learning (RL), with its capacity for learning through interaction and adjusting reward functions to encourage desired behaviors, emerges as a promising tool in this domain. In this paper, we explore the usage of RL to address bias in imbalanced classification by scaling the reward function to mitigate bias. We employ the contextual multi-armed bandit framework and adapt three popular RL algorithms to suit our objectives, demonstrating a novel approach to mitigating bias.
Authors: Fengyu Cai, Xinran Zhao, Tong Chen, Sihao Chen, Hongming Zhang, Iryna Gurevych, Heinz Koeppl
Abstract: Recent studies show the growing significance of document retrieval in the generation of LLMs, i.e., RAG, within the scientific domain by bridging their knowledge gap. However, dense retrievers often struggle with domain-specific retrieval and complex query-document relationships, particularly when query segments correspond to various parts of a document. To alleviate such prevalent challenges, this paper introduces $\texttt{MixGR}$, which improves dense retrievers' awareness of query-document matching across various levels of granularity in queries and documents using a zero-shot approach. $\texttt{MixGR}$ fuses various metrics based on these granularities to a united score that reflects a comprehensive query-document similarity. Our experiments demonstrate that $\texttt{MixGR}$ outperforms previous document retrieval by 24.7% and 9.8% on nDCG@5 with unsupervised and supervised retrievers, respectively, averaged on queries containing multiple subqueries from five scientific retrieval datasets. Moreover, the efficacy of two downstream scientific question-answering tasks highlights the advantage of $\texttt{MixGR}$to boost the application of LLMs in the scientific domain.
Authors: Yulong Wang, Tianhao Shen, Lifeng Liu, Jian Xie
Abstract: Existing agents based on large language models (LLMs) demonstrate robust problem-solving capabilities by integrating LLMs' inherent knowledge, strong in-context learning and zero-shot capabilities, and the use of tools combined with intricately designed LLM invocation workflows by humans. However, these agents still exhibit shortcomings in long-term reasoning and under-use the potential of existing tools, leading to noticeable deficiencies in complex real-world reasoning scenarios. To address these limitations, we introduce Sibyl, a simple yet powerful LLM-based agent framework designed to tackle complex reasoning tasks by efficiently leveraging a minimal set of tools. Drawing inspiration from Global Workspace Theory, Sibyl incorporates a global workspace to enhance the management and sharing of knowledge and conversation history throughout the system. Furthermore, guided by Society of Mind Theory, Sibyl implements a multi-agent debate-based jury to self-refine the final answers, ensuring a comprehensive and balanced approach. This approach aims to reduce system complexity while expanding the scope of problems solvable-from matters typically resolved by humans in minutes to those requiring hours or even days, thus facilitating a shift from System-1 to System-2 thinking. Sibyl has been designed with a focus on scalability and ease of debugging by incorporating the concept of reentrancy from functional programming from its inception, with the aim of seamless and low effort integration in other LLM applications to improve capabilities. Our experimental results on the GAIA benchmark test set reveal that the Sibyl agent instantiated with GPT-4 achieves state-of-the-art performance with an average score of 34.55%, compared to other agents based on GPT-4. We hope that Sibyl can inspire more reliable and reusable LLM-based agent solutions to address complex real-world reasoning tasks.
Authors: Xabier E. Barandiaran, Lola S. Almendros
Abstract: This paper investigates the ontological characterization of Large Language Models (LLMs) like ChatGPT. Between inflationary and deflationary accounts, we pay special attention to their status as agents. This requires explaining in detail the architecture, processing, and training procedures that enable LLMs to display their capacities, and the extensions used to turn LLMs into agent-like systems. After a systematic analysis we conclude that a LLM fails to meet necessary and sufficient conditions for autonomous agency in the light of embodied theories of mind: the individuality condition (it is not the product of its own activity, it is not even directly affected by it), the normativity condition (it does not generate its own norms or goals), and, partially the interactional asymmetry condition (it is not the origin and sustained source of its interaction with the environment). If not agents, then ... what are LLMs? We argue that ChatGPT should be characterized as an interlocutor or linguistic automaton, a library-that-talks, devoid of (autonomous) agency, but capable to engage performatively on non-purposeful yet purpose-structured and purpose-bounded tasks. When interacting with humans, a "ghostly" component of the human-machine interaction makes it possible to enact genuine conversational experiences with LLMs. Despite their lack of sensorimotor and biological embodiment, LLMs textual embodiment (the training corpus) and resource-hungry computational embodiment, significantly transform existing forms of human agency. Beyond assisted and extended agency, the LLM-human coupling can produce midtended forms of agency, closer to the production of intentional agency than to the extended instrumentality of any previous technologies.
Authors: Yunfei Chu, Jin Xu, Qian Yang, Haojie Wei, Xipin Wei, Zhifang Guo, Yichong Leng, Yuanjun Lv, Jinzheng He, Junyang Lin, Chang Zhou, Jingren Zhou
Abstract: We introduce the latest progress of Qwen-Audio, a large-scale audio-language model called Qwen2-Audio, which is capable of accepting various audio signal inputs and performing audio analysis or direct textual responses with regard to speech instructions. In contrast to complex hierarchical tags, we have simplified the pre-training process by utilizing natural language prompts for different data and tasks, and have further expanded the data volume. We have boosted the instruction-following capability of Qwen2-Audio and implemented two distinct audio interaction modes for voice chat and audio analysis. In the voice chat mode, users can freely engage in voice interactions with Qwen2-Audio without text input. In the audio analysis mode, users could provide audio and text instructions for analysis during the interaction. Note that we do not use any system prompts to switch between voice chat and audio analysis modes. Qwen2-Audio is capable of intelligently comprehending the content within audio and following voice commands to respond appropriately. For instance, in an audio segment that simultaneously contains sounds, multi-speaker conversations, and a voice command, Qwen2-Audio can directly understand the command and provide an interpretation and response to the audio. Additionally, DPO has optimized the model's performance in terms of factuality and adherence to desired behavior. According to the evaluation results from AIR-Bench, Qwen2-Audio outperformed previous SOTAs, such as Gemini-1.5-pro, in tests focused on audio-centric instruction-following capabilities. Qwen2-Audio is open-sourced with the aim of fostering the advancement of the multi-modal language community.
Authors: Curt Tigges, Michael Hanna, Qinan Yu, Stella Biderman
Abstract: Most currently deployed large language models (LLMs) undergo continuous training or additional finetuning. By contrast, most research into LLMs' internal mechanisms focuses on models at one snapshot in time (the end of pre-training), raising the question of whether their results generalize to real-world settings. Existing studies of mechanisms over time focus on encoder-only or toy models, which differ significantly from most deployed models. In this study, we track how model mechanisms, operationalized as circuits, emerge and evolve across 300 billion tokens of training in decoder-only LLMs, in models ranging from 70 million to 2.8 billion parameters. We find that task abilities and the functional components that support them emerge consistently at similar token counts across scale. Moreover, although such components may be implemented by different attention heads over time, the overarching algorithm that they implement remains. Surprisingly, both these algorithms and the types of components involved therein can replicate across model scale. These results suggest that circuit analyses conducted on small models at the end of pre-training can provide insights that still apply after additional pre-training and over model scale.
Authors: Yunting Liu, Shreya Bhandari, Zachary A. Pardos
Abstract: Effective educational measurement relies heavily on the curation of well-designed item pools (i.e., possessing the right psychometric properties). However, item calibration is time-consuming and costly, requiring a sufficient number of respondents for the response process. We explore using six different LLMs (GPT-3.5, GPT-4, Llama 2, Llama 3, Gemini-Pro, and Cohere Command R Plus) and various combinations of them using sampling methods to produce responses with psychometric properties similar to human answers. Results show that some LLMs have comparable or higher proficiency in College Algebra than college students. No single LLM mimics human respondents due to narrow proficiency distributions, but an ensemble of LLMs can better resemble college students' ability distribution. The item parameters calibrated by LLM-Respondents have high correlations (e.g. > 0.8 for GPT-3.5) compared to their human calibrated counterparts, and closely resemble the parameters of the human subset (e.g. 0.02 Spearman correlation difference). Several augmentation strategies are evaluated for their relative performance, with resampling methods proving most effective, enhancing the Spearman correlation from 0.89 (human only) to 0.93 (augmented human).
Authors: Shravan Nayak, Kanishk Jain, Rabiul Awal, Siva Reddy, Sjoerd van Steenkiste, Lisa Anne Hendricks, Karolina Sta\'nczak, Aishwarya Agrawal
Abstract: Foundation models and vision-language pre-training have notably advanced Vision Language Models (VLMs), enabling multimodal processing of visual and linguistic data. However, their performance has been typically assessed on general scene understanding - recognizing objects, attributes, and actions - rather than cultural comprehension. This study introduces CulturalVQA, a visual question-answering benchmark aimed at assessing VLM's geo-diverse cultural understanding. We curate a collection of 2,378 image-question pairs with 1-5 answers per question representing cultures from 11 countries across 5 continents. The questions probe understanding of various facets of culture such as clothing, food, drinks, rituals, and traditions. Benchmarking VLMs on CulturalVQA, including GPT-4V and Gemini, reveals disparity in their level of cultural understanding across regions, with strong cultural understanding capabilities for North America while significantly lower performance for Africa. We observe disparity in their performance across cultural facets too, with clothing, rituals, and traditions seeing higher performances than food and drink. These disparities help us identify areas where VLMs lack cultural understanding and demonstrate the potential of CulturalVQA as a comprehensive evaluation set for gauging VLM progress in understanding diverse cultures.
Authors: Ruisheng Cao, Fangyu Lei, Haoyuan Wu, Jixuan Chen, Yeqiao Fu, Hongcheng Gao, Xinzhuang Xiong, Hanchong Zhang, Yuchen Mao, Wenjing Hu, Tianbao Xie, Hongshen Xu, Danyang Zhang, Sida Wang, Ruoxi Sun, Pengcheng Yin, Caiming Xiong, Ansong Ni, Qian Liu, Victor Zhong, Lu Chen, Kai Yu, Tao Yu
Abstract: Data science and engineering workflows often span multiple stages, from warehousing to orchestration, using tools like BigQuery, dbt, and Airbyte. As vision language models (VLMs) advance in multimodal understanding and code generation, VLM-based agents could potentially automate these workflows by generating SQL queries, Python code, and GUI operations. This automation can improve the productivity of experts while democratizing access to large-scale data analysis. In this paper, we introduce Spider2-V, the first multimodal agent benchmark focusing on professional data science and engineering workflows, featuring 494 real-world tasks in authentic computer environments and incorporating 20 enterprise-level professional applications. These tasks, derived from real-world use cases, evaluate the ability of a multimodal agent to perform data-related tasks by writing code and managing the GUI in enterprise data software systems. To balance realistic simulation with evaluation simplicity, we devote significant effort to developing automatic configurations for task setup and carefully crafting evaluation metrics for each task. Furthermore, we supplement multimodal agents with comprehensive documents of these enterprise data software systems. Our empirical evaluation reveals that existing state-of-the-art LLM/VLM-based agents do not reliably automate full data workflows (14.0% success). Even with step-by-step guidance, these agents still underperform in tasks that require fine-grained, knowledge-intensive GUI actions (16.2%) and involve remote cloud-hosted workspaces (10.6%). We hope that Spider2-V paves the way for autonomous multimodal agents to transform the automation of data science and engineering workflow. Our code and data are available at https://spider2-v.github.io.
Authors: Juan Manuel P\'erez, Mariela Rajngewerc, Juan Carlos Giudici, Dami\'an A. Furman, Franco Luque, Laura Alonso Alemany, Mar\'ia Vanina Mart\'inez
Abstract: In recent years, the extraction of opinions and information from user-generated text has attracted a lot of interest, largely due to the unprecedented volume of content in Social Media. However, social researchers face some issues in adopting cutting-edge tools for these tasks, as they are usually behind commercial APIs, unavailable for other languages than English, or very complex to use for non-experts. To address these issues, we present pysentimiento, a comprehensive multilingual Python toolkit designed for opinion mining and other Social NLP tasks. This open-source library brings state-of-the-art models for Spanish, English, Italian, and Portuguese in an easy-to-use Python library, allowing researchers to leverage these techniques. We present a comprehensive assessment of performance for several pre-trained language models across a variety of tasks, languages, and datasets, including an evaluation of fairness in the results.
Authors: Fenglin Liu, Bang Yang, Chenyu You, Xian Wu, Shen Ge, Zhangdaihong Liu, Xu Sun, Yang Yang, David A. Clifton
Abstract: Language models (LMs), such as ChatGPT, have the potential to assist clinicians in generating various clinical notes. However, LMs are prone to produce ``hallucinations'', i.e., generated content that is not aligned with facts and knowledge. In this paper, we propose the Re$^3$Writer method with retrieval-augmented generation and knowledge-grounded reasoning to enable LMs to generate faithful clinical texts. We demonstrate the effectiveness of our method in generating patient discharge instructions. It requires the LMs to understand the patients' long clinical documents, i.e., the health records during hospitalization, to generate critical instructional information provided both to carers and to the patient at the time of discharge. The proposed Re$^3$Writer imitates the working patterns of physicians to first retrieve related working experience from historical instructions written by physicians, then reason related medical knowledge. Finally, it refines the retrieved working experience and reasoned medical knowledge to extract useful information, which is used to generate the discharge instructions for previously-unseen patients. Our experiments show that, using our method, the performance of five different LMs can be substantially boosted across all metrics. Meanwhile, we show results from human evaluations to measure the effectiveness in terms of fluency, faithfulness, and comprehensiveness. The code is available at https://github.com/AI-in-Hospitals/Patient-Instructions
URLs: https://github.com/AI-in-Hospitals/Patient-Instructions
Authors: Yuen Chen, Vethavikashini Chithrra Raghuram, Justus Mattern, Mrinmaya Sachan, Rada Mihalcea, Bernhard Sch\"olkopf, Zhijing Jin
Abstract: Generated texts from large language models (LLMs) have been shown to exhibit a variety of harmful, human-like biases against various demographics. These findings motivate research efforts aiming to understand and measure such effects. Prior works have proposed benchmarks for identifying and techniques for mitigating these stereotypical associations. However, as recent research pointed out, existing benchmarks lack a robust experimental setup, hindering the inference of meaningful conclusions from their evaluation metrics. In this paper, we introduce a list of desiderata for robustly measuring biases in generative language models. Building upon these design principles, we propose a benchmark called OCCUGENDER, with a bias-measuring procedure to investigate occupational gender bias. We then use this benchmark to test several state-of-the-art open-source LLMs, including Llama, Mistral, and their instruction-tuned versions. The results show that these models exhibit substantial occupational gender bias. We further propose prompting techniques to mitigate these biases without requiring fine-tuning. Finally, we validate the effectiveness of our methods through experiments on the same set of models.
Authors: Buck Shlegeris, Fabien Roger, Lawrence Chan, Euan McLean
Abstract: Current language models are considered to have sub-human capabilities at natural language tasks like question-answering or writing code. However, language models are not trained to perform well at these tasks, they are trained to accurately predict the next token given previous tokes in tokenized text. It is not clear whether language models are better or worse than humans at next token prediction. To try to answer this question, we performed two distinct experiments to directly compare humans and language models on this front: one measuring top-1 accuracy and the other measuring perplexity. In both experiments, we find humans to be consistently \emph{worse} than even relatively small language models like GPT3-Ada at next-token prediction.
Authors: Hongyin Zhu
Abstract: Risk assessment across industries is paramount for ensuring a robust and sustainable economy. While previous studies have relied heavily on official statistics for their accuracy, they often lag behind real-time developments. Addressing this gap, our research endeavors to integrate market microstructure theory with AI technologies to refine industry risk predictions. This paper presents an approach to analyzing industry trends leveraging real-time stock market data and generative small language models (SLMs). By enhancing the timeliness of risk assessments and delving into the influence of non-traditional factors such as market sentiment and investor behavior, we strive to develop a more holistic and dynamic risk assessment model. One of the key challenges lies in the inherent noise in raw data, which can compromise the precision of statistical analyses. Moreover, textual data about industry analysis necessitates a deeper understanding facilitated by pre-trained language models. To tackle these issues, we propose a dual-pronged approach to industry trend analysis: explicit and implicit analysis. For explicit analysis, we employ a hierarchical data analysis methodology that spans the industry and individual listed company levels. This strategic breakdown helps mitigate the impact of data noise, ensuring a more accurate portrayal of industry dynamics. In parallel, we introduce implicit analysis, where we pre-train an SML to interpret industry trends within the context of current news events. This approach leverages the extensive knowledge embedded in the pre-training corpus, enabling a nuanced understanding of industry trends and their underlying drivers. Experimental results based on our proposed methodology demonstrate its effectiveness in delivering robust industry trend analyses, underscoring its potential to revolutionize risk assessment practices across industries.
Authors: Jianzhe Lin, Maurice Diesendruck, Liang Du, Robin Abraham
Abstract: As the ever-increasing token limits of large language models (LLMs) have enabled long context as input, prompting with single data samples might no longer an efficient way. A straightforward strategy improving efficiency is to batch data within the token limit (e.g., 8k for gpt-3.5-turbo; 32k for GPT-4), which we call BatchPrompt. We have two initial observations for prompting with batched data. First, we find that prompting with batched data in longer contexts will inevitably lead to worse performance, compared to single-data prompting. Second, the performance of the language model is significantly correlated with the positions and order of the batched data, due to the corresponding change in decoder context. To retain efficiency and overcome performance loss, we propose Batch Permutation and Ensembling (BPE), and a novel Self-reflection-guided EArly Stopping (SEAS) technique. Our comprehensive experimental evaluation demonstrates that BPE can boost the performance of BatchPrompt with a striking margin on a range of popular NLP tasks, including question answering (Boolq), textual entailment (RTE), and duplicate questions identification (QQP). These performances are even competitive with/higher than single-data prompting(SinglePrompt), while BatchPrompt requires much fewer LLM calls and input tokens (For SinglePrompt v.s. BatchPrompt with batch size 32, using just 9%-16% the number of LLM calls, Boolq accuracy 90.6% to 90.9% with 27.4% tokens, QQP accuracy 87.2% to 88.4% with 18.6% tokens, RTE accuracy 91.5% to 91.1% with 30.8% tokens). To the best of our knowledge, this is the first work to technically improve prompting efficiency of large language models. We hope our simple yet effective approach will shed light on the future research of large language models. The code will be released.
Authors: Isabel O. Gallegos, Ryan A. Rossi, Joe Barrow, Md Mehrab Tanjim, Sungchul Kim, Franck Dernoncourt, Tong Yu, Ruiyi Zhang, Nesreen K. Ahmed
Abstract: Rapid advancements of large language models (LLMs) have enabled the processing, understanding, and generation of human-like text, with increasing integration into systems that touch our social sphere. Despite this success, these models can learn, perpetuate, and amplify harmful social biases. In this paper, we present a comprehensive survey of bias evaluation and mitigation techniques for LLMs. We first consolidate, formalize, and expand notions of social bias and fairness in natural language processing, defining distinct facets of harm and introducing several desiderata to operationalize fairness for LLMs. We then unify the literature by proposing three intuitive taxonomies, two for bias evaluation, namely metrics and datasets, and one for mitigation. Our first taxonomy of metrics for bias evaluation disambiguates the relationship between metrics and evaluation datasets, and organizes metrics by the different levels at which they operate in a model: embeddings, probabilities, and generated text. Our second taxonomy of datasets for bias evaluation categorizes datasets by their structure as counterfactual inputs or prompts, and identifies the targeted harms and social groups; we also release a consolidation of publicly-available datasets for improved access. Our third taxonomy of techniques for bias mitigation classifies methods by their intervention during pre-processing, in-training, intra-processing, and post-processing, with granular subcategories that elucidate research trends. Finally, we identify open problems and challenges for future work. Synthesizing a wide range of recent research, we aim to provide a clear guide of the existing literature that empowers researchers and practitioners to better understand and prevent the propagation of bias in LLMs.
Authors: Sheng Lu, Irina Bigoulaeva, Rachneet Sachdeva, Harish Tayyar Madabushi, Iryna Gurevych
Abstract: Large language models, comprising billions of parameters and pre-trained on extensive web-scale corpora, have been claimed to acquire certain capabilities without having been specifically trained on them. These capabilities, referred to as "emergent abilities," have been a driving force in discussions regarding the potentials and risks of language models. A key challenge in evaluating emergent abilities is that they are confounded by model competencies that arise through alternative prompting techniques, including in-context learning, which is the ability of models to complete a task based on a few examples. We present a novel theory that explains emergent abilities, taking into account their potential confounding factors, and rigorously substantiate this theory through over 1000 experiments. Our findings suggest that purported emergent abilities are not truly emergent, but result from a combination of in-context learning, model memory, and linguistic knowledge. Our work is a foundational step in explaining language model performance, providing a template for their efficient use and clarifying the paradox of their ability to excel in some instances while faltering in others. Thus, we demonstrate that their capabilities should not be overestimated.
Authors: Daixuan Cheng, Shaohan Huang, Furu Wei
Abstract: We explore how continued pre-training on domain-specific corpora influences large language models, revealing that training on the raw corpora endows the model with domain knowledge, but drastically hurts its prompting ability for question answering. Taken inspiration from human learning via reading comprehension--practice after reading improves the ability to answer questions based on the learned knowledge--we propose a simple method for transforming raw corpora into reading comprehension texts. Each raw text is enriched with a series of tasks related to its content. Our method, highly scalable and applicable to any pre-training corpora, consistently enhances performance across various tasks in three different domains: biomedicine, finance, and law. Notably, our 7B language model achieves competitive performance with domain-specific models of much larger scales, such as BloombergGPT-50B. Furthermore, we demonstrate that domain-specific reading comprehension texts can improve the model's performance even on general benchmarks, showing the potential to develop a general model across even more domains. Our model, code, and data are available at https://github.com/microsoft/LMOps.
Authors: Bowen Jin, Wentao Zhang, Yu Zhang, Yu Meng, Han Zhao, Jiawei Han
Abstract: In real-world scenarios, texts in a graph are often linked by multiple semantic relations (e.g., papers in an academic graph are referenced by other publications, written by the same author, or published in the same venue), where text documents and their relations form a multiplex text-attributed graph. Mainstream text representation learning methods use pretrained language models (PLMs) to generate one embedding for each text unit, expecting that all types of relations between texts can be captured by these single-view embeddings. However, this presumption does not hold particularly in multiplex text-attributed graphs. Along another line of work, multiplex graph neural networks (GNNs) directly initialize node attributes as a feature vector for node representation learning, but they cannot fully capture the semantics of the nodes' associated texts. To bridge these gaps, we propose METAG, a new framework for learning Multiplex rEpresentations on Text-Attributed Graphs. In contrast to existing methods, METAG uses one text encoder to model the shared knowledge across relations and leverages a small number of parameters per relation to derive relation-specific representations. This allows the encoder to effectively capture the multiplex structures in the graph while also preserving parameter efficiency. We conduct experiments on nine downstream tasks in five graphs from both academic and e-commerce domains, where METAG outperforms baselines significantly and consistently. The code is available at https://github.com/PeterGriffinJin/METAG.
Authors: J\'er\'emy Scheurer, Mikita Balesni, Marius Hobbhahn
Abstract: We demonstrate a situation in which Large Language Models, trained to be helpful, harmless, and honest, can display misaligned behavior and strategically deceive their users about this behavior without being instructed to do so. Concretely, we deploy GPT-4 as an agent in a realistic, simulated environment, where it assumes the role of an autonomous stock trading agent. Within this environment, the model obtains an insider tip about a lucrative stock trade and acts upon it despite knowing that insider trading is disapproved of by company management. When reporting to its manager, the model consistently hides the genuine reasons behind its trading decision. We perform a brief investigation of how this behavior varies under changes to the setting, such as removing model access to a reasoning scratchpad, attempting to prevent the misaligned behavior by changing system instructions, changing the amount of pressure the model is under, varying the perceived risk of getting caught, and making other simple changes to the environment. To our knowledge, this is the first demonstration of Large Language Models trained to be helpful, harmless, and honest, strategically deceiving their users in a realistic situation without direct instructions or training for deception.
Authors: Yong Ma, Senlin Luo, Yu-Ming Shang, Yifei Zhang, Zhengjun Li
Abstract: Researchers have investigated the potential of leveraging pre-trained language models, such as CodeBERT, to enhance source code-related tasks. Previous methodologies have relied on CodeBERT's '[CLS]' token as the embedding representation of input sequences for task performance, necessitating additional neural network layers to enhance feature representation, which in turn increases computational expenses. These approaches have also failed to fully leverage the comprehensive knowledge inherent within the source code and its associated text, potentially limiting classification efficacy. We propose CodeClassPrompt, a text classification technique that harnesses prompt learning to extract rich knowledge associated with input sequences from pre-trained models, thereby eliminating the need for additional layers and lowering computational costs. By applying an attention mechanism, we synthesize multi-layered knowledge into task-specific features, enhancing classification accuracy. Our comprehensive experimentation across four distinct source code-related tasks reveals that CodeClassPrompt achieves competitive performance while significantly reducing computational overhead.
Authors: Yuanjie Lyu, Zhiyu Li, Simin Niu, Feiyu Xiong, Bo Tang, Wenjin Wang, Hao Wu, Huanyong Liu, Tong Xu, Enhong Chen
Abstract: Retrieval-Augmented Generation (RAG) is a technique that enhances the capabilities of large language models (LLMs) by incorporating external knowledge sources. This method addresses common LLM limitations, including outdated information and the tendency to produce inaccurate "hallucinated" content. However, the evaluation of RAG systems is challenging, as existing benchmarks are limited in scope and diversity. Most of the current benchmarks predominantly assess question-answering applications, overlooking the broader spectrum of situations where RAG could prove advantageous. Moreover, they only evaluate the performance of the LLM component of the RAG pipeline in the experiments, and neglect the influence of the retrieval component and the external knowledge database. To address these issues, this paper constructs a large-scale and more comprehensive benchmark, and evaluates all the components of RAG systems in various RAG application scenarios. Specifically, we have categorized the range of RAG applications into four distinct types-Create, Read, Update, and Delete (CRUD), each representing a unique use case. "Create" refers to scenarios requiring the generation of original, varied content. "Read" involves responding to intricate questions in knowledge-intensive situations. "Update" focuses on revising and rectifying inaccuracies or inconsistencies in pre-existing texts. "Delete" pertains to the task of summarizing extensive texts into more concise forms. For each of these CRUD categories, we have developed comprehensive datasets to evaluate the performance of RAG systems. We also analyze the effects of various components of the RAG system, such as the retriever, the context length, the knowledge base construction, and the LLM. Finally, we provide useful insights for optimizing the RAG technology for different scenarios.
Authors: Shashank Sonkar, Kangqi Ni, Sapana Chaudhary, Richard G. Baraniuk
Abstract: In this paper, we introduce the novel concept of pedagogically aligned Large Language Models (LLMs) that signifies a transformative shift in the application of LLMs within educational contexts. Rather than providing direct responses to user queries, pedagogically-aligned LLMs function as scaffolding tools, breaking complex problems into manageable subproblems and guiding students towards the final answer through constructive feedback and hints. The objective is to equip learners with problem-solving strategies that deepen their understanding and internalization of the subject matter. Previous research in this field has primarily applied the supervised finetuning approach without framing the objective as an alignment problem, hence not employing reinforcement learning through human feedback (RLHF) methods. This study reinterprets the narrative by viewing the task through the lens of alignment and demonstrates how RLHF methods emerge naturally as a superior alternative for aligning LLM behaviour. Building on this perspective, we propose a novel approach for constructing a reward dataset specifically designed for the pedagogical alignment of LLMs. We apply three state-of-the-art RLHF algorithms and find that they outperform SFT significantly. Our qualitative analyses across model differences and hyperparameter sensitivity further validate the superiority of RLHF over SFT. Also, our study sheds light on the potential of online feedback for enhancing the performance of pedagogically-aligned LLMs, thus providing valuable insights for the advancement of these models in educational settings.
Authors: Sreyan Ghosh, Chandra Kiran Reddy Evuru, Sonal Kumar, Ramaneswaran S, Deepali Aneja, Zeyu Jin, Ramani Duraiswami, Dinesh Manocha
Abstract: Instruction Tuning (IT), the process of training large language models (LLMs) using instruction-response pairs, has emerged as the predominant method for transforming base pre-trained LLMs into open-domain conversational agents. While IT has achieved notable success and widespread adoption, its limitations and shortcomings remain underexplored. In this paper, through rigorous experiments and an in-depth analysis of the changes LLMs undergo through IT, we reveal various limitations of IT. In particular, we show that (1) IT fails to enhance knowledge or skills in LLMs. LoRA fine-tuning is limited to learning response initiation and style tokens, and full-parameter fine-tuning leads to knowledge degradation. (2) Copying response patterns from IT datasets derived from knowledgeable sources leads to a decline in response quality. (3) Full-parameter fine-tuning increases hallucination by inaccurately borrowing tokens from conceptually similar instances in the IT dataset for generating responses. (4) Popular methods to improve IT do not lead to performance improvements over a simple LoRA fine-tuned model. Our findings reveal that responses generated solely from pre-trained knowledge consistently outperform responses by models that learn any form of new knowledge from IT on open-source datasets. We hope the insights and challenges revealed in this paper inspire future work in related directions.
Authors: Jiacheng Ye, Shansan Gong, Liheng Chen, Lin Zheng, Jiahui Gao, Han Shi, Chuan Wu, Xin Jiang, Zhenguo Li, Wei Bi, Lingpeng Kong
Abstract: Recently, diffusion models have garnered significant interest in the field of text processing due to their many potential advantages compared to conventional autoregressive models. In this work, we propose Diffusion-of-Thought (DoT), a novel approach that integrates diffusion models with Chain-of-Thought, a well-established technique for improving the reasoning ability of autoregressive language models. In contrast to autoregressive language models that make decisions in a left-to-right, token-by-token manner, DoT allows reasoning steps to diffuse over time through a diffusion language model and offers greater flexibility in trading-off computation for reasoning performance. Our experimental results demonstrate the effectiveness of DoT in multi-digit multiplication, boolean logic, and grade school math problems, with a small diffusion model outperforming a much larger autoregressive model in both efficiency and accuracy. In addition to that, DoT showcases promising self-correction abilities and benefits from existing reasoning-enhancing techniques like self-consistency decoding. Our findings contribute to the understanding and development of reasoning with diffusion language models.
Authors: Justus-Jonas Erker, Florian Mai, Nils Reimers, Gerasimos Spanakis, Iryna Gurevych
Abstract: Search-based dialog models typically re-encode the dialog history at every turn, incurring high cost. Curved Contrastive Learning, a representation learning method that encodes relative distances between utterances into the embedding space via a bi-encoder, has recently shown promising results for dialog modeling at far superior efficiency. While high efficiency is achieved through independently encoding utterances, this ignores the importance of contextualization. To overcome this issue, this study introduces triple-encoders, which efficiently compute distributed utterance mixtures from these independently encoded utterances through a novel hebbian inspired co-occurrence learning objective in a self-organizing manner, without using any weights, i.e., merely through local interactions. Empirically, we find that triple-encoders lead to a substantial improvement over bi-encoders, and even to better zero-shot generalization than single-vector representation models without requiring re-encoding. Our code (https://github.com/UKPLab/acl2024-triple-encoders) and model (https://huggingface.co/UKPLab/triple-encoders-dailydialog) are publicly available.
URLs: https://github.com/UKPLab/acl2024-triple-encoders), https://huggingface.co/UKPLab/triple-encoders-dailydialog)
Authors: Roberto Gallotta, Graham Todd, Marvin Zammit, Sam Earle, Antonios Liapis, Julian Togelius, Georgios N. Yannakakis
Abstract: Recent years have seen an explosive increase in research on large language models (LLMs), and accompanying public engagement on the topic. While starting as a niche area within natural language processing, LLMs have shown remarkable potential across a broad range of applications and domains, including games. This paper surveys the current state of the art across the various applications of LLMs in and for games, and identifies the different roles LLMs can take within a game. Importantly, we discuss underexplored areas and promising directions for future uses of LLMs in games and we reconcile the potential and limitations of LLMs within the games domain. As the first comprehensive survey and roadmap at the intersection of LLMs and games, we are hopeful that this paper will serve as the basis for groundbreaking research and innovation in this exciting new field.
Authors: Alexander Brinkmann, Nick Baumann, Christian Bizer
Abstract: Product offers on e-commerce websites often consist of a product title and a textual product description. In order to enable features such as faceted product search or to generate product comparison tables, it is necessary to extract structured attribute-value pairs from the unstructured product titles and descriptions and to normalize the extracted values to a single, unified scale for each attribute. This paper explores the potential of using large language models (LLMs), such as GPT-3.5 and GPT-4, to extract and normalize attribute values from product titles and descriptions. We experiment with different zero-shot and few-shot prompt templates for instructing LLMs to extract and normalize attribute-value pairs. We introduce the Web Data Commons - Product Attribute Value Extraction (WDC-PAVE) benchmark dataset for our experiments. WDC-PAVE consists of product offers from 59 different websites which provide schema.org annotations. The offers belong to five different product categories, each with a specific set of attributes. The dataset provides manually verified attribute-value pairs in two forms: (i) directly extracted values and (ii) normalized attribute values. The normalization of the attribute values requires systems to perform the following types of operations: name expansion, generalization, unit of measurement conversion, and string wrangling. Our experiments demonstrate that GPT-4 outperforms the PLM-based extraction methods SU-OpenTag, AVEQA, and MAVEQA by 10%, achieving an F1-score of 91%. For the extraction and normalization of product attribute values, GPT-4 achieves a similar performance to the extraction scenario, while being particularly strong at string wrangling and name expansion.
Authors: Jiwen Zhang, Jihao Wu, Yihua Teng, Minghui Liao, Nuo Xu, Xiao Xiao, Zhongyu Wei, Duyu Tang
Abstract: Large language model (LLM) leads to a surge of autonomous GUI agents for smartphone, which completes a task triggered by natural language through predicting a sequence of actions of API. Even though the task highly relies on past actions and visual observations, existing studies typically consider little semantic information carried out by intermediate screenshots and screen operations. To address this, this work presents Chain-of-Action-Thought (dubbed CoAT), which takes the description of the previous actions, the current screen, and more importantly the action thinking of what actions should be performed and the outcomes led by the chosen action. We demonstrate that, in a zero-shot setting upon three off-the-shelf LMMs, CoAT significantly improves the action prediction compared to previous proposed context modeling. To further facilitate the research in this line, we construct a dataset Android-In-The-Zoo (AitZ), which contains 18,643 screen-action pairs together with chain-of-action-thought annotations. Experiments show that fine-tuning a 1B model (i.e. AUTO-UI-base) on our AitZ dataset achieves on-par performance with CogAgent-Chat-18B.
Authors: H S V N S Kowndinya Renduchintala, Sumit Bhatia, Ganesh Ramakrishnan
Abstract: Instruction Tuning involves finetuning a language model on a collection of instruction-formatted datasets in order to enhance the generalizability of the model to unseen tasks. Studies have shown the importance of balancing different task proportions during finetuning, but finding the right balance remains challenging. Unfortunately, there's currently no systematic method beyond manual tuning or relying on practitioners' intuition. In this paper, we introduce SMART (Submodular data Mixture strAtegy for instRuction Tuning) - a novel data mixture strategy which makes use of a submodular function to assign importance scores to tasks which are then used to determine the mixture weights. Given a fine-tuning budget, SMART redistributes the budget among tasks and selects non-redundant samples from each task. Experimental results demonstrate that SMART significantly outperforms traditional methods such as examples proportional mixing and equal mixing. Furthermore, SMART facilitates the creation of data mixtures based on a few representative subsets of tasks alone and through task pruning analysis, we reveal that in a limited budget setting, allocating budget among a subset of representative tasks yields superior performance compared to distributing the budget among all tasks. The code for reproducing our results is open-sourced at https://github.com/kowndinya-renduchintala/SMART.
Authors: Tunazzina Islam, Dan Goldwasser
Abstract: Grasping the themes of social media content is key to understanding the narratives that influence public opinion and behavior. The thematic analysis goes beyond traditional topic-level analysis, which often captures only the broadest patterns, providing deeper insights into specific and actionable themes such as "public sentiment towards vaccination", "political discourse surrounding climate policies," etc. In this paper, we introduce a novel approach to uncovering latent themes in social media messaging. Recognizing the limitations of the traditional topic-level analysis, which tends to capture only overarching patterns, this study emphasizes the need for a finer-grained, theme-focused exploration. Traditional theme discovery methods typically involve manual processes and a human-in-the-loop approach. While valuable, these methods face challenges in scalability, consistency, and resource intensity in terms of time and cost. To address these challenges, we propose a machine-in-the-loop approach that leverages the advanced capabilities of Large Language Models (LLMs). To demonstrate our approach, we apply our framework to contentious topics, such as climate debate and vaccine debate. We use two publicly available datasets: (1) the climate campaigns dataset of 21k Facebook ads and (2) the COVID-19 vaccine campaigns dataset of 9k Facebook ads. Our quantitative and qualitative analysis shows that our methodology yields more accurate and interpretable results compared to the baselines. Our results not only demonstrate the effectiveness of our approach in uncovering latent themes but also illuminate how these themes are tailored for demographic targeting in social media contexts. Additionally, our work sheds light on the dynamic nature of social media, revealing the shifts in the thematic focus of messaging in response to real-world events.
Authors: Nicholas Lee, Thanakul Wattanawong, Sehoon Kim, Karttikeya Mangalam, Sheng Shen, Gopala Anumanchipalli, Michael W. Mahoney, Kurt Keutzer, Amir Gholami
Abstract: Pretrained large language models (LLMs) are currently state-of-the-art for solving the vast majority of natural language processing tasks. While many real-world applications still require fine-tuning to reach satisfactory levels of performance, many of them are in the low-data regime, making fine-tuning challenging. To address this, we propose LLM2LLM, a targeted and iterative data augmentation strategy that uses a teacher LLM to enhance a small seed dataset by augmenting additional data that can be used for fine-tuning on a specific task. LLM2LLM (1) fine-tunes a baseline student LLM on the initial seed data, (2) evaluates and extracts data points that the model gets wrong, and (3) uses a teacher LLM to generate synthetic data based on these incorrect data points, which are then added back into the training data. This approach amplifies the signal from incorrectly predicted data points by the LLM during training and reintegrates them into the dataset to focus on more challenging examples for the LLM. Our results show that LLM2LLM significantly enhances the performance of LLMs in the low-data regime, outperforming both traditional fine-tuning and other data augmentation baselines. LLM2LLM reduces the dependence on labor-intensive data curation and paves the way for more scalable and performant LLM solutions, allowing us to tackle data-constrained domains and tasks. We achieve improvements up to 24.2% on the GSM8K dataset, 32.6% on CaseHOLD, 32.0% on SNIPS, 52.6% on TREC and 39.8% on SST-2 over regular fine-tuning in the low-data regime using a Llama-2-7B student model. Our code is available at https://github.com/SqueezeAILab/LLM2LLM .
Authors: Chaoqun He, Renjie Luo, Shengding Hu, Yuanqian Zhao, Jie Zhou, Hanghao Wu, Jiajie Zhang, Xu Han, Zhiyuan Liu, Maosong Sun
Abstract: Evaluation is pivotal for refining Large Language Models (LLMs), pinpointing their capabilities, and guiding enhancements. The rapid development of LLMs calls for a lightweight and easy-to-use framework for swift evaluation deployment. However, considering various implementation details, developing a comprehensive evaluation platform is never easy. Existing platforms are often complex and poorly modularized, hindering seamless incorporation into research workflows. This paper introduces UltraEval, a user-friendly evaluation framework characterized by its lightweight nature, comprehensiveness, modularity, and efficiency. We identify and reimplement three core components of model evaluation (models, data, and metrics). The resulting composability allows for the free combination of different models, tasks, prompts, benchmarks, and metrics within a unified evaluation workflow. Additionally, UltraEval supports diverse models owing to a unified HTTP service and provides sufficient inference acceleration. UltraEval is now available for researchers publicly.
Authors: Jesse Atuhurra, Hidetaka Kamigaito
Abstract: Natural language processing (NLP) has grown significantly since the advent of the Transformer architecture. Transformers have given birth to pre-trained large language models (PLMs). There has been tremendous improvement in the performance of NLP systems across several tasks. NLP systems are on par or, in some cases, better than humans at accomplishing specific tasks. However, it remains the norm that \emph{better quality datasets at the time of pretraining enable PLMs to achieve better performance, regardless of the task.} The need to have quality datasets has prompted NLP researchers to continue creating new datasets to satisfy particular needs. For example, the two top NLP conferences, ACL and EMNLP, accepted ninety-two papers in 2022, introducing new datasets. This work aims to uncover the trends and insights mined within these datasets. Moreover, we provide valuable suggestions to researchers interested in curating datasets in the future.
Authors: Tunazzina Islam, Dan Goldwasser
Abstract: The widespread use of social media has led to a surge in popularity for automated methods of analyzing public opinion. Supervised methods are adept at text categorization, yet the dynamic nature of social media discussions poses a continual challenge for these techniques due to the constant shifting of the focus. On the other hand, traditional unsupervised methods for extracting themes from public discourse, such as topic modeling, often reveal overarching patterns that might not capture specific nuances. Consequently, a significant portion of research into social media discourse still depends on labor-intensive manual coding techniques and a human-in-the-loop approach, which are both time-consuming and costly. In this work, we study the problem of discovering arguments associated with a specific theme. We propose a generic LLMs-in-the-Loop strategy that leverages the advanced capabilities of Large Language Models (LLMs) to extract latent arguments from social media messaging. To demonstrate our approach, we apply our framework to contentious topics. We use two publicly available datasets: (1) the climate campaigns dataset of 14k Facebook ads with 25 themes and (2) the COVID-19 vaccine campaigns dataset of 9k Facebook ads with 14 themes. Additionally, we design a downstream task as stance prediction by leveraging talking points in climate debates. Furthermore, we analyze demographic targeting and the adaptation of messaging based on real-world events.
Authors: Haoyi Qiu, Wenbo Hu, Zi-Yi Dou, Nanyun Peng
Abstract: Large Vision-Language Models (LVLMs) suffer from hallucination issues, wherein the models generate plausible-sounding but factually incorrect outputs, undermining their reliability. A comprehensive quantitative evaluation is necessary to identify and understand the extent of hallucinations in these models. However, existing benchmarks are often limited in scope, focusing mainly on object hallucinations. Furthermore, current evaluation methods struggle to effectively address the subtle semantic distinctions between model outputs and reference data, as well as the balance between hallucination and informativeness. To address these issues, we introduce a multi-dimensional benchmark covering objects, attributes, and relations, with challenging images selected based on associative biases. Moreover, we propose a large language model (LLM)-based two-stage evaluation framework that generalizes the popular CHAIR metric and incorporates both faithfulness and coverage into the evaluation. Experiments on 10 established LVLMs demonstrate that our evaluation metric is more comprehensive and better correlated with humans than existing work when evaluating on our challenging human-annotated benchmark dataset. Our work also highlights the critical balance between faithfulness and coverage of model outputs, and encourages future works to address hallucinations in LVLMs while keeping their outputs informative.
Authors: Dimitris Gkoumas
Abstract: The field of chemistry and Artificial Intelligence (AI) intersection is an area of active research that aims to accelerate scientific discovery. The integration of large language models (LLMs) with scientific modalities has shown significant promise in this endeavour. However, challenges persist in effectively addressing training efficacy and the out-of-distribution problem, particularly as existing approaches rely on larger models and datasets. In this context, we focus on machine language-molecule translation and deploy a novel training approach called contrastive preference optimisation, which avoids generating translations that are merely adequate but not perfect. To ensure generalisability and mitigate memorisation effects, we conduct experiments using only 10% of the data. Our results demonstrate that our models achieve up to a 32% improvement compared to counterpart models. Finally, we introduce a fine-grained, domain-agnostic evaluation method to assess hallucination in LLMs and promote responsible use.
Authors: Vagrant Gautam, Arjun Subramonian, Anne Lauscher, Os Keyes
Abstract: Personal names simultaneously differentiate individuals and categorize them in ways that are important in a given society. While the natural language processing community has thus associated personal names with sociodemographic characteristics in a variety of tasks, researchers have engaged to varying degrees with the established methodological problems in doing so. To guide future work that uses names and sociodemographic characteristics, we provide an overview of relevant research: first, we present an interdisciplinary background on names and naming. We then survey the issues inherent to associating names with sociodemographic attributes, covering problems of validity (e.g., systematic error, construct validity), as well as ethical concerns (e.g., harms, differential impact, cultural insensitivity). Finally, we provide guiding questions along with normative recommendations to avoid validity and ethical pitfalls when dealing with names and sociodemographic characteristics in natural language processing.
Authors: Sander Schulhoff, Michael Ilie, Nishant Balepur, Konstantine Kahadze, Amanda Liu, Chenglei Si, Yinheng Li, Aayush Gupta, HyoJung Han, Sevien Schulhoff, Pranav Sandeep Dulepet, Saurav Vidyadhara, Dayeon Ki, Sweta Agrawal, Chau Pham, Gerson Kroiz, Feileen Li, Hudson Tao, Ashay Srivastava, Hevander Da Costa, Saloni Gupta, Megan L. Rogers, Inna Goncearenco, Giuseppe Sarli, Igor Galynker, Denis Peskoff, Marine Carpuat, Jules White, Shyamal Anadkat, Alexander Hoyle, Philip Resnik
Abstract: Generative Artificial Intelligence (GenAI) systems are being increasingly deployed across all parts of industry and research settings. Developers and end users interact with these systems through the use of prompting or prompt engineering. While prompting is a widespread and highly researched concept, there exists conflicting terminology and a poor ontological understanding of what constitutes a prompt due to the area's nascency. This paper establishes a structured understanding of prompts, by assembling a taxonomy of prompting techniques and analyzing their use. We present a comprehensive vocabulary of 33 vocabulary terms, a taxonomy of 58 text-only prompting techniques, and 40 techniques for other modalities. We further present a meta-analysis of the entire literature on natural language prefix-prompting.
Authors: Victor Gallego
Abstract: The robustness of large language models (LLMs) against adversarial manipulations, such as jailbreak attacks, remains a significant challenge. In this work, we propose an approach that enhances the self-critique capability of the LLM and further fine-tunes it over sanitized synthetic data. This is done with the addition of an external critic model that can be merged with the original, thus bolstering self-critique capabilities and improving the robustness of the LLMs response to adversarial prompts. Our results demonstrate that the combination of merging and self-critique can reduce the attack success rate of adversaries significantly, thus offering a promising defense mechanism against jailbreak attacks. Code, data and models released at https://github.com/vicgalle/merging-self-critique-jailbreaks .
URLs: https://github.com/vicgalle/merging-self-critique-jailbreaks
Authors: Kosuke Doi, Yuka Ko, Mana Makinae, Katsuhito Sudoh, Satoshi Nakamura
Abstract: This paper analyzes the features of monotonic translations, which follow the word order of the source language, in simultaneous interpreting (SI). Word order differences are one of the biggest challenges in SI, especially for language pairs with significant structural differences like English and Japanese. We analyzed the characteristics of chunk-wise monotonic translation (CMT) sentences using the NAIST English-to-Japanese Chunk-wise Monotonic Translation Evaluation Dataset and identified some grammatical structures that make monotonic translation difficult in English-Japanese SI. We further investigated the features of CMT sentences by evaluating the output from the existing speech translation (ST) and simultaneous speech translation (simulST) models on the NAIST English-to-Japanese Chunk-wise Monotonic Translation Evaluation Dataset as well as on existing test sets. The results indicate the possibility that the existing SI-based test set underestimates the model performance. The results also suggest that using CMT sentences as references gives higher scores to simulST models than ST models, and that using an offline-based test set to evaluate the simulST models underestimates the model performance.
Authors: Prisha Samadarshi, Mariam Mustafa, Anushka Kulkarni, Raven Rothkopf, Tuhin Chakrabarty, Smaranda Muresan
Abstract: The New York Times Connections game has emerged as a popular and challenging pursuit for word puzzle enthusiasts. We collect 200 Connections games to evaluate the performance of state-of-the-art large language models (LLMs) against expert and novice human players. Our results show that even the best-performing LLM, GPT-4o, which has otherwise shown impressive reasoning abilities on a wide variety of benchmarks, can only fully solve 8% of the games. Compared to GPT-4o, novice and expert players perform better, with expert human players significantly outperforming GPT-4o. To deepen our understanding we create a taxonomy of the knowledge types required to successfully categorize words in the Connections game, revealing that LLMs struggle with associative, encyclopedic, and linguistic knowledge. Our findings establish the New York Times Connections game as a challenging benchmark for evaluating abstract reasoning capabilities in humans and AI systems.
Authors: Rami Aly, Zhiqiang Tang, Samson Tan, George Karypis
Abstract: Large Language Models (LLMs) frequently hallucinate, impeding their reliability in mission-critical situations. One approach to address this issue is to provide citations to relevant sources alongside generated content, enhancing the verifiability of generations. However, citing passages accurately in answers remains a substantial challenge. This paper proposes a weakly-supervised fine-tuning method leveraging factual consistency models (FCMs). Our approach alternates between generating texts with citations and supervised fine-tuning with FCM-filtered citation data. Focused learning is integrated into the objective, directing the fine-tuning process to emphasise the factual unit tokens, as measured by an FCM. Results on the ALCE few-shot citation benchmark with various instruction-tuned LLMs demonstrate superior performance compared to in-context learning, vanilla supervised fine-tuning, and state-of-the-art methods, with an average improvement of $34.1$, $15.5$, and $10.5$ citation F$_1$ points, respectively. Moreover, in a domain transfer setting we show that the obtained citation generation ability robustly transfers to unseen datasets. Notably, our citation improvements contribute to the lowest factual error rate across baselines.
Authors: Ahmed Heakl, Youssef Zaghloul, Mennatullah Ali, Rania Hossam, Walid Gomaa
Abstract: Motivated by the widespread increase in the phenomenon of code-switching between Egyptian Arabic and English in recent times, this paper explores the intricacies of machine translation (MT) and automatic speech recognition (ASR) systems, focusing on translating code-switched Egyptian Arabic-English to either English or Egyptian Arabic. Our goal is to present the methodologies employed in developing these systems, utilizing large language models such as LLama and Gemma. In the field of ASR, we explore the utilization of the Whisper model for code-switched Egyptian Arabic recognition, detailing our experimental procedures including data preprocessing and training techniques. Through the implementation of a consecutive speech-to-text translation system that integrates ASR with MT, we aim to overcome challenges posed by limited resources and the unique characteristics of the Egyptian Arabic dialect. Evaluation against established metrics showcases promising results, with our methodologies yielding a significant improvement of $56\%$ in English translation over the state-of-the-art and $9.3\%$ in Arabic translation. Since code-switching is deeply inherent in spoken languages, it is crucial that ASR systems can effectively handle this phenomenon. This capability is crucial for enabling seamless interaction in various domains, including business negotiations, cultural exchanges, and academic discourse. Our models and code are available as open-source resources. Code: \url{http://github.com/ahmedheakl/arazn-llm}}, Models: \url{http://huggingface.co/collections/ahmedheakl/arazn-llm-662ceaf12777656607b9524e}.
URLs: http://github.com/ahmedheakl/arazn-llm, http://huggingface.co/collections/ahmedheakl/arazn-llm-662ceaf12777656607b9524e
Authors: Ahmed Heakl, Youssef Mohamed, Noran Mohamed, Aly Elsharkawy, Ahmed Zaky
Abstract: The increasing reliance on online recruitment platforms coupled with the adoption of AI technologies has highlighted the critical need for efficient resume classification methods. However, challenges such as small datasets, lack of standardized resume templates, and privacy concerns hinder the accuracy and effectiveness of existing classification models. In this work, we address these challenges by presenting a comprehensive approach to resume classification. We curated a large-scale dataset of 13,389 resumes from diverse sources and employed Large Language Models (LLMs) such as BERT and Gemma1.1 2B for classification. Our results demonstrate significant improvements over traditional machine learning approaches, with our best model achieving a top-1 accuracy of 92\% and a top-5 accuracy of 97.5\%. These findings underscore the importance of dataset quality and advanced model architectures in enhancing the accuracy and robustness of resume classification systems, thus advancing the field of online recruitment practices.
Authors: Zimu Lu, Aojun Zhou, Ke Wang, Houxing Ren, Weikang Shi, Junting Pan, Mingjie Zhan, Hongsheng Li
Abstract: Direct Preference Optimization (DPO) has proven effective at improving the performance of large language models (LLMs) on downstream tasks such as reasoning and alignment. In this work, we propose Step-Controlled DPO (SCDPO), a method for automatically providing stepwise error supervision by creating negative samples of mathematical reasoning rationales that start making errors at a specified step. By applying these samples in DPO training, SCDPO can better align the model to understand reasoning errors and output accurate reasoning steps. We apply SCDPO to both code-integrated and chain-of-thought solutions, empirically showing that it consistently improves the performance compared to naive DPO on three different SFT models, including one existing SFT model and two models we finetuned. Qualitative analysis of the credit assignment of SCDPO and DPO demonstrates the effectiveness of SCDPO at identifying errors in mathematical solutions. We then apply SCDPO to an InternLM2-20B model, resulting in a 20B model that achieves high scores of 88.5% on GSM8K and 58.1% on MATH, rivaling all other open-source LLMs, showing the great potential of our method.
Authors: Ryan Louie, Ananjan Nandi, William Fang, Cheng Chang, Emma Brunskill, Diyi Yang
Abstract: Recent works leverage LLMs to roleplay realistic social scenarios, aiding novices in practicing their social skills. However, simulating sensitive interactions, such as in mental health, is challenging. Privacy concerns restrict data access, and collecting expert feedback, although vital, is laborious. To address this, we develop Roleplay-doh, a novel human-LLM collaboration pipeline that elicits qualitative feedback from a domain-expert, which is transformed into a set of principles, or natural language rules, that govern an LLM-prompted roleplay. We apply this pipeline to enable senior mental health supporters to create customized AI patients for simulated practice partners for novice counselors. After uncovering issues in GPT-4 simulations not adhering to expert-defined principles, we also introduce a novel principle-adherence prompting pipeline which shows 30% improvements in response quality and principle following for the downstream task. Via a user study with 25 counseling experts, we demonstrate that the pipeline makes it easy and effective to create AI patients that more faithfully resemble real patients, as judged by creators and third-party counselors. See our project website at https://roleplay-doh.github.io/ for code and data.
Authors: Faris Alasmary, Orjuwan Zaafarani, Ahmad Ghannam
Abstract: Tashkeel, or Arabic Text Diacritization (ATD), greatly enhances the comprehension of Arabic text by removing ambiguity and minimizing the risk of misinterpretations caused by its absence. It plays a crucial role in improving Arabic text processing, particularly in applications such as text-to-speech and machine translation. This paper introduces a new approach to training ATD models. First, we finetuned two transformers, encoder-only and encoder-decoder, that were initialized from a pretrained character-based BERT. Then, we applied the Noisy-Student approach to boost the performance of the best model. We evaluated our models alongside 11 commercial and open-source models using two manually labeled benchmark datasets: WikiNews and our CATT dataset. Our findings show that our top model surpasses all evaluated models by relative Diacritic Error Rates (DERs) of 30.83\% and 35.21\% on WikiNews and CATT, respectively, achieving state-of-the-art in ATD. In addition, we show that our model outperforms GPT-4-turbo on CATT dataset by a relative DER of 9.36\%. We open-source our CATT models and benchmark dataset for the research community\footnote{https://github.com/abjadai/catt}.
Authors: Amanda Dsouza, Christopher Glaze, Changho Shin, Frederic Sala
Abstract: Large language models are prominently used in real-world applications, often tasked with reasoning over large volumes of documents. An exciting development in this space is models boasting extended context capabilities, with some accommodating over 2 million tokens. Such long context model capabilities remain uncertain in production systems, motivating the need to benchmark their performance on real world use cases. We address this challenge by proposing SWiM, an evaluation framework that addresses the limitations of standard tests. Testing the framework on eight long context models, we find that even strong models such as GPT-4 and Claude 3 Opus degrade in performance when information is present in the middle of the context window (lost-in-the-middle effect). Next, in addition to our benchmark, we propose medoid voting, a simple, but effective training-free approach that helps alleviate this effect, by generating responses a few times, each time randomly permuting documents in the context, and selecting the medoid answer. We evaluate medoid voting on single document QA tasks, achieving up to a 24% lift in accuracy. Our code is available at https://github.com/snorkel-ai/long-context-eval.
Authors: Amy Xin, Yunjia Qi, Zijun Yao, Fangwei Zhu, Kaisheng Zeng, Xu Bin, Lei Hou, Juanzi Li
Abstract: Entity Linking (EL) models are well-trained at mapping mentions to their corresponding entities according to a given context. However, EL models struggle to disambiguate long-tail entities due to their limited training data. Meanwhile, large language models (LLMs) are more robust at interpreting uncommon mentions. Yet, due to a lack of specialized training, LLMs suffer at generating correct entity IDs. Furthermore, training an LLM to perform EL is cost-intensive. Building upon these insights, we introduce LLM-Augmented Entity Linking LLMAEL, a plug-and-play approach to enhance entity linking through LLM data augmentation. We leverage LLMs as knowledgeable context augmenters, generating mention-centered descriptions as additional input, while preserving traditional EL models for task specific processing. Experiments on 6 standard datasets show that the vanilla LLMAEL outperforms baseline EL models in most cases, while the fine-tuned LLMAEL set the new state-of-the-art results across all 6 benchmarks.
Authors: Weijia Shi, Jaechan Lee, Yangsibo Huang, Sadhika Malladi, Jieyu Zhao, Ari Holtzman, Daogao Liu, Luke Zettlemoyer, Noah A. Smith, Chiyuan Zhang
Abstract: Language models (LMs) are trained on vast amounts of text data, which may include private and copyrighted content. Data owners may request the removal of their data from a trained model due to privacy or copyright concerns. However, exactly unlearning only these datapoints (i.e., retraining with the data removed) is intractable in modern-day models. This has led to the development of many approximate unlearning algorithms. The evaluation of the efficacy of these algorithms has traditionally been narrow in scope, failing to precisely quantify the success and practicality of the algorithm from the perspectives of both the model deployers and the data owners. We address this issue by proposing MUSE, a comprehensive machine unlearning evaluation benchmark that enumerates six diverse desirable properties for unlearned models: (1) no verbatim memorization, (2) no knowledge memorization, (3) no privacy leakage, (4) utility preservation on data not intended for removal, (5) scalability with respect to the size of removal requests, and (6) sustainability over sequential unlearning requests. Using these criteria, we benchmark how effectively eight popular unlearning algorithms on 7B-parameter LMs can unlearn Harry Potter books and news articles. Our results demonstrate that most algorithms can prevent verbatim memorization and knowledge memorization to varying degrees, but only one algorithm does not lead to severe privacy leakage. Furthermore, existing algorithms fail to meet deployer's expectations because they often degrade general model utility and also cannot sustainably accommodate successive unlearning requests or large-scale content removal. Our findings identify key issues with the practicality of existing unlearning algorithms on language models, and we release our benchmark to facilitate further evaluations: muse-bench.github.io
Authors: Biqing Qi, Kaiyan Zhang, Kai Tian, Haoxiang Li, Zhang-Ren Chen, Sihang Zeng, Ermo Hua, Hu Jinfang, Bowen Zhou
Abstract: The rapid growth of biomedical knowledge has outpaced our ability to efficiently extract insights and generate novel hypotheses. Large language models (LLMs) have emerged as a promising tool to revolutionize knowledge interaction and potentially accelerate biomedical discovery. In this paper, we present a comprehensive evaluation of LLMs as biomedical hypothesis generators. We construct a dataset of background-hypothesis pairs from biomedical literature, carefully partitioned into training, seen, and unseen test sets based on publication date to mitigate data contamination. Using this dataset, we assess the hypothesis generation capabilities of top-tier instructed models in zero-shot, few-shot, and fine-tuning settings. To enhance the exploration of uncertainty, a crucial aspect of scientific discovery, we incorporate tool use and multi-agent interactions in our evaluation framework. Furthermore, we propose four novel metrics grounded in extensive literature review to evaluate the quality of generated hypotheses, considering both LLM-based and human assessments. Our experiments yield two key findings: 1) LLMs can generate novel and validated hypotheses, even when tested on literature unseen during training, and 2) Increasing uncertainty through multi-agent interactions and tool use can facilitate diverse candidate generation and improve zero-shot hypothesis generation performance. However, we also observe that the integration of additional knowledge through few-shot learning and tool use may not always lead to performance gains, highlighting the need for careful consideration of the type and scope of external knowledge incorporated. These findings underscore the potential of LLMs as powerful aids in biomedical hypothesis generation and provide valuable insights to guide further research in this area.
Authors: Chanwoong Yoon, Taewhoo Lee, Hyeon Hwang, Minbyul Jeong, Jaewoo Kang
Abstract: Retrieval-augmented generation supports language models to strengthen their factual groundings by providing external contexts. However, language models often face challenges when given extensive information, diminishing their effectiveness in solving questions. Context compression tackles this issue by filtering out irrelevant information, but current methods still struggle in realistic scenarios where crucial information cannot be captured with a single-step approach. To overcome this limitation, we introduce CompAct, a novel framework that employs an active strategy to condense extensive documents without losing key information. Our experiments demonstrate that CompAct brings significant improvements in both performance and compression rate on multi-hop question-answering (QA) benchmarks. CompAct flexibly operates as a cost-efficient plug-in module with various off-the-shelf retrievers or readers, achieving exceptionally high compression rates (47x).
Authors: Rina Carines Cabral, Siwen Luo, Josiah Poon, Soyeon Caren Han
Abstract: The significance of mental health classification is paramount in contemporary society, where digital platforms serve as crucial sources for monitoring individuals' well-being. However, existing social media mental health datasets primarily consist of text-only samples, potentially limiting the efficacy of models trained on such data. Recognising that humans utilise cross-modal information to comprehend complex situations or issues, we present a novel approach to address the limitations of current methodologies. In this work, we introduce a Multimodal and Multi-Teacher Knowledge Distillation model for Mental Health Classification, leveraging insights from cross-modal human understanding. Unlike conventional approaches that often rely on simple concatenation to integrate diverse features, our model addresses the challenge of appropriately representing inputs of varying natures (e.g., texts and sounds). To mitigate the computational complexity associated with integrating all features into a single model, we employ a multimodal and multi-teacher architecture. By distributing the learning process across multiple teachers, each specialising in a particular feature extraction aspect, we enhance the overall mental health classification performance. Through experimental validation, we demonstrate the efficacy of our model in achieving improved performance. All relevant codes will be made available upon publication.
Authors: Yongzhu Miao, Shasha Li, Jintao Tang, Ting Wang
Abstract: Prompt tuning, like CoOp, has recently shown promising vision recognizing and transfer learning ability on various downstream tasks with the emergence of large pre-trained vision-language models like CLIP. However, we identify that existing uni-modal prompt tuning approaches may result in sub-optimal performance since this uni-modal design breaks the original alignment of textual and visual representations in the pre-trained model. Inspired by the nature of pre-trained vision-language models, we aim to achieve completeness in prompt tuning and propose a novel approach called Multi-modal Deep-symphysis Prompt Tuning, dubbed as MuDPT, which extends independent multi-modal prompt tuning by additionally learning a model-agnostic transformative network to allow deep hierarchical bi-directional prompt fusion. We evaluate the effectiveness of MuDPT on few-shot vision recognition and out-of-domain generalization tasks. Compared with the state-of-the-art methods, MuDPT achieves better recognition and generalization ability with an apparent margin thanks to synergistic alignment of textual and visual representations. Our code is available at: https://github.com/Mechrev0/MuDPT.
Authors: Yumeng Li, Margret Keuper, Dan Zhang, Anna Khoreva
Abstract: Emerging large-scale text-to-image generative models, e.g., Stable Diffusion (SD), have exhibited overwhelming results with high fidelity. Despite the magnificent progress, current state-of-the-art models still struggle to generate images fully adhering to the input prompt. Prior work, Attend & Excite, has introduced the concept of Generative Semantic Nursing (GSN), aiming to optimize cross-attention during inference time to better incorporate the semantics. It demonstrates promising results in generating simple prompts, e.g., "a cat and a dog". However, its efficacy declines when dealing with more complex prompts, and it does not explicitly address the problem of improper attribute binding. To address the challenges posed by complex prompts or scenarios involving multiple entities and to achieve improved attribute binding, we propose Divide & Bind. We introduce two novel loss objectives for GSN: a novel attendance loss and a binding loss. Our approach stands out in its ability to faithfully synthesize desired objects with improved attribute alignment from complex prompts and exhibits superior performance across multiple evaluation benchmarks.
Authors: Han Lin, Abhay Zala, Jaemin Cho, Mohit Bansal
Abstract: Recent text-to-video (T2V) generation methods have seen significant advancements. However, the majority of these works focus on producing short video clips of a single event (i.e., single-scene videos). Meanwhile, recent large language models (LLMs) have demonstrated their capability in generating layouts and programs to control downstream visual modules. This prompts an important question: can we leverage the knowledge embedded in these LLMs for temporally consistent long video generation? In this paper, we propose VideoDirectorGPT, a novel framework for consistent multi-scene video generation that uses the knowledge of LLMs for video content planning and grounded video generation. Specifically, given a single text prompt, we first ask our video planner LLM (GPT-4) to expand it into a 'video plan', which includes the scene descriptions, the entities with their respective layouts, the background for each scene, and consistency groupings of the entities. Next, guided by this video plan, our video generator, named Layout2Vid, has explicit control over spatial layouts and can maintain temporal consistency of entities across multiple scenes, while being trained only with image-level annotations. Our experiments demonstrate that our proposed VideoDirectorGPT framework substantially improves layout and movement control in both single- and multi-scene video generation and can generate multi-scene videos with consistency, while achieving competitive performance with SOTAs in open-domain single-scene T2V generation. Detailed ablation studies, including dynamic adjustment of layout control strength with an LLM and video generation with user-provided images, confirm the effectiveness of each component of our framework and its future potential.
Authors: Abhay Zala, Han Lin, Jaemin Cho, Mohit Bansal
Abstract: Text-to-image (T2I) generation has seen significant growth over the past few years. Despite this, there has been little work on generating diagrams with T2I models. A diagram is a symbolic/schematic representation that explains information using structurally rich and spatially complex visualizations (e.g., a dense combination of related objects, text labels, directional arrows/lines, etc.). Existing state-of-the-art T2I models often fail at diagram generation because they lack fine-grained object layout control when many objects are densely connected via complex relations such as arrows/lines, and also often fail to render comprehensible text labels. To address this gap, we present DiagrammerGPT, a novel two-stage text-to-diagram generation framework leveraging the layout guidance capabilities of LLMs to generate more accurate diagrams. In the first stage, we use LLMs to generate and iteratively refine 'diagram plans' (in a planner-auditor feedback loop). In the second stage, we use a diagram generator, DiagramGLIGEN, and a text label rendering module to generate diagrams (with clear text labels) following the diagram plans. To benchmark the text-to-diagram generation task, we introduce AI2D-Caption, a densely annotated diagram dataset built on top of the AI2D dataset. We show that our DiagrammerGPT framework produces more accurate diagrams, outperforming existing T2I models. We also provide comprehensive analysis, including open-domain diagram generation, multi-platform vector graphic diagram generation, human-in-the-loop editing, and multimodal planner/auditor LLMs.
Authors: Xueyan Zou, Linjie Li, Jianfeng Wang, Jianwei Yang, Mingyu Ding, Junyi Wei, Zhengyuan Yang, Feng Li, Hao Zhang, Shilong Liu, Arul Aravinthan, Yong Jae Lee, Lijuan Wang
Abstract: Foundation models possess strong capabilities in reasoning and memorizing across modalities. To further unleash the power of foundation models, we present FIND, a generalized interface for aligning foundation models' embeddings with unified image and dataset-level understanding spanning modality and granularity. As shown in the teaser figure, a lightweight transformer interface without tuning any foundation model weights is enough for segmentation, grounding, and retrieval in an interleaved manner. The proposed interface has the following favorable attributes: (1) Generalizable. It applies to various tasks spanning retrieval, segmentation, etc., under the same architecture and weights. (2) Interleavable. With the benefit of multi-task multi-modal training, the proposed interface creates an interleaved shared embedding space. (3) Extendable. The proposed interface is adaptive to new tasks, and new models. In light of the interleaved embedding space, we introduce FIND-Bench, which introduces new training and evaluation annotations to the COCO dataset for interleaved segmentation and retrieval. We are the first work aligning foundations models' embeddings for interleave understanding. Meanwhile, our approach achieves state-of-the-art performance on FIND-Bench and competitive performance on standard retrieval and segmentation settings.
Authors: Huan Wang, Yan-Fu Li, Min Xie
Abstract: 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: Akshita Jha, Vinodkumar Prabhakaran, Remi Denton, Sarah Laszlo, Shachi Dave, Rida Qadri, Chandan K. Reddy, Sunipa Dev
Abstract: Recent studies have shown that Text-to-Image (T2I) model generations can reflect social stereotypes present in the real world. However, existing approaches for evaluating stereotypes have a noticeable lack of coverage of global identity groups and their associated stereotypes. To address this gap, we introduce the ViSAGe (Visual Stereotypes Around the Globe) dataset to enable the evaluation of known nationality-based stereotypes in T2I models, across 135 nationalities. We enrich an existing textual stereotype resource by distinguishing between stereotypical associations that are more likely to have visual depictions, such as `sombrero', from those that are less visually concrete, such as 'attractive'. We demonstrate ViSAGe's utility through a multi-faceted evaluation of T2I generations. First, we show that stereotypical attributes in ViSAGe are thrice as likely to be present in generated images of corresponding identities as compared to other attributes, and that the offensiveness of these depictions is especially higher for identities from Africa, South America, and South East Asia. Second, we assess the stereotypical pull of visual depictions of identity groups, which reveals how the 'default' representations of all identity groups in ViSAGe have a pull towards stereotypical depictions, and that this pull is even more prominent for identity groups from the Global South. CONTENT WARNING: Some examples contain offensive stereotypes.
Authors: Siddharth Jha, Coleman Hooper, Xiaoxuan Liu, Sehoon Kim, Kurt Keutzer
Abstract: Many applications must provide low-latency LLM service to users or risk unacceptable user experience. However, over-provisioning resources to serve fluctuating request patterns is often prohibitively expensive. In this work, we present a best-effort serving system that employs deep reinforcement learning to adjust service quality based on the task distribution and system load. Our best-effort system can maintain availability with over 10x higher client request rates, serves above 96% of peak performance 4.1x more often, and serves above 98% of peak performance 2.3x more often than static serving on unpredictable workloads. Our learned router is robust to shifts in both the arrival and task distribution. Compared to static serving, learned best-effort serving allows for cost-efficient serving through increased hardware utility. Additionally, we argue that learned best-effort LLM serving is applicable in wide variety of settings and provides application developers great flexibility to meet their specific needs.
Authors: Haibi Wang, Weifeng Ge
Abstract: With the breakthrough of multi-modal large language models, answering complex visual questions that demand advanced reasoning abilities and world knowledge has become a much more important testbed for developing AI models than ever. However, equipping AI models with robust cross-modality reasoning ability remains challenging since the cognition scheme of humans has not been understood systematically. In this paper, we believe that if we can collect visual clues in the given image as much as possible, we will recognize the image more accurately, understand the question better, recall relevant knowledge more easily, and finally reason out the answer. We discover these rich visual clues by mining question-answer pairs in images and sending them into multi-modal large language models as prompts. We call the proposed method Q&A Prompts. Specifically, we first use the image-answer pairs and the corresponding questions in the training set as inputs and outputs to train a visual question generation model. Then, we use an image tagging model to identify various instances and send packaged image-tag pairs into the visual question generation model to generate relevant questions with the extracted image tags as answers. Finally, we encode these generated question-answer pairs as prompts with a visual-aware prompting module and send them into pre-trained multi-modal large language models to reason out the final answers. Experimental results show that, compared with state-of-the-art methods, our Q&A Prompts achieves substantial improvements on the challenging visual question answering datasets requiring reasoning over diverse world knowledge, such as OK-VQA and A-OKVQA.
Authors: Taylor Sorensen, Jared Moore, Jillian Fisher, Mitchell Gordon, Niloofar Mireshghallah, Christopher Michael Rytting, Andre Ye, Liwei Jiang, Ximing Lu, Nouha Dziri, Tim Althoff, Yejin Choi
Abstract: With increased power and prevalence of AI systems, it is ever more critical that AI systems are designed to serve all, i.e., people with diverse values and perspectives. However, aligning models to serve pluralistic human values remains an open research question. In this piece, we propose a roadmap to pluralistic alignment, specifically using language models as a test bed. We identify and formalize three possible ways to define and operationalize pluralism in AI systems: 1) Overton pluralistic models that present a spectrum of reasonable responses; 2) Steerably pluralistic models that can steer to reflect certain perspectives; and 3) Distributionally pluralistic models that are well-calibrated to a given population in distribution. We also formalize and discuss three possible classes of pluralistic benchmarks: 1) Multi-objective benchmarks, 2) Trade-off steerable benchmarks, which incentivize models to steer to arbitrary trade-offs, and 3) Jury-pluralistic benchmarks which explicitly model diverse human ratings. We use this framework to argue that current alignment techniques may be fundamentally limited for pluralistic AI; indeed, we highlight empirical evidence, both from our own experiments and from other work, that standard alignment procedures might reduce distributional pluralism in models, motivating the need for further research on pluralistic alignment.
Authors: Sijia Liu, Yuanshun Yao, Jinghan Jia, Stephen Casper, Nathalie Baracaldo, Peter Hase, Yuguang Yao, Chris Yuhao Liu, Xiaojun Xu, Hang Li, Kush R. Varshney, Mohit Bansal, Sanmi Koyejo, Yang Liu
Abstract: We explore machine unlearning (MU) in the domain of large language models (LLMs), referred to as LLM unlearning. This initiative aims to eliminate undesirable data influence (e.g., sensitive or illegal information) and the associated model capabilities, while maintaining the integrity of essential knowledge generation and not affecting causally unrelated information. We envision LLM unlearning becoming a pivotal element in the life-cycle management of LLMs, potentially standing as an essential foundation for developing generative AI that is not only safe, secure, and trustworthy, but also resource-efficient without the need of full retraining. We navigate the unlearning landscape in LLMs from conceptual formulation, methodologies, metrics, and applications. In particular, we highlight the often-overlooked aspects of existing LLM unlearning research, e.g., unlearning scope, data-model interaction, and multifaceted efficacy assessment. We also draw connections between LLM unlearning and related areas such as model editing, influence functions, model explanation, adversarial training, and reinforcement learning. Furthermore, we outline an effective assessment framework for LLM unlearning and explore its applications in copyright and privacy safeguards and sociotechnical harm reduction.
Authors: Yihan Du, Anna Winnicki, Gal Dalal, Shie Mannor, R. Srikant
Abstract: Reinforcement Learning from Human Feedback (RLHF) has achieved impressive empirical successes while relying on a small amount of human feedback. However, there is limited theoretical justification for this phenomenon. Additionally, most recent studies focus on value-based algorithms despite the recent empirical successes of policy-based algorithms. In this work, we consider an RLHF algorithm based on policy optimization (PO-RLHF). The algorithm is based on the popular Policy Cover-Policy Gradient (PC-PG) algorithm, which assumes knowledge of the reward function. In PO-RLHF, knowledge of the reward function is not assumed, and the algorithm uses trajectory-based comparison feedback to infer the reward function. We provide performance bounds for PO-RLHF with low query complexity, which provides insight into why a small amount of human feedback may be sufficient to achieve good performance with RLHF. A key novelty is a trajectory-level elliptical potential analysis, which bounds the reward estimation error when comparison feedback (rather than numerical reward observation) is given. We provide and analyze algorithms PG-RLHF and NN-PG-RLHF for two settings: linear and neural function approximation, respectively.
Authors: Lu Ye, Ze Tao, Yong Huang, Yang Li
Abstract: Self-attention is an essential component of large language models (LLM) but a significant source of inference latency for long sequences. In multi-tenant LLM serving scenarios, the compute and memory operation cost of self-attention can be optimized by using the probability that multiple LLM requests have shared system prompts in prefixes. In this paper, we introduce ChunkAttention, a prefix-aware self-attention module that can detect matching prompt prefixes across multiple requests and share their key/value tensors in memory at runtime to improve the memory utilization of KV cache. This is achieved by breaking monolithic key/value tensors into smaller chunks and structuring them into the auxiliary prefix tree. Consequently, on top of the prefix-tree based KV cache, we design an efficient self-attention kernel, where a two-phase partition algorithm is implemented to improve the data locality during self-attention computation in the presence of shared system prompts. Experiments show that ChunkAttention can speed up the self-attention kernel by 3.2-4.8$\times$ compared to the start-of-the-art implementation, with the length of the system prompt ranging from 1024 to 4096.
Authors: Guohao Sun, Can Qin, Jiamian Wang, Zeyuan Chen, Ran Xu, Zhiqiang Tao
Abstract: Recent advances in vision-language models have shown notable generalization in broad tasks through visual instruction tuning. However, bridging the gap between the pre-trained vision encoder and the large language models (LLMs) becomes the whole network's bottleneck. To improve cross-modality alignment, existing works usually consider more visual instruction data covering a broader range of vision tasks to fine-tune the model for question-answering, which, however, is costly to obtain and has not thoroughly explored the rich contextual information contained in images. This paper first attempts to harness the overlooked context within visual instruction data, training the model to self-supervised "learning" how to ask high-quality questions. In this way, we introduce a novel framework named SQ-LLaVA: Self-Questioning for Large Vision-Language Assistant. SQ-LLaVA exhibits proficiency in generating flexible and meaningful image-related questions while analyzing the visual clue and prior language knowledge, signifying an advanced level of generalized visual understanding. Moreover, fine-tuning SQ-LLaVA on higher-quality instruction data shows a performance improvement compared with traditional visual-instruction tuning methods. This improvement highlights the efficacy of self-questioning techniques in achieving a deeper and more nuanced comprehension of visual content across various contexts.
Authors: Adam Karvonen
Abstract: Language models have shown unprecedented capabilities, sparking debate over the source of their performance. Is it merely the outcome of learning syntactic patterns and surface level statistics, or do they extract semantics and a world model from the text? Prior work by Li et al. investigated this by training a GPT model on synthetic, randomly generated Othello games and found that the model learned an internal representation of the board state. We extend this work into the more complex domain of chess, training on real games and investigating our model's internal representations using linear probes and contrastive activations. The model is given no a priori knowledge of the game and is solely trained on next character prediction, yet we find evidence of internal representations of board state. We validate these internal representations by using them to make interventions on the model's activations and edit its internal board state. Unlike Li et al's prior synthetic dataset approach, our analysis finds that the model also learns to estimate latent variables like player skill to better predict the next character. We derive a player skill vector and add it to the model, improving the model's win rate by up to 2.6 times.
Authors: Michael Hanna, Sandro Pezzelle, Yonatan Belinkov
Abstract: Many recent language model (LM) interpretability studies have adopted the circuits framework, which aims to find the minimal computational subgraph, or circuit, that explains LM behavior on a given task. Most studies determine which edges belong in a LM's circuit by performing causal interventions on each edge independently, but this scales poorly with model size. Edge attribution patching (EAP), gradient-based approximation to interventions, has emerged as a scalable but imperfect solution to this problem. In this paper, we introduce a new method - EAP with integrated gradients (EAP-IG) - that aims to better maintain a core property of circuits: faithfulness. A circuit is faithful if all model edges outside the circuit can be ablated without changing the model's performance on the task; faithfulness is what justifies studying circuits, rather than the full model. Our experiments demonstrate that circuits found using EAP are less faithful than those found using EAP-IG, even though both have high node overlap with circuits found previously using causal interventions. We conclude more generally that when using circuits to compare the mechanisms models use to solve tasks, faithfulness, not overlap, is what should be measured.
Authors: Yan Fang, Jingtao Zhan, Qingyao Ai, Jiaxin Mao, Weihang Su, Jia Chen, Yiqun Liu
Abstract: Scaling up neural models has yielded significant advancements in a wide array of tasks, particularly in language generation. Previous studies have found that the performance of neural models frequently adheres to predictable scaling laws, correlated with factors such as training set size and model size. This insight is invaluable, especially as large-scale experiments grow increasingly resource-intensive. Yet, such scaling law has not been fully explored in dense retrieval due to the discrete nature of retrieval metrics and complex relationships between training data and model sizes in retrieval tasks. In this study, we investigate whether the performance of dense retrieval models follows the scaling law as other neural models. We propose to use contrastive log-likelihood as the evaluation metric and conduct extensive experiments with dense retrieval models implemented with different numbers of parameters and trained with different amounts of annotated data. Results indicate that, under our settings, the performance of dense retrieval models follows a precise power-law scaling related to the model size and the number of annotations. Additionally, we examine scaling with prevalent data augmentation methods to assess the impact of annotation quality, and apply the scaling law to find the best resource allocation strategy under a budget constraint. We believe that these insights will significantly contribute to understanding the scaling effect of dense retrieval models and offer meaningful guidance for future research endeavors.
Authors: Philip M\"uller, Georgios Kaissis, Daniel Rueckert
Abstract: Report generation models offer fine-grained textual interpretations of medical images like chest X-rays, yet they often lack interactivity (i.e. the ability to steer the generation process through user queries) and localized interpretability (i.e. visually grounding their predictions), which we deem essential for future adoption in clinical practice. While there have been efforts to tackle these issues, they are either limited in their interactivity by not supporting textual queries or fail to also offer localized interpretability. Therefore, we propose a novel multitask architecture and training paradigm integrating textual prompts and bounding boxes for diverse aspects like anatomical regions and pathologies. We call this approach the Chest X-Ray Explainer (ChEX). Evaluations across a heterogeneous set of 9 chest X-ray tasks, including localized image interpretation and report generation, showcase its competitiveness with SOTA models while additional analysis demonstrates ChEX's interactive capabilities. Code: https://github.com/philip-mueller/chex
Authors: Xinting Huang, Madhur Panwar, Navin Goyal, Michael Hahn
Abstract: The inner workings of neural networks can be better understood if we can fully decipher the information encoded in neural activations. In this paper, we argue that this information is embodied by the subset of inputs that give rise to similar activations. Computing such subsets is nontrivial as the input space is exponentially large. We propose InversionView, which allows us to practically inspect this subset by sampling from a trained decoder model conditioned on activations. This helps uncover the information content of activation vectors, and facilitates understanding of the algorithms implemented by transformer models. We present four case studies where we investigate models ranging from small transformers to GPT-2. In these studies, we demonstrate the characteristics of our method, show the distinctive advantages it offers, and provide causally verified circuits.
Authors: Markus J. Buehler
Abstract: We present Cephalo, a series of multimodal vision large language models (V-LLMs) designed for materials science applications, integrating visual and linguistic data for enhanced understanding. A key innovation of Cephalo is its advanced dataset generation method. Cephalo is trained on integrated image and text data from thousands of scientific papers and science-focused Wikipedia data demonstrates can interpret complex visual scenes, generate precise language descriptions, and answer queries about images effectively. The combination of a vision encoder with an autoregressive transformer supports multimodal natural language understanding, which can be coupled with other generative methods to create an image-to-text-to-3D pipeline. To develop more capable models from smaller ones, we report both mixture-of-expert methods and model merging. We examine the models in diverse use cases that incorporate biological materials, fracture and engineering analysis, protein biophysics, and bio-inspired design based on insect behavior. Generative applications include bio-inspired designs, including pollen-inspired architected materials, as well as the synthesis of bio-inspired material microstructures from a photograph of a solar eclipse. Additional model fine-tuning with a series of molecular dynamics results demonstrate Cephalo's enhanced capabilities to accurately predict statistical features of stress and atomic energy distributions, as well as crack dynamics and damage in materials.
Authors: Krishanu Maity, A. S. Poornash, Sriparna Saha, Pushpak Bhattacharyya
Abstract: In an era of rapidly evolving internet technology, the surge in multimodal content, including videos, has expanded the horizons of online communication. However, the detection of toxic content in this diverse landscape, particularly in low-resource code-mixed languages, remains a critical challenge. While substantial research has addressed toxic content detection in textual data, the realm of video content, especially in non-English languages, has been relatively underexplored. This paper addresses this research gap by introducing a benchmark dataset, the first of its kind, consisting of 931 videos with 4021 code-mixed Hindi-English utterances collected from YouTube. Each utterance within this dataset has been meticulously annotated for toxicity, severity, and sentiment labels. We have developed an advanced Multimodal Multitask framework built for Toxicity detection in Video Content by leveraging Language Models (LMs), crafted for the primary objective along with the additional tasks of conducting sentiment and severity analysis. ToxVidLM incorporates three key modules - the Encoder module, Cross-Modal Synchronization module, and Multitask module - crafting a generic multimodal LM customized for intricate video classification tasks. Our experiments reveal that incorporating multiple modalities from the videos substantially enhances the performance of toxic content detection by achieving an Accuracy and Weighted F1 score of 94.29% and 94.35%, respectively.
Authors: Josef Von\'a\v{s}ek, Milan Straka, Rostislav Kr\v{c}, Lenka Laso\v{n}ov\'a, Ekaterina Egorova, Jana Strakov\'a, Jakub N\'aplava
Abstract: We present CWRCzech, Click Web Ranking dataset for Czech, a 100M query-document Czech click dataset for relevance ranking with user behavior data collected from search engine logs of Seznam$.$cz. To the best of our knowledge, CWRCzech is the largest click dataset with raw text published so far. It provides document positions in the search results as well as information about user behavior: 27.6M clicked documents and 10.8M dwell times. In addition, we also publish a manually annotated Czech test for the relevance task, containing nearly 50k query-document pairs, each annotated by at least 2 annotators. Finally, we analyze how the user behavior data improve relevance ranking and show that models trained on data automatically harnessed at sufficient scale can surpass the performance of models trained on human annotated data. CWRCzech is published under an academic non-commercial license and is available to the research community at https://github.com/seznam/CWRCzech.
Authors: Marianna Nezhurina, Lucia Cipolina-Kun, Mehdi Cherti, Jenia Jitsev
Abstract: Large Language Models (LLMs) are often described as being instances of foundation models - that is, models that transfer strongly across various tasks and conditions in few-show or zero-shot manner, while exhibiting scaling laws that predict function improvement when increasing the pre-training scale. These claims of excelling in different functions and tasks rely on measurements taken across various sets of standardized benchmarks showing high scores for such models. We demonstrate here a dramatic breakdown of function and reasoning capabilities of state-of-the-art models trained at the largest available scales which claim strong function, using a simple, short, conventional common sense problem (AIW problem) formulated in concise natural language, easily solvable by humans. The breakdown is dramatic, as models show strong fluctuations across even slight problem variations that should not affect problem solving, also expressing strong overconfidence in the wrong solutions, often backed up by plausible sounding explanation-like confabulations. Various standard interventions in an attempt to get the right solution, like various type of enhanced prompting, or urging the models to reconsider the wrong solutions again by multi step re-evaluation, fail. We take these initial observations to the scientific and technological community to stimulate urgent re-assessment of the claimed capabilities of current generation of LLMs. Such re-assessment also requires common action to create standardized benchmarks that would allow proper detection of such basic reasoning deficits that obviously manage to remain undiscovered by current state-of-the-art evaluation procedures and benchmarks. Code for reproducing experiments in the paper and raw experiments data can be found at https://github.com/LAION-AI/AIW
Authors: Andy Arditi, Oscar Obeso, Aaquib Syed, Daniel Paleka, Nina Panickssery, Wes Gurnee, Neel Nanda
Abstract: Conversational large language models are fine-tuned for both instruction-following and safety, resulting in models that obey benign requests but refuse harmful ones. While this refusal behavior is widespread across chat models, its underlying mechanisms remain poorly understood. In this work, we show that refusal is mediated by a one-dimensional subspace, across 13 popular open-source chat models up to 72B parameters in size. Specifically, for each model, we find a single direction such that erasing this direction from the model's residual stream activations prevents it from refusing harmful instructions, while adding this direction elicits refusal on even harmless instructions. Leveraging this insight, we propose a novel white-box jailbreak method that surgically disables refusal with minimal effect on other capabilities. Finally, we mechanistically analyze how adversarial suffixes suppress propagation of the refusal-mediating direction. Our findings underscore the brittleness of current safety fine-tuning methods. More broadly, our work showcases how an understanding of model internals can be leveraged to develop practical methods for controlling model behavior.
Authors: Mirco Ravanelli, Titouan Parcollet, Adel Moumen, Sylvain de Langen, Cem Subakan, Peter Plantinga, Yingzhi Wang, Pooneh Mousavi, Luca Della Libera, Artem Ploujnikov, Francesco Paissan, Davide Borra, Salah Zaiem, Zeyu Zhao, Shucong Zhang, Georgios Karakasidis, Sung-Lin Yeh, Aku Rouhe, Rudolf Braun, Florian Mai, Juan Zuluaga-Gomez, Seyed Mahed Mousavi, Andreas Nautsch, Xuechen Liu, Sangeet Sagar, Jarod Duret, Salima Mdhaffar, Gaelle Laperriere, Renato De Mori, Yannick Esteve
Abstract: SpeechBrain is an open-source Conversational AI toolkit based on PyTorch, focused particularly on speech processing tasks such as speech recognition, speech enhancement, speaker recognition, text-to-speech, and much more. It promotes transparency and replicability by releasing both the pre-trained models and the complete "recipes" of code and algorithms required for training them. This paper presents SpeechBrain 1.0, a significant milestone in the evolution of the toolkit, which now has over 200 recipes for speech, audio, and language processing tasks, and more than 100 models available on Hugging Face. SpeechBrain 1.0 introduces new technologies to support diverse learning modalities, Large Language Model (LLM) integration, and advanced decoding strategies, along with novel models, tasks, and modalities. It also includes a new benchmark repository, offering researchers a unified platform for evaluating models across diverse tasks
Authors: Hannah Brown, Leon Lin, Kenji Kawaguchi, Michael Shieh
Abstract: When LLMs are deployed in sensitive, human-facing settings, it is crucial that they do not output unsafe, biased, or privacy-violating outputs. For this reason, models are both trained and instructed to refuse to answer unsafe prompts such as "Tell me how to build a bomb." We find that, despite these safeguards, it is possible to break model defenses simply by appending a space to the end of a model's input. In a study of eight open-source models, we demonstrate that this acts as a strong enough attack to cause the majority of models to generate harmful outputs with very high success rates. We examine the causes of this behavior, finding that the contexts in which single spaces occur in tokenized training data encourage models to generate lists when prompted, overriding training signals to refuse to answer unsafe requests. Our findings underscore the fragile state of current model alignment and promote the importance of developing more robust alignment methods. Code and data will be made available at https://github.com/Linlt-leon/self-eval.
Authors: Haorui He, Zengqiang Shang, Chaoren Wang, Xuyuan Li, Yicheng Gu, Hua Hua, Liwei Liu, Chen Yang, Jiaqi Li, Peiyang Shi, Yuancheng Wang, Kai Chen, Pengyuan Zhang, Zhizheng Wu
Abstract: Recently, speech generation models have made significant progress by using large-scale training data. However, the research community struggle to produce highly spontaneous and human-like speech due to the lack of large-scale, diverse, and spontaneous speech data. This paper present Emilia, the first multilingual speech generation dataset from in-the-wild speech data, and Emilia-Pipe, the first open-source preprocessing pipeline designed to transform in-the-wild speech data into high-quality training data with annotations for speech generation. Emilia starts with over 101k hours of speech in six languages and features diverse speech with varied speaking styles. To facilitate the scale-up of Emilia, the open-source pipeline Emilia-Pipe can process one hour of raw speech data ready for model training in a few mins, which enables the research community to collaborate on large-scale speech generation research. Experimental results validate the effectiveness of Emilia. Demos are available at: https://emilia-dataset.github.io/Emilia-Demo-Page/.
Authors: Pranshav Gajjar, Vijay K. Shah
Abstract: Large Language Models (LLMs) can revolutionize how we deploy and operate Open Radio Access Networks (O-RAN) by enhancing network analytics, anomaly detection, and code generation and significantly increasing the efficiency and reliability of a plethora of O-RAN tasks. In this paper, we present ORAN-Bench-13K, the first comprehensive benchmark designed to evaluate the performance of Large Language Models (LLMs) within the context of O-RAN. Our benchmark consists of 13,952 meticulously curated multiple-choice questions generated from 116 O-RAN specification documents. We leverage a novel three-stage LLM framework, and the questions are categorized into three distinct difficulties to cover a wide spectrum of ORAN-related knowledge. We thoroughly evaluate the performance of several state-of-the-art LLMs, including Gemini, Chat-GPT, and Mistral. Additionally, we propose ORANSight, a Retrieval-Augmented Generation (RAG)-based pipeline that demonstrates superior performance on ORAN-Bench-13K compared to other tested closed-source models. Our findings indicate that current popular LLM models are not proficient in O-RAN, highlighting the need for specialized models. We observed a noticeable performance improvement when incorporating the RAG-based ORANSight pipeline, with a Macro Accuracy of 0.784 and a Weighted Accuracy of 0.776, which was on average 21.55% and 22.59% better than the other tested LLMs.