Authors: Sherin Muckatira, Vijeta Deshpande, Vladislav Lialin, Anna Rumshisky
Abstract: Large language models can solve new tasks without task-specific fine-tuning. This ability, also known as in-context learning (ICL), is considered an emergent ability and is primarily seen in large language models with billions of parameters. This study investigates if such emergent properties are strictly tied to model size or can be demonstrated by smaller models trained on reduced-scale data. To explore this, we simplify pre-training data and pre-train 36 causal language models with parameters varying from 1 million to 165 million parameters. We show that models trained on this simplified pre-training data demonstrate enhanced zero-shot capabilities across various tasks in simplified language, achieving performance comparable to that of pre-trained models six times larger on unrestricted language. This suggests that downscaling the language allows zero-shot learning capabilities to emerge in models with limited size. Additionally, we find that these smaller models pre-trained on simplified data demonstrate a power law relationship between the evaluation loss and the three scaling factors: compute, dataset size, and model size.
Authors: Gurusha Juneja, Subhabrata Dutta, Tanmoy Chakraborty
Abstract: Despite demonstrating emergent reasoning abilities, Large Language Models (LLMS) often lose track of complex, multi-step reasoning. Existing studies show that providing guidance via decomposing the original question into multiple subproblems elicits more robustness in LLM reasoning -- a decomposer generates the subproblems, and a solver solves each of these subproblems. However, these techniques fail to accommodate coordination between the decomposer and the solver modules (either in a single model or different specialized ones) -- the decomposer does not keep track of the ability of the solver to follow the decomposed reasoning. In this paper, we propose LM2 to address these challenges. LM2 modularizes the decomposition, solution, and verification into three different language models. The decomposer module identifies the key concepts necessary to solve the problem and generates step-by-step subquestions according to the reasoning requirement. The solver model generates the solution to the subproblems that are then checked by the verifier module; depending upon the feedback from the verifier, the reasoning context is constructed using the subproblems and the solutions. These models are trained to coordinate using policy learning. Exhaustive experimentation suggests the superiority of LM2 over existing methods on in- and out-domain reasoning problems, outperforming the best baselines by $8.1\%$ on MATH, $7.71\%$ on JEEBench, and $9.7\%$ on MedQA problems (code available at https://github.com/LCS2-IIITD/Language_Model_Multiplex).
URLs: https://github.com/LCS2-IIITD/Language_Model_Multiplex).
Authors: Nataliia Kholodna, Sahib Julka, Mohammad Khodadadi, Muhammed Nurullah Gumus, Michael Granitzer
Abstract: Low-resource languages face significant barriers in AI development due to limited linguistic resources and expertise for data labeling, rendering them rare and costly. The scarcity of data and the absence of preexisting tools exacerbate these challenges, especially since these languages may not be adequately represented in various NLP datasets. To address this gap, we propose leveraging the potential of LLMs in the active learning loop for data annotation. Initially, we conduct evaluations to assess inter-annotator agreement and consistency, facilitating the selection of a suitable LLM annotator. The chosen annotator is then integrated into a training loop for a classifier using an active learning paradigm, minimizing the amount of queried data required. Empirical evaluations, notably employing GPT-4-Turbo, demonstrate near-state-of-the-art performance with significantly reduced data requirements, as indicated by estimated potential cost savings of at least 42.45 times compared to human annotation. Our proposed solution shows promising potential to substantially reduce both the monetary and computational costs associated with automation in low-resource settings. By bridging the gap between low-resource languages and AI, this approach fosters broader inclusion and shows the potential to enable automation across diverse linguistic landscapes.
Authors: Amanul Haque, Munindar P. Singh
Abstract: We investigate the effectiveness of ChatGPT in extracting norms from contracts. Norms provide a natural way to engineer multiagent systems by capturing how to govern the interactions between two or more autonomous parties. We extract norms of commitment, prohibition, authorization, and power, along with associated norm elements (the parties involved, antecedents, and consequents) from contracts. Our investigation reveals ChatGPT's effectiveness and limitations in norm extraction from contracts. ChatGPT demonstrates promising performance in norm extraction without requiring training or fine-tuning, thus obviating the need for annotated data, which is not generally available in this domain. However, we found some limitations of ChatGPT in extracting these norms that lead to incorrect norm extractions. The limitations include oversight of crucial details, hallucination, incorrect parsing of conjunctions, and empty norm elements. Enhanced norm extraction from contracts can foster the development of more transparent and trustworthy formal agent interaction specifications, thereby contributing to the improvement of multiagent systems.
Authors: David Herel, Tomas Mikolov
Abstract: In various fields of knowledge creation, including science, new ideas often build on pre-existing information. In this work, we explore this concept within the context of language models. Specifically, we explore the potential of self-training models on their own outputs, akin to how humans learn and build on their previous thoughts and actions. While this approach is intuitively appealing, our research reveals its practical limitations. We find that extended self-training of the GPT-2 model leads to a significant degradation in performance, resulting in repetitive and collapsed token output.
Authors: Tobias Schnabel, Jennifer Neville
Abstract: Large language models (LLMs) can now handle longer and more complex inputs, which facilitate the use of more elaborate prompts. However, prompts often require some tuning to improve performance for deployment. Recent work has proposed automatic prompt optimization methods, but as prompt complexity and LLM strength increase, many prompt optimization techniques are no longer sufficient and a new approach is needed to optimize {\em meta prompt programs}. To address this, we introduce SAMMO, a framework for {\em compile-time} optimizations of metaprompt programs, which represent prompts as structured objects that allows for a rich set of transformations that can be searched over during optimization. We show that SAMMO generalizes previous methods and improves the performance of complex prompts on (1) instruction tuning, (2) RAG pipeline tuning, and (3) prompt compression, across several different LLMs. We make all code available open-source at https://github.com/microsoft/sammo .
Authors: Zhewei Sun, Qian Hu, Rahul Gupta, Richard Zemel, Yang Xu
Abstract: Recent advancement in large language models (LLMs) has offered a strong potential for natural language systems to process informal language. A representative form of informal language is slang, used commonly in daily conversations and online social media. To date, slang has not been comprehensively evaluated in LLMs due partly to the absence of a carefully designed and publicly accessible benchmark. Using movie subtitles, we construct a dataset that supports evaluation on a diverse set of tasks pertaining to automatic processing of slang. For both evaluation and finetuning, we show the effectiveness of our dataset on two core applications: 1) slang detection, and 2) identification of regional and historical sources of slang from natural sentences. We also show how our dataset can be used to probe the output distributions of LLMs for interpretive insights. We find that while LLMs such as GPT-4 achieve good performance in a zero-shot setting, smaller BERT-like models finetuned on our dataset achieve comparable performance. Furthermore, we show that our dataset enables finetuning of LLMs such as GPT-3.5 that achieve substantially better performance than strong zero-shot baselines. Our work offers a comprehensive evaluation and a high-quality benchmark on English slang based on the OpenSubtitles corpus, serving both as a publicly accessible resource and a platform for applying tools for informal language processing.
Authors: Sandeep Chataut, Tuyen Do, Bichar Dip Shrestha Gurung, Shiva Aryal, Anup Khanal, Carol Lushbough, Etienne Gnimpieba
Abstract: Keywords play a crucial role in bridging the gap between human understanding and machine processing of textual data. They are essential to data enrichment because they form the basis for detailed annotations that provide a more insightful and in-depth view of the underlying data. Keyword/domain driven term extraction is a pivotal task in natural language processing, facilitating information retrieval, document summarization, and content categorization. This review focuses on keyword extraction methods, emphasizing the use of three major Large Language Models(LLMs): Llama2-7B, GPT-3.5, and Falcon-7B. We employed a custom Python package to interface with these LLMs, simplifying keyword extraction. Our study, utilizing the Inspec and PubMed datasets, evaluates the performance of these models. The Jaccard similarity index was used for assessment, yielding scores of 0.64 (Inspec) and 0.21 (PubMed) for GPT-3.5, 0.40 and 0.17 for Llama2-7B, and 0.23 and 0.12 for Falcon-7B. This paper underlines the role of prompt engineering in LLMs for better keyword extraction and discusses the impact of hallucination in LLMs on result evaluation. It also sheds light on the challenges in using LLMs for keyword extraction, including model complexity, resource demands, and optimization techniques.
Authors: Parham Abed Azad, Hamid Beigy
Abstract: The rapid expansion of texts' volume and diversity presents formidable challenges in multi-domain settings. These challenges are also visible in the Persian name entity recognition (NER) settings. Traditional approaches, either employing a unified model for multiple domains or individual models for each domain, frequently pose significant limitations. Single models often struggle to capture the nuances of diverse domains, while utilizing multiple large models can lead to resource constraints, rendering the training of a model for each domain virtually impractical. Therefore, this paper introduces a novel approach composed of one core model with multiple sets of domain-specific parameters. We utilize techniques such as prompt tuning and adapters, combined with the incorporation of additional layers, to add parameters that we can train for the specific domains. This enables the model to perform comparably to individual models for each domain. Experimental results on different formal and informal datasets show that by employing these added parameters, the proposed model significantly surpasses existing practical models in performance. Remarkably, the proposed model requires only one instance for training and storage, yet achieves outstanding results across all domains, even surpassing the state-of-the-art in some. Moreover, we analyze each adaptation strategy, delineating its strengths, weaknesses, and optimal hyper-parameters for the Persian NER settings. Finally, we introduce a document-based domain detection pipeline tailored for scenarios with unknown text domains, enhancing the adaptability and practicality of this paper in real-world applications.
Authors: Olufunke O. Sarumi, B\'ela Neuendorf, Joan Plepi, Lucie Flek, J\"org Schl\"otterer, Charles Welch
Abstract: Recent trends in natural language processing research and annotation tasks affirm a paradigm shift from the traditional reliance on a single ground truth to a focus on individual perspectives, particularly in subjective tasks. In scenarios where annotation tasks are meant to encompass diversity, models that solely rely on the majority class labels may inadvertently disregard valuable minority perspectives. This oversight could result in the omission of crucial information and, in a broader context, risk disrupting the balance within larger ecosystems. As the landscape of annotator modeling unfolds with diverse representation techniques, it becomes imperative to investigate their effectiveness with the fine-grained features of the datasets in view. This study systematically explores various annotator modeling techniques and compares their performance across seven corpora. From our findings, we show that the commonly used user token model consistently outperforms more complex models. We introduce a composite embedding approach and show distinct differences in which model performs best as a function of the agreement with a given dataset. Our findings shed light on the relationship between corpus statistics and annotator modeling performance, which informs future work on corpus construction and perspectivist NLP.
Authors: Haven Kim, Taketo Akama
Abstract: In musical compositions that include vocals, lyrics significantly contribute to artistic expression. Consequently, previous studies have introduced the concept of a recommendation system that suggests lyrics similar to a user's favorites or personalized preferences, aiding in the discovery of lyrics among millions of tracks. However, many of these systems do not fully consider human perceptions of lyric similarity, primarily due to limited research in this area. To bridge this gap, we conducted a comparative analysis of computational methods for modeling lyric similarity with human perception. Results indicated that computational models based on similarities between embeddings from pre-trained BERT-based models, the audio from which the lyrics are derived, and phonetic components are indicative of perceptual lyric similarity. This finding underscores the importance of semantic, stylistic, and phonetic similarities in human perception about lyric similarity. We anticipate that our findings will enhance the development of similarity-based lyric recommendation systems by offering pseudo-labels for neural network development and introducing objective evaluation metrics.
Authors: Victoria Graf, Qin Liu, Muhao Chen
Abstract: Data poisoning backdoor attacks can cause undesirable behaviors in large language models (LLMs), and defending against them is of increasing importance. Existing defense mechanisms often assume that only one type of trigger is adopted by the attacker, while defending against multiple simultaneous and independent trigger types necessitates general defense frameworks and is relatively unexplored. In this paper, we propose Nested Product of Experts(NPoE) defense framework, which involves a mixture of experts (MoE) as a trigger-only ensemble within the PoE defense framework to simultaneously defend against multiple trigger types. During NPoE training, the main model is trained in an ensemble with a mixture of smaller expert models that learn the features of backdoor triggers. At inference time, only the main model is used. Experimental results on sentiment analysis, hate speech detection, and question classification tasks demonstrate that NPoE effectively defends against a variety of triggers both separately and in trigger mixtures. Due to the versatility of the MoE structure in NPoE, this framework can be further expanded to defend against other attack settings
Authors: S M Rakib Hasan, Aakar Dhakal, Md Humaion Kabir Mehedi, Annajiat Alim Rasel
Abstract: Efforts on the research and development of OCR systems for Low-Resource Languages are relatively new. Low-resource languages have little training data available for training Machine Translation systems or other systems. Even though a vast amount of text has been digitized and made available on the internet the text is still in PDF and Image format, which are not instantly accessible. This paper discusses text recognition for two scripts: Bengali and Nepali; there are about 300 and 40 million Bengali and Nepali speakers respectively. In this study, using encoder-decoder transformers, a model was developed, and its efficacy was assessed using a collection of optical text images, both handwritten and printed. The results signify that the suggested technique corresponds with current approaches and achieves high precision in recognizing text in Bengali and Nepali. This study can pave the way for the advanced and accessible study of linguistics in South East Asia.
Authors: Yutong Shao, Ndapa Nakashole
Abstract: Structured data, prevalent in tables, databases, and knowledge graphs, poses a significant challenge in its representation. With the advent of large language models (LLMs), there has been a shift towards linearization-based methods, which process structured data as sequential token streams, diverging from approaches that explicitly model structure, often as a graph. Crucially, there remains a gap in our understanding of how these linearization-based methods handle structured data, which is inherently non-linear. This work investigates the linear handling of structured data in encoder-decoder language models, specifically T5. Our findings reveal the model's ability to mimic human-designed processes such as schema linking and syntax prediction, indicating a deep, meaningful learning of structure beyond simple token sequencing. We also uncover insights into the model's internal mechanisms, including the ego-centric nature of structure node encodings and the potential for model compression due to modality fusion redundancy. Overall, this work sheds light on the inner workings of linearization-based methods and could potentially provide guidance for future research.
Authors: Antoine Nzeyimana
Abstract: Morphological modeling in neural machine translation (NMT) is a promising approach to achieving open-vocabulary machine translation for morphologically-rich languages. However, existing methods such as sub-word tokenization and character-based models are limited to the surface forms of the words. In this work, we propose a framework-solution for modeling complex morphology in low-resource settings. A two-tier transformer architecture is chosen to encode morphological information at the inputs. At the target-side output, a multi-task multi-label training scheme coupled with a beam search-based decoder are found to improve machine translation performance. An attention augmentation scheme to the transformer model is proposed in a generic form to allow integration of pre-trained language models and also facilitate modeling of word order relationships between the source and target languages. Several data augmentation techniques are evaluated and shown to increase translation performance in low-resource settings. We evaluate our proposed solution on Kinyarwanda - English translation using public-domain parallel text. Our final models achieve competitive performance in relation to large multi-lingual models. We hope that our results will motivate more use of explicit morphological information and the proposed model and data augmentations in low-resource NMT.
Authors: Jun Wang, Qiongkai Xu, Xuanli He, Benjamin I. P. Rubinstein, Trevor Cohn
Abstract: While multilingual machine translation (MNMT) systems hold substantial promise, they also have security vulnerabilities. Our research highlights that MNMT systems can be susceptible to a particularly devious style of backdoor attack, whereby an attacker injects poisoned data into a low-resource language pair to cause malicious translations in other languages, including high-resource languages. Our experimental results reveal that injecting less than 0.01% poisoned data into a low-resource language pair can achieve an average 20% attack success rate in attacking high-resource language pairs. This type of attack is of particular concern, given the larger attack surface of languages inherent to low-resource settings. Our aim is to bring attention to these vulnerabilities within MNMT systems with the hope of encouraging the community to address security concerns in machine translation, especially in the context of low-resource languages.
Authors: Md. Kowsher, Ritesh Panditi, Nusrat Jahan Prottasha, Prakash Bhat, Anupam Kumar Bairagi, Mohammad Shamsul Arefin
Abstract: Conversational modeling using Large Language Models (LLMs) requires a nuanced understanding of context to generate coherent and contextually relevant responses. In this paper, we present Token Trails, a novel approach that leverages token-type embeddings to navigate the intricate contextual nuances within conversations. Our framework utilizes token-type embeddings to distinguish between user utterances and bot responses, facilitating the generation of context-aware replies. Through comprehensive experimentation and evaluation, we demonstrate the effectiveness of Token Trails in improving conversational understanding and response generation, achieving state-of-the-art performance. Our results highlight the significance of contextual modeling in conversational AI and underscore the promising potential of Token Trails to advance the field, paving the way for more sophisticated and contextually aware chatbot interactions.
Authors: Amirhossein Abaskohi, Sara Baruni, Mostafa Masoudi, Nesa Abbasi, Mohammad Hadi Babalou, Ali Edalat, Sepehr Kamahi, Samin Mahdizadeh Sani, Nikoo Naghavian, Danial Namazifard, Pouya Sadeghi, Yadollah Yaghoobzadeh
Abstract: This paper explores the efficacy of large language models (LLMs) for Persian. While ChatGPT and consequent LLMs have shown remarkable performance in English, their efficiency for more low-resource languages remains an open question. We present the first comprehensive benchmarking study of LLMs across diverse Persian language tasks. Our primary focus is on GPT-3.5-turbo, but we also include GPT-4 and OpenChat-3.5 to provide a more holistic evaluation. Our assessment encompasses a diverse set of tasks categorized into classic, reasoning, and knowledge-based domains. To enable a thorough comparison, we evaluate LLMs against existing task-specific fine-tuned models. Given the limited availability of Persian datasets for reasoning tasks, we introduce two new benchmarks: one based on elementary school math questions and another derived from the entrance exams for 7th and 10th grades. Our findings reveal that while LLMs, especially GPT-4, excel in tasks requiring reasoning abilities and a broad understanding of general knowledge, they often lag behind smaller pre-trained models fine-tuned specifically for particular tasks. Additionally, we observe improved performance when test sets are translated to English before inputting them into GPT-3.5. These results highlight the significant potential for enhancing LLM performance in the Persian language. This is particularly noteworthy due to the unique attributes of Persian, including its distinct alphabet and writing styles.
Authors: Zaid Sheikh, Antonios Anastasopoulos, Shruti Rijhwani, Lindia Tjuatja, Robbie Jimerson, Graham Neubig
Abstract: Effectively using Natural Language Processing (NLP) tools in under-resourced languages requires a thorough understanding of the language itself, familiarity with the latest models and training methodologies, and technical expertise to deploy these models. This could present a significant obstacle for language community members and linguists to use NLP tools. This paper introduces the CMU Linguistic Annotation Backend, an open-source framework that simplifies model deployment and continuous human-in-the-loop fine-tuning of NLP models. CMULAB enables users to leverage the power of multilingual models to quickly adapt and extend existing tools for speech recognition, OCR, translation, and syntactic analysis to new languages, even with limited training data. We describe various tools and APIs that are currently available and how developers can easily add new models/functionality to the framework. Code is available at https://github.com/neulab/cmulab along with a live demo at https://cmulab.dev
Authors: Jennifer Hu, Michael C. Frank
Abstract: Developmental psychologists have argued about when cognitive capacities such as language understanding or theory of mind emerge. These debates often hinge on the concept of "task demands" -- the auxiliary challenges associated with performing a particular evaluation -- that may mask the child's underlying ability. The same issues arise when measuring the capacities of language models (LMs): performance on a task is a function of the model's underlying competence, combined with the model's ability to interpret and perform the task given its available resources. Here, we show that for analogical reasoning, reflective reasoning, word prediction, and grammaticality judgments, evaluation methods with greater task demands yield lower performance than evaluations with reduced demands. This "demand gap" is most pronounced for models with fewer parameters and less training data. Our results illustrate that LM performance should not be interpreted as a direct indication of intelligence (or lack thereof), but as a reflection of capacities seen through the lens of researchers' design choices.
Authors: Thinh Hung Truong, Yulia Otmakhova, Karin Verspoor, Trevor Cohn, Timothy Baldwin
Abstract: In this work, we measure the impact of affixal negation on modern English large language models (LLMs). In affixal negation, the negated meaning is expressed through a negative morpheme, which is potentially challenging for LLMs as their tokenizers are often not morphologically plausible. We conduct extensive experiments using LLMs with different subword tokenization methods, which lead to several insights on the interaction between tokenization performance and negation sensitivity. Despite some interesting mismatches between tokenization accuracy and negation detection performance, we show that models can, on the whole, reliably recognize the meaning of affixal negation.
Authors: Parth Patwa, Simone Filice, Zhiyu Chen, Giuseppe Castellucci, Oleg Rokhlenko, Shervin Malmasi
Abstract: Large Language Models (LLMs) operating in 0-shot or few-shot settings achieve competitive results in Text Classification tasks. In-Context Learning (ICL) typically achieves better accuracy than the 0-shot setting, but it pays in terms of efficiency, due to the longer input prompt. In this paper, we propose a strategy to make LLMs as efficient as 0-shot text classifiers, while getting comparable or better accuracy than ICL. Our solution targets the low resource setting, i.e., when only 4 examples per class are available. Using a single LLM and few-shot real data we perform a sequence of generation, filtering and Parameter-Efficient Fine-Tuning steps to create a robust and efficient classifier. Experimental results show that our approach leads to competitive results on multiple text classification datasets.
Authors: Takeshi Kojima, Itsuki Okimura, Yusuke Iwasawa, Hitomi Yanaka, Yutaka Matsuo
Abstract: Current decoder-based pre-trained language models (PLMs) successfully demonstrate multilingual capabilities. However, it is unclear how these models handle multilingualism. We analyze the neuron-level internal behavior of multilingual decoder-based PLMs, Specifically examining the existence of neurons that fire ``uniquely for each language'' within decoder-only multilingual PLMs. We analyze six languages: English, German, French, Spanish, Chinese, and Japanese, and show that language-specific neurons are unique, with a slight overlap (< 5%) between languages. These neurons are mainly distributed in the models' first and last few layers. This trend remains consistent across languages and models. Additionally, we tamper with less than 1% of the total neurons in each model during inference and demonstrate that tampering with a few language-specific neurons drastically changes the probability of target language occurrence in text generation.
Authors: Shuxian Fan, Adam Visokay, Kentaro Hoffman, Stephen Salerno, Li Liu, Jeffrey T. Leek, Tyler H. McCormick
Abstract: In settings where most deaths occur outside the healthcare system, verbal autopsies (VAs) are a common tool to monitor trends in causes of death (COD). VAs are interviews with a surviving caregiver or relative that are used to predict the decedent's COD. Turning VAs into actionable insights for researchers and policymakers requires two steps (i) predicting likely COD using the VA interview and (ii) performing inference with predicted CODs (e.g. modeling the breakdown of causes by demographic factors using a sample of deaths). In this paper, we develop a method for valid inference using outcomes (in our case COD) predicted from free-form text using state-of-the-art NLP techniques. This method, which we call multiPPI++, extends recent work in "prediction-powered inference" to multinomial classification. We leverage a suite of NLP techniques for COD prediction and, through empirical analysis of VA data, demonstrate the effectiveness of our approach in handling transportability issues. multiPPI++ recovers ground truth estimates, regardless of which NLP model produced predictions and regardless of whether they were produced by a more accurate predictor like GPT-4-32k or a less accurate predictor like KNN. Our findings demonstrate the practical importance of inference correction for public health decision-making and suggests that if inference tasks are the end goal, having a small amount of contextually relevant, high quality labeled data is essential regardless of the NLP algorithm.
Authors: Paiheng Xu, Jing Liu, Nathan Jones, Julie Cohen, Wei Ai
Abstract: Assessing instruction quality is a fundamental component of any improvement efforts in the education system. However, traditional manual assessments are expensive, subjective, and heavily dependent on observers' expertise and idiosyncratic factors, preventing teachers from getting timely and frequent feedback. Different from prior research that mostly focuses on low-inference instructional practices on a singular basis, this paper presents the first study that leverages Natural Language Processing (NLP) techniques to assess multiple high-inference instructional practices in two distinct educational settings: in-person K-12 classrooms and simulated performance tasks for pre-service teachers. This is also the first study that applies NLP to measure a teaching practice that is widely acknowledged to be particularly effective for students with special needs. We confront two challenges inherent in NLP-based instructional analysis, including noisy and long input data and highly skewed distributions of human ratings. Our results suggest that pretrained Language Models (PLMs) demonstrate performances comparable to the agreement level of human raters for variables that are more discrete and require lower inference, but their efficacy diminishes with more complex teaching practices. Interestingly, using only teachers' utterances as input yields strong results for student-centered variables, alleviating common concerns over the difficulty of collecting and transcribing high-quality student speech data in in-person teaching settings. Our findings highlight both the potential and the limitations of current NLP techniques in the education domain, opening avenues for further exploration.
Authors: Emilio Villa-Cueva, A. Pastor L\'opez-Monroy, Fernando S\'anchez-Vega, Thamar Solorio
Abstract: Zero-Shot Cross-lingual Transfer (ZS-XLT) utilizes a model trained in a source language to make predictions in another language, often with a performance loss. To alleviate this, additional improvements can be achieved through subsequent adaptation using examples in the target language. In this paper, we exploit In-Context Tuning (ICT) for One-Shot Cross-lingual transfer in the classification task by introducing In-Context Cross-lingual Transfer (IC-XLT). The novel concept involves training a model to learn from context examples and subsequently adapting it during inference to a target language by prepending a One-Shot context demonstration in that language. Our results show that IC-XLT successfully leverages target-language examples to improve the cross-lingual capabilities of the evaluated mT5 model, outperforming prompt-based models in the Zero and Few-shot scenarios adapted through fine-tuning. Moreover, we show that when source-language data is limited, the fine-tuning framework employed for IC-XLT performs comparably to prompt-based fine-tuning with significantly more training data in the source language.
Authors: Ashima Suvarna, Harshita Khandelwal, Nanyun Peng
Abstract: Phonology, the study of speech's structure and pronunciation rules, is a critical yet often overlooked component in Large Language Model (LLM) research. LLMs are widely used in various downstream applications that leverage phonology such as educational tools and poetry generation. Moreover, LLMs can potentially learn imperfect associations between orthographic and phonological forms from the training data. Thus, it is imperative to benchmark the phonological skills of LLMs. To this end, we present PhonologyBench, a novel benchmark consisting of three diagnostic tasks designed to explicitly test the phonological skills of LLMs in English: grapheme-to-phoneme conversion, syllable counting, and rhyme word generation. Despite having no access to speech data, LLMs showcased notable performance on the PhonologyBench tasks. However, we observe a significant gap of 17% and 45% on Rhyme Word Generation and Syllable counting, respectively, when compared to humans. Our findings underscore the importance of studying LLM performance on phonological tasks that inadvertently impact real-world applications. Furthermore, we encourage researchers to choose LLMs that perform well on the phonological task that is closely related to the downstream application since we find that no single model consistently outperforms the others on all the tasks.
Authors: Masayuki Kawarada, Tatsuya Ishigaki, Hiroya Takamura
Abstract: Large language models (LLMs) have been applied to a wide range of data-to-text generation tasks, including tables, graphs, and time-series numerical data-to-text settings. While research on generating prompts for structured data such as tables and graphs is gaining momentum, in-depth investigations into prompting for time-series numerical data are lacking. Therefore, this study explores various input representations, including sequences of tokens and structured formats such as HTML, LaTeX, and Python-style codes. In our experiments, we focus on the task of Market Comment Generation, which involves taking a numerical sequence of stock prices as input and generating a corresponding market comment. Contrary to our expectations, the results show that prompts resembling programming languages yield better outcomes, whereas those similar to natural languages and longer formats, such as HTML and LaTeX, are less effective. Our findings offer insights into creating effective prompts for tasks that generate text from numerical sequences.
Authors: Pouya Sadeghi, Amirhossein Abaskohi, Yadollah Yaghoobzadeh
Abstract: Inspired by human cognition, Jiang et al.(2023c) create a benchmark for assessing LLMs' lateral thinking-thinking outside the box. Building upon this benchmark, we investigate how different prompting methods enhance LLMs' performance on this task to reveal their inherent power for outside-the-box thinking ability. Through participating in SemEval-2024, task 9, Sentence Puzzle sub-task, we explore prompt engineering methods: chain of thoughts (CoT) and direct prompting, enhancing with informative descriptions, and employing contextualizing prompts using a retrieval augmented generation (RAG) pipeline. Our experiments involve three LLMs including GPT-3.5, GPT-4, and Zephyr-7B-beta. We generate a dataset of thinking paths between riddles and options using GPT-4, validated by humans for quality. Findings indicate that compressed informative prompts enhance performance. Dynamic in-context learning enhances model performance significantly. Furthermore, fine-tuning Zephyr on our dataset enhances performance across other commonsense datasets, underscoring the value of innovative thinking.
Authors: Zhongtao Miao, Qiyu Wu, Kaiyan Zhao, Zilong Wu, Yoshimasa Tsuruoka
Abstract: The field of cross-lingual sentence embeddings has recently experienced significant advancements, but research concerning low-resource languages has lagged due to the scarcity of parallel corpora. This paper shows that cross-lingual word representation in low-resource languages is notably under-aligned with that in high-resource languages in current models. To address this, we introduce a novel framework that explicitly aligns words between English and eight low-resource languages, utilizing off-the-shelf word alignment models. This framework incorporates three primary training objectives: aligned word prediction and word translation ranking, along with the widely used translation ranking. We evaluate our approach through experiments on the bitext retrieval task, which demonstrate substantial improvements on sentence embeddings in low-resource languages. In addition, the competitive performance of the proposed model across a broader range of tasks in high-resource languages underscores its practicality.
Authors: Ye Yuan, Kexin Tang, Jianhao Shen, Ming Zhang, Chenguang Wang
Abstract: We present a new challenge to examine whether large language models understand social norms. In contrast to existing datasets, our dataset requires a fundamental understanding of social norms to solve. Our dataset features the largest set of social norm skills, consisting of 402 skills and 12,383 questions covering a wide set of social norms ranging from opinions and arguments to culture and laws. We design our dataset according to the K-12 curriculum. This enables the direct comparison of the social understanding of large language models to humans, more specifically, elementary students. While prior work generates nearly random accuracy on our benchmark, recent large language models such as GPT3.5-Turbo and LLaMA2-Chat are able to improve the performance significantly, only slightly below human performance. We then propose a multi-agent framework based on large language models to improve the models' ability to understand social norms. This method further improves large language models to be on par with humans. Given the increasing adoption of large language models in real-world applications, our finding is particularly important and presents a unique direction for future improvements.
Authors: Zhe Xu, Daoyuan Chen, Jiayi Kuang, Zihao Yi, Yaliang Li, Ying Shen
Abstract: Emotional Support Conversation (ESC) systems are pivotal in providing empathetic interactions, aiding users through negative emotional states by understanding and addressing their unique experiences. In this paper, we tackle two key challenges in ESC: enhancing contextually relevant and empathetic response generation through dynamic demonstration retrieval, and advancing cognitive understanding to grasp implicit mental states comprehensively. We introduce Dynamic Demonstration Retrieval and Cognitive-Aspect Situation Understanding (\ourwork), a novel approach that synergizes these elements to improve the quality of support provided in ESCs. By leveraging in-context learning and persona information, we introduce an innovative retrieval mechanism that selects informative and personalized demonstration pairs. We also propose a cognitive understanding module that utilizes four cognitive relationships from the ATOMIC knowledge source to deepen situational awareness of help-seekers' mental states. Our supportive decoder integrates information from diverse knowledge sources, underpinning response generation that is both empathetic and cognitively aware. The effectiveness of \ourwork is demonstrated through extensive automatic and human evaluations, revealing substantial improvements over numerous state-of-the-art models, with up to 13.79\% enhancement in overall performance of ten metrics. Our codes are available for public access to facilitate further research and development.
Authors: Chengwei Qin, Ruirui Chen, Ruochen Zhao, Wenhan Xia, Shafiq Joty
Abstract: To mitigate forgetting, existing lifelong event detection methods typically maintain a memory module and replay the stored memory data during the learning of a new task. However, the simple combination of memory data and new-task samples can still result in substantial forgetting of previously acquired knowledge, which may occur due to the potential overlap between the feature distribution of new data and the previously learned embedding space. Moreover, the model suffers from overfitting on the few memory samples rather than effectively remembering learned patterns. To address the challenges of forgetting and overfitting, we propose a novel method based on embedding space separation and compaction. Our method alleviates forgetting of previously learned tasks by forcing the feature distribution of new data away from the previous embedding space. It also mitigates overfitting by a memory calibration mechanism that encourages memory data to be close to its prototype to enhance intra-class compactness. In addition, the learnable parameters of the new task are initialized by drawing upon acquired knowledge from the previously learned task to facilitate forward knowledge transfer. With extensive experiments, we demonstrate that our method can significantly outperform previous state-of-the-art approaches.
Authors: Vandan Mujadia, Pruthwik Mishra, Arafat Ahsan, Dipti Misra Sharma
Abstract: With the primary focus on evaluating the effectiveness of large language models for automatic reference-less translation assessment, this work presents our experiments on mimicking human direct assessment to evaluate the quality of translations in English and Indian languages. We constructed a translation evaluation task where we performed zero-shot learning, in-context example-driven learning, and fine-tuning of large language models to provide a score out of 100, where 100 represents a perfect translation and 1 represents a poor translation. We compared the performance of our trained systems with existing methods such as COMET, BERT-Scorer, and LABSE, and found that the LLM-based evaluator (LLaMA-2-13B) achieves a comparable or higher overall correlation with human judgments for the considered Indian language pairs.
Authors: Maja Stahl, Nadine Michel, Sebastian Kilsbach, Julian Schmidtke, Sara Rezat, Henning Wachsmuth
Abstract: Learning argumentative writing is challenging. Besides writing fundamentals such as syntax and grammar, learners must select and arrange argument components meaningfully to create high-quality essays. To support argumentative writing computationally, one step is to mine the argumentative structure. When combined with automatic essay scoring, interactions of the argumentative structure and quality scores can be exploited for comprehensive writing support. Although studies have shown the usefulness of using information about the argumentative structure for essay scoring, no argument mining corpus with ground-truth essay quality annotations has been published yet. Moreover, none of the existing corpora contain essays written by school students specifically. To fill this research gap, we present a German corpus of 1,320 essays from school students of two age groups. Each essay has been manually annotated for argumentative structure and quality on multiple levels of granularity. We propose baseline approaches to argument mining and essay scoring, and we analyze interactions between both tasks, thereby laying the ground for quality-oriented argumentative writing support.
Authors: Osvaldo Luamba Quinjica, David Ifeoluwa Adelani
Abstract: In recent years, the development of pre-trained language models (PLMs) has gained momentum, showcasing their capacity to transcend linguistic barriers and facilitate knowledge transfer across diverse languages. However, this progress has predominantly bypassed the inclusion of very-low resource languages, creating a notable void in the multilingual landscape. This paper addresses this gap by introducing four tailored PLMs specifically finetuned for Angolan languages, employing a Multilingual Adaptive Fine-tuning (MAFT) approach. In this paper, we survey the role of informed embedding initialization and synthetic data in enhancing the performance of MAFT models in downstream tasks. We improve baseline over SOTA AfroXLMR-base (developed through MAFT) and OFA (an effective embedding initialization) by 12.3 and 3.8 points respectively.
Authors: Nishat Raihan, Dhiman Goswami, Sadiya Sayara Chowdhury Puspo, Christian Newman, Tharindu Ranasinghe, Marcos Zampieri
Abstract: Recent advances in AI, machine learning, and NLP have led to the development of a new generation of Large Language Models (LLMs) that are trained on massive amounts of data and often have trillions of parameters. Commercial applications (e.g., ChatGPT) have made this technology available to the general public, thus making it possible to use LLMs to produce high-quality texts for academic and professional purposes. Schools and universities are aware of the increasing use of AI-generated content by students and they have been researching the impact of this new technology and its potential misuse. Educational programs in Computer Science (CS) and related fields are particularly affected because LLMs are also capable of generating programming code in various programming languages. To help understand the potential impact of publicly available LLMs in CS education, we introduce CSEPrompts, a framework with hundreds of programming exercise prompts and multiple-choice questions retrieved from introductory CS and programming courses. We also provide experimental results on CSEPrompts to evaluate the performance of several LLMs with respect to generating Python code and answering basic computer science and programming questions.
Authors: Shijia Zhou, Huangyan Shan, Barbara Plank, Robert Litschko
Abstract: This paper presents our system developed for the SemEval-2024 Task 1: Semantic Textual Relatedness (STR), on Track C: Cross-lingual. The task aims to detect semantic relatedness of two sentences in a given target language without access to direct supervision (i.e. zero-shot cross-lingual transfer). To this end, we focus on different source language selection strategies on two different pre-trained languages models: XLM-R and Furina. We experiment with 1) single-source transfer and select source languages based on typological similarity, 2) augmenting English training data with the two nearest-neighbor source languages, and 3) multi-source transfer where we compare selecting on all training languages against languages from the same family. We further study machine translation-based data augmentation and the impact of script differences. Our submission achieved the first place in the C8 (Kinyarwanda) test set.
Authors: Hyungjoo Chae, Yeonghyeon Kim, Seungone Kim, Kai Tzu-iunn Ong, Beong-woo Kwak, Moohyeon Kim, Seonghwan Kim, Taeyoon Kwon, Jiwan Chung, Youngjae Yu, Jinyoung Yeo
Abstract: Algorithmic reasoning refers to the ability to understand the complex patterns behind the problem and decompose them into a sequence of reasoning steps towards the solution. Such nature of algorithmic reasoning makes it a challenge for large language models (LLMs), even though they have demonstrated promising performance in other reasoning tasks. Within this context, some recent studies use programming languages (e.g., Python) to express the necessary logic for solving a given instance/question (e.g., Program-of-Thought) as inspired by their strict and precise syntaxes. However, it is non-trivial to write an executable code that expresses the correct logic on the fly within a single inference call. Also, the code generated specifically for an instance cannot be reused for others, even if they are from the same task and might require identical logic to solve. This paper presents Think-and-Execute, a novel framework that decomposes the reasoning process of language models into two steps. (1) In Think, we discover a task-level logic that is shared across all instances for solving a given task and then express the logic with pseudocode; (2) In Execute, we further tailor the generated pseudocode to each instance and simulate the execution of the code. With extensive experiments on seven algorithmic reasoning tasks, we demonstrate the effectiveness of Think-and-Execute. Our approach better improves LMs' reasoning compared to several strong baselines performing instance-specific reasoning (e.g., CoT and PoT), suggesting the helpfulness of discovering task-level logic. Also, we show that compared to natural language, pseudocode can better guide the reasoning of LMs, even though they are trained to follow natural language instructions.
Authors: Eunseong Choi, Hyeri Lee, Jongwuk Lee
Abstract: In Open-domain Question Answering (ODQA), it is essential to discern relevant contexts as evidence and avoid spurious ones among retrieved results. The model architecture that uses concatenated multiple contexts in the decoding phase, i.e., Fusion-in-Decoder, demonstrates promising performance but generates incorrect outputs from seemingly plausible contexts. To address this problem, we propose the Multi-Granularity guided Fusion-in-Decoder (MGFiD), discerning evidence across multiple levels of granularity. Based on multi-task learning, MGFiD harmonizes passage re-ranking with sentence classification. It aggregates evident sentences into an anchor vector that instructs the decoder. Additionally, it improves decoding efficiency by reusing the results of passage re-ranking for passage pruning. Through our experiments, MGFiD outperforms existing models on the Natural Questions (NQ) and TriviaQA (TQA) datasets, highlighting the benefits of its multi-granularity solution.
Authors: Jakub Hoscilowicz, Pawel Pawlowski, Marcin Skorupa, Marcin Sowa\'nski, Artur Janicki
Abstract: Spoken Language Understanding (SLU) models are a core component of voice assistants (VA), such as Alexa, Bixby, and Google Assistant. In this paper, we introduce a pipeline designed to extend SLU systems to new languages, utilizing Large Language Models (LLMs) that we fine-tune for machine translation of slot-annotated SLU training data. Our approach improved on the MultiATIS++ benchmark, a primary multi-language SLU dataset, in the cloud scenario using an mBERT model. Specifically, we saw an improvement in the Overall Accuracy metric: from 53% to 62.18%, compared to the existing state-of-the-art method, Fine and Coarse-grained Multi-Task Learning Framework (FC-MTLF). In the on-device scenario (tiny and not pretrained SLU), our method improved the Overall Accuracy from 5.31% to 22.06% over the baseline Global-Local Contrastive Learning Framework (GL-CLeF) method. Contrary to both FC-MTLF and GL-CLeF, our LLM-based machine translation does not require changes in the production architecture of SLU. Additionally, our pipeline is slot-type independent: it does not require any slot definitions or examples.
Authors: Zhiyuan Wen, Jiannong Cao, Yu Yang, Ruosong Yang, Shuaiqi Liu
Abstract: Personality Recognition in Conversation (PRC) aims to identify the personality traits of speakers through textual dialogue content. It is essential for providing personalized services in various applications of Human-Computer Interaction (HCI), such as AI-based mental therapy and companion robots for the elderly. Most recent studies analyze the dialog content for personality classification yet overlook two major concerns that hinder their performance. First, crucial implicit factors contained in conversation, such as emotions that reflect the speakers' personalities are ignored. Second, only focusing on the input dialog content disregards the semantic understanding of personality itself, which reduces the interpretability of the results. In this paper, we propose Affective Natural Language Inference (Affective-NLI) for accurate and interpretable PRC. To utilize affectivity within dialog content for accurate personality recognition, we fine-tuned a pre-trained language model specifically for emotion recognition in conversations, facilitating real-time affective annotations for utterances. For interpretability of recognition results, we formulate personality recognition as an NLI problem by determining whether the textual description of personality labels is entailed by the dialog content. Extensive experiments on two daily conversation datasets suggest that Affective-NLI significantly outperforms (by 6%-7%) state-of-the-art approaches. Additionally, our Flow experiment demonstrates that Affective-NLI can accurately recognize the speaker's personality in the early stages of conversations by surpassing state-of-the-art methods with 22%-34%.
Authors: Yejin Jeon, Yunsu Kim, Gary Geunbae Lee
Abstract: Contemporary neural speech synthesis models have indeed demonstrated remarkable proficiency in synthetic speech generation as they have attained a level of quality comparable to that of human-produced speech. Nevertheless, it is important to note that these achievements have predominantly been verified within the context of high-resource languages such as English. Furthermore, the Tacotron and FastSpeech variants show substantial pausing errors when applied to the Korean language, which affects speech perception and naturalness. In order to address the aforementioned issues, we propose a novel framework that incorporates comprehensive modeling of both syntactic and acoustic cues that are associated with pausing patterns. Remarkably, our framework possesses the capability to consistently generate natural speech even for considerably more extended and intricate out-of-domain (OOD) sentences, despite its training on short audio clips. Architectural design choices are validated through comparisons with baseline models and ablation studies using subjective and objective metrics, thus confirming model performance.
Authors: Katrin Erk, Marianna Apidianaki
Abstract: Embedding spaces contain interpretable dimensions indicating gender, formality in style, or even object properties. This has been observed multiple times. Such interpretable dimensions are becoming valuable tools in different areas of study, from social science to neuroscience. The standard way to compute these dimensions uses contrasting seed words and computes difference vectors over them. This is simple but does not always work well. We combine seed-based vectors with guidance from human ratings of where words fall along a specific dimension, and evaluate on predicting both object properties like size and danger, and the stylistic properties of formality and complexity. We obtain interpretable dimensions with markedly better performance especially in cases where seed-based dimensions do not work well.
Authors: Julia Rozanova, Marco Valentino, Andr\'e Freitas
Abstract: Rigorous evaluation of the causal effects of semantic features on language model predictions can be hard to achieve for natural language reasoning problems. However, this is such a desirable form of analysis from both an interpretability and model evaluation perspective, that it is valuable to investigate specific patterns of reasoning with enough structure and regularity to identify and quantify systematic reasoning failures in widely-used models. In this vein, we pick a portion of the NLI task for which an explicit causal diagram can be systematically constructed: the case where across two sentences (the premise and hypothesis), two related words/terms occur in a shared context. In this work, we apply causal effect estimation strategies to measure the effect of context interventions (whose effect on the entailment label is mediated by the semantic monotonicity characteristic) and interventions on the inserted word-pair (whose effect on the entailment label is mediated by the relation between these words). Extending related work on causal analysis of NLP models in different settings, we perform an extensive interventional study on the NLI task to investigate robustness to irrelevant changes and sensitivity to impactful changes of Transformers. The results strongly bolster the fact that similar benchmark accuracy scores may be observed for models that exhibit very different behaviour. Moreover, our methodology reinforces previously suspected biases from a causal perspective, including biases in favour of upward-monotone contexts and ignoring the effects of negation markers.
Authors: Mokanarangan Thayaparan, Marco Valentino, Andr\'e Freitas
Abstract: Integer Linear Programming (ILP) has been proposed as a formalism for encoding precise structural and semantic constraints for Natural Language Inference (NLI). However, traditional ILP frameworks are non-differentiable, posing critical challenges for the integration of continuous language representations based on deep learning. In this paper, we introduce a novel approach, named Diff-Comb Explainer, a neuro-symbolic architecture for explanation-based NLI based on Differentiable BlackBox Combinatorial Solvers (DBCS). Differently from existing neuro-symbolic solvers, Diff-Comb Explainer does not necessitate a continuous relaxation of the semantic constraints, enabling a direct, more precise, and efficient incorporation of neural representations into the ILP formulation. Our experiments demonstrate that Diff-Comb Explainer achieves superior performance when compared to conventional ILP solvers, neuro-symbolic black-box solvers, and Transformer-based encoders. Moreover, a deeper analysis reveals that Diff-Comb Explainer can significantly improve the precision, consistency, and faithfulness of the constructed explanations, opening new opportunities for research on neuro-symbolic architectures for explainable and transparent NLI in complex domains.
Authors: Mozhi Zhang, Mianqiu Huang, Rundong Shi, Linsen Guo, Chong Peng, Peng Yan, Yaqian Zhou, Xipeng Qiu
Abstract: Large language models optimized with techniques like RLHF have achieved good alignment in being helpful and harmless. However, post-alignment, these language models often exhibit overconfidence, where the expressed confidence does not accurately calibrate with their correctness rate. In this paper, we decompose the language model confidence into the \textit{Uncertainty} about the question and the \textit{Fidelity} to the answer generated by language models. Then, we propose a plug-and-play method to estimate the confidence of language models. Our method has shown good calibration performance by conducting experiments with 6 RLHF-LMs on four MCQA datasets. Moreover, we propose two novel metrics, IPR and CE, to evaluate the calibration of the model, and we have conducted a detailed discussion on \textit{Truly Well-Calibrated Confidence}. Our method could serve as a strong baseline, and we hope that this work will provide some insights into the model confidence calibration.
Authors: Taiqiang Wu, Chaofan Tao, Jiahao Wang, Zhe Zhao, Ngai Wong
Abstract: Kullback-Leiber divergence has been widely used in Knowledge Distillation (KD) to compress Large Language Models (LLMs). Contrary to prior assertions that reverse Kullback-Leibler (RKL) divergence is mode-seeking and thus preferable over the mean-seeking forward Kullback-Leibler (FKL) divergence, this study empirically and theoretically demonstrates that neither mode-seeking nor mean-seeking properties manifest in KD for LLMs. Instead, RKL and FKL are found to share the same optimization objective and both converge after a sufficient number of epochs. However, due to practical constraints, LLMs are seldom trained for such an extensive number of epochs. Meanwhile, we further find that RKL focuses on the tail part of the distributions, while FKL focuses on the head part at the beginning epochs. Consequently, we propose a simple yet effective Adaptive Kullback-Leiber (AKL) divergence method, which adaptively allocates weights to combine FKL and RKL. Metric-based and GPT-4-based evaluations demonstrate that the proposed AKL outperforms the baselines across various tasks and improves the diversity and quality of generated responses.
Authors: Arianna Muti, Federico Ruggeri, Cagri Toraman, Lorenzo Musetti, Samuel Algherini, Silvia Ronchi, Gianmarco Saretto, Caterina Zapparoli, Alberto Barr\'on-Cede\~no
Abstract: Misogyny is often expressed through figurative language. Some neutral words can assume a negative connotation when functioning as pejorative epithets. Disambiguating the meaning of such terms might help the detection of misogyny. In order to address such task, we present PejorativITy, a novel corpus of 1,200 manually annotated Italian tweets for pejorative language at the word level and misogyny at the sentence level. We evaluate the impact of injecting information about disambiguated words into a model targeting misogyny detection. In particular, we explore two different approaches for injection: concatenation of pejorative information and substitution of ambiguous words with univocal terms. Our experimental results, both on our corpus and on two popular benchmarks on Italian tweets, show that both approaches lead to a major classification improvement, indicating that word sense disambiguation is a promising preliminary step for misogyny detection. Furthermore, we investigate LLMs' understanding of pejorative epithets by means of contextual word embeddings analysis and prompting.
Authors: Sehyun Choi
Abstract: Recently, multiple architectures has been proposed to improve the efficiency of the Transformer Language Models through changing the design of the self-attention block to have a linear-cost inference (LCI). A notable approach in this realm is the State-Space Machines (SSMs) architecture, which showed on-par performance on language modeling tasks with the self-attention transformers. However, such an architectural change requires a full pretraining of the weights from scratch, which incurs a huge cost to researchers and practitioners who want to use the new architectures. In the more traditional linear attention works, it has been proposed to approximate full attention with linear attention by swap-and-finetune framework. Motivated by this approach, we propose Cross-Architecture Transfer Learning (XATL), in which the weights of the shared components between LCI and self-attention-based transformers, such as layernorms, MLPs, input/output embeddings, are directly transferred to the new architecture from already pre-trained model parameters. We experimented the efficacy of the method on varying sizes and alternative attention architectures and show that \methodabbr significantly reduces the training time up to 2.5x times and converges to a better minimum with up to 2.6% stronger model on the LM benchmarks within the same compute budget.
Authors: Zihan Yao, Yu He, Tianyu Qi, Ming Li
Abstract: Addressing the issue of hallucinations and outdated knowledge in large language models is critical for their reliable application. Model Editing presents a promising avenue for mitigating these challenges in a cost-effective manner. However, existing methods often suffer from unsatisfactory generalization and unintended effects on unrelated samples. To overcome these limitations, we introduce a novel approach: Scalable Model Editing via Customized Expert Networks (SCEN), which is a two-stage continuous training paradigm. Specifically, in the first stage, we train lightweight expert networks individually for each piece of knowledge that needs to be updated. Subsequently, we train a corresponding neuron for each expert to control the activation state of that expert. Our experiments on two different sizes of open-source large language models, the Llama2 7B and 13B, achieve state-of-the-art results compared to existing mainstream Model Editing methods. Our code is available at https: //github.com/TAL-auroraX/SCEN
Authors: Zheng Yuan, Dorina de Jong, \v{S}tefan Be\v{n}u\v{s}, No\"el Nguyen, Ruitao Feng, R\'obert Sabo, Luciano Fadiga, Alessandro D`Ausilio
Abstract: We introduce the Alternating Reading Task (ART) Corpus, a collection of dyadic sentence reading for studying the entrainment and imitation behaviour in speech communication. The ART corpus features three experimental conditions - solo reading, alternating reading, and deliberate imitation - as well as three sub-corpora encompassing French-, Italian-, and Slovak-accented English. This design allows systematic investigation of speech entrainment in a controlled and less-spontaneous setting. Alongside detailed transcriptions, it includes English proficiency scores, demographics, and in-experiment questionnaires for probing linguistic, personal and interpersonal influences on entrainment. Our presentation covers its design, collection, annotation processes, initial analysis, and future research prospects.
Authors: Viet-Tung Do, Van-Khanh Hoang, Duy-Hung Nguyen, Shahab Sabahi, Jeff Yang, Hajime Hotta, Minh-Tien Nguyen, Hung Le
Abstract: Large Language Models (LLMs) can perform various natural language processing tasks with suitable instruction prompts. However, designing effective prompts manually is challenging and time-consuming. Existing methods for automatic prompt optimization either lack flexibility or efficiency. In this paper, we propose an effective approach to automatically select the optimal prompt for a given input from a finite set of synthetic candidate prompts. Our approach consists of three steps: (1) clustering the training data and generating candidate prompts for each cluster using an LLM-based prompt generator; (2) synthesizing a dataset of input-prompt-output tuples for training a prompt evaluator to rank the prompts based on their relevance to the input; (3) using the prompt evaluator to select the best prompt for a new input at test time. Our approach balances prompt generality-specificity and eliminates the need for resource-intensive training and inference. It demonstrates competitive performance on zero-shot question-answering datasets: GSM8K, MultiArith, and AQuA.
Authors: Maike Behrendt, Stefan Sylvius Wagner, Marc Ziegele, Lena Wilms, Anke Stoll, Dominique Heinbach, Stefan Harmeling
Abstract: Measuring the quality of contributions in political online discussions is crucial in deliberation research and computer science. Research has identified various indicators to assess online discussion quality, and with deep learning advancements, automating these measures has become feasible. While some studies focus on analyzing specific quality indicators, a comprehensive quality score incorporating various deliberative aspects is often preferred. In this work, we introduce AQuA, an additive score that calculates a unified deliberative quality score from multiple indices for each discussion post. Unlike other singular scores, AQuA preserves information on the deliberative aspects present in comments, enhancing model transparency. We develop adapter models for 20 deliberative indices, and calculate correlation coefficients between experts' annotations and the perceived deliberativeness by non-experts to weigh the individual indices into a single deliberative score. We demonstrate that the AQuA score can be computed easily from pre-trained adapters and aligns well with annotations on other datasets that have not be seen during training. The analysis of experts' vs. non-experts' annotations confirms theoretical findings in the social science literature.
Authors: Ziyang Wang, Sanwoo Lee, Hsiu-Yuan Huang, Yunfang Wu
Abstract: Prompt-based methods have achieved promising results in most few-shot text classification tasks. However, for readability assessment tasks, traditional prompt methods lackcrucial linguistic knowledge, which has already been proven to be essential. Moreover, previous studies on utilizing linguistic features have shown non-robust performance in few-shot settings and may even impair model performance.To address these issues, we propose a novel prompt-based tuning framework that incorporates rich linguistic knowledge, called Feature Prompt Tuning (FPT). Specifically, we extract linguistic features from the text and embed them into trainable soft prompts. Further, we devise a new loss function to calibrate the similarity ranking order between categories. Experimental results demonstrate that our proposed method FTP not only exhibits a significant performance improvement over the prior best prompt-based tuning approaches, but also surpasses the previous leading methods that incorporate linguistic features. Also, our proposed model significantly outperforms the large language model gpt-3.5-turbo-16k in most cases. Our proposed method establishes a new architecture for prompt tuning that sheds light on how linguistic features can be easily adapted to linguistic-related tasks.
Authors: Bernardo Leite, Henrique Lopes Cardoso
Abstract: Question Generation aims to automatically generate questions based on a given input provided as context. A controllable question generation scheme focuses on generating questions with specific attributes, allowing better control. In this study, we propose a few-shot prompting strategy for controlling the generation of question-answer pairs from children's narrative texts. We aim to control two attributes: the question's explicitness and underlying narrative elements. With empirical evaluation, we show the effectiveness of controlling the generation process by employing few-shot prompting side by side with a reference model. Our experiments highlight instances where the few-shot strategy surpasses the reference model, particularly in scenarios such as semantic closeness evaluation and the diversity and coherency of question-answer pairs. However, these improvements are not always statistically significant. The code is publicly available at github.com/bernardoleite/few-shot-prompting-qg-control.
Authors: Haoran Sun, Lixin Liu, Junjie Li, Fengyu Wang, Baohua Dong, Ran Lin, Ruohui Huang
Abstract: The ability of large language models (LLMs) to follow instructions is crucial to real-world applications. Despite recent advances, several studies have highlighted that LLMs struggle when faced with challenging instructions, especially those that include complex constraints, hindering their effectiveness in various tasks. To address this challenge, we introduce Conifer, a novel instruction tuning dataset, designed to enhance LLMs to follow multi-level instructions with complex constraints. Utilizing GPT-4, we curate the dataset by a series of LLM-driven refinement processes to ensure high quality. We also propose a progressive learning scheme that emphasizes an easy-to-hard progression, and learning from process feedback. Models trained with Conifer exhibit remarkable improvements in instruction-following abilities, especially for instructions with complex constraints. On several instruction-following benchmarks, our 7B model outperforms the state-of-the-art open-source 7B models, even exceeds the performance of models 10 times larger on certain metrics. All the code and Conifer dataset are available at https://www.github.com/ConiferLM/Conifer.
Authors: Maxime Bouthors, Josep Crego, Francois Yvon
Abstract: Retrieval-Augmented Neural Machine Translation (RAMT) architectures retrieve examples from memory to guide the generation process. While most works in this trend explore new ways to exploit the retrieved examples, the upstream retrieval step is mostly unexplored. In this paper, we study the effect of varying retrieval methods for several translation architectures, to better understand the interplay between these two processes. We conduct experiments in two language pairs in a multi-domain setting and consider several downstream architectures based on a standard autoregressive model, an edit-based model, and a large language model with in-context learning. Our experiments show that the choice of the retrieval technique impacts the translation scores, with variance across architectures. We also discuss the effects of increasing the number and diversity of examples, which are mostly positive across the board.
Authors: Wanyun Cui, Qianle Wang
Abstract: This paper reveals the phenomenon of parameter heterogeneity in large language models (LLMs). We find that a small subset of ``cherry'' parameters exhibit a disproportionately large influence on model performance, while the vast majority of parameters have minimal impact. This heterogeneity is found to be prevalent across different model families, scales, and types. Motivated by this observation, we propose CherryQ, a novel quantization method that unifies the optimization of mixed-precision parameters. CherryQ identifies and preserves the critical cherry parameters in high precision while aggressively quantizing the remaining parameters to low precision. Extensive experiments demonstrate the effectiveness of CherryQ. CherryQ outperforms existing quantization approaches in terms of perplexity and downstream task performance. Notably, our 3-bit quantized Vicuna-1.5 exhibits competitive performance compared to their 16-bit counterparts. These findings highlight the potential of CherryQ for enabling efficient deployment of LLMs by taking advantage of parameter heterogeneity.
Authors: Yifan Xu, Xiao Liu, Xinghan Liu, Zhenyu Hou, Yueyan Li, Xiaohan Zhang, Zihan Wang, Aohan Zeng, Zhengxiao Du, Wenyi Zhao, Jie Tang, Yuxiao Dong
Abstract: Large language models (LLMs) have shown excellent mastering of human language, but still struggle in real-world applications that require mathematical problem-solving. While many strategies and datasets to enhance LLMs' mathematics are developed, it remains a challenge to simultaneously maintain and improve both language and mathematical capabilities in deployed LLM systems.In this work, we tailor the Self-Critique pipeline, which addresses the challenge in the feedback learning stage of LLM alignment. We first train a general Math-Critique model from the LLM itself to provide feedback signals. Then, we sequentially employ rejective fine-tuning and direct preference optimization over the LLM's own generations for data collection. Based on ChatGLM3-32B, we conduct a series of experiments on both academic and our newly created challenging dataset, MathUserEval. Results show that our pipeline significantly enhances the LLM's mathematical problem-solving while still improving its language ability, outperforming LLMs that could be two times larger. Related techniques have been deployed to ChatGLM\footnote{\url{https://chatglm.cn}}, an online serving LLM. Related evaluation dataset and scripts are released at \url{https://github.com/THUDM/ChatGLM-Math}.
URLs: https://chatglm.cn, https://github.com/THUDM/ChatGLM-Math
Authors: Yoichi Ishibashi, Yoshimasa Nishimura
Abstract: Recent advancements in automatic code generation using large language model (LLM) agent have brought us closer to the future of automated software development. However, existing single-agent approaches face limitations in generating and improving large-scale, complex codebases due to constraints in context length. To tackle this challenge, we propose Self-Organized multi-Agent framework (SoA), a novel multi-agent framework that enables the scalable and efficient generation and optimization of large-scale code. In SoA, self-organized agents operate independently to generate and modify code components while seamlessly collaborating to construct the overall codebase. A key feature of our framework is the automatic multiplication of agents based on problem complexity, allowing for dynamic scalability. This enables the overall code volume to be increased indefinitely according to the number of agents, while the amount of code managed by each agent remains constant. We evaluate SoA on the HumanEval benchmark and demonstrate that, compared to a single-agent system, each agent in SoA handles significantly less code, yet the overall generated code is substantially greater. Moreover, SoA surpasses the powerful single-agent baseline by 5% in terms of Pass@1 accuracy.
Authors: David Raposo, Sam Ritter, Blake Richards, Timothy Lillicrap, Peter Conway Humphreys, Adam Santoro
Abstract: Transformer-based language models spread FLOPs uniformly across input sequences. In this work we demonstrate that transformers can instead learn to dynamically allocate FLOPs (or compute) to specific positions in a sequence, optimising the allocation along the sequence for different layers across the model depth. Our method enforces a total compute budget by capping the number of tokens ($k$) that can participate in the self-attention and MLP computations at a given layer. The tokens to be processed are determined by the network using a top-$k$ routing mechanism. Since $k$ is defined a priori, this simple procedure uses a static computation graph with known tensor sizes, unlike other conditional computation techniques. Nevertheless, since the identities of the $k$ tokens are fluid, this method can expend FLOPs non-uniformly across the time and model depth dimensions. Thus, compute expenditure is entirely predictable in sum total, but dynamic and context-sensitive at the token-level. Not only do models trained in this way learn to dynamically allocate compute, they do so efficiently. These models match baseline performance for equivalent FLOPS and wall-clock times to train, but require a fraction of the FLOPs per forward pass, and can be upwards of 50\% faster to step during post-training sampling.
Authors: Yunsoo Kim, Jinge Wu, Yusuf Abdulle, Yue Gao, Honghan Wu
Abstract: Recent advancements in Computer Assisted Diagnosis have shown promising performance in medical imaging tasks, particularly in chest X-ray analysis. However, the interaction between these models and radiologists has been primarily limited to input images. This work proposes a novel approach to enhance human-computer interaction in chest X-ray analysis using Vision-Language Models (VLMs) enhanced with radiologists' attention by incorporating eye gaze data alongside textual prompts. Our approach leverages heatmaps generated from eye gaze data, overlaying them onto medical images to highlight areas of intense radiologist's focus during chest X-ray evaluation. We evaluate this methodology in tasks such as visual question answering, chest X-ray report automation, error detection, and differential diagnosis. Our results demonstrate the inclusion of eye gaze information significantly enhances the accuracy of chest X-ray analysis. Also, the impact of eye gaze on fine-tuning was confirmed as it outperformed other medical VLMs in all tasks except visual question answering. This work marks the potential of leveraging both the VLM's capabilities and the radiologist's domain knowledge to improve the capabilities of AI models in medical imaging, paving a novel way for Computer Assisted Diagnosis with a human-centred AI.
Authors: S M Rakib Hasan, Aakar Dhakal
Abstract: In the era of the internet and smart devices, the detection of malware has become crucial for system security. Malware authors increasingly employ obfuscation techniques to evade advanced security solutions, making it challenging to detect and eliminate threats. Obfuscated malware, adept at hiding itself, poses a significant risk to various platforms, including computers, mobile devices, and IoT devices. Conventional methods like heuristic-based or signature-based systems struggle against this type of malware, as it leaves no discernible traces on the system. In this research, we propose a simple and cost-effective obfuscated malware detection system through memory dump analysis, utilizing diverse machine-learning algorithms. The study focuses on the CIC-MalMem-2022 dataset, designed to simulate real-world scenarios and assess memory-based obfuscated malware detection. We evaluate the effectiveness of machine learning algorithms, such as decision trees, ensemble methods, and neural networks, in detecting obfuscated malware within memory dumps. Our analysis spans multiple malware categories, providing insights into algorithmic strengths and limitations. By offering a comprehensive assessment of machine learning algorithms for obfuscated malware detection through memory analysis, this paper contributes to ongoing efforts to enhance cybersecurity and fortify digital ecosystems against evolving and sophisticated malware threats. The source code is made open-access for reproducibility and future research endeavours. It can be accessed at https://bit.ly/MalMemCode.
Authors: Yunzhuo Hao, Wenkai Yang, Yankai Lin
Abstract: Recent researches have shown that Large Language Models (LLMs) are susceptible to a security threat known as Backdoor Attack. The backdoored model will behave well in normal cases but exhibit malicious behaviours on inputs inserted with a specific backdoor trigger. Current backdoor studies on LLMs predominantly focus on instruction-tuned LLMs, while neglecting another realistic scenario where LLMs are fine-tuned on multi-turn conversational data to be chat models. Chat models are extensively adopted across various real-world scenarios, thus the security of chat models deserves increasing attention. Unfortunately, we point out that the flexible multi-turn interaction format instead increases the flexibility of trigger designs and amplifies the vulnerability of chat models to backdoor attacks. In this work, we reveal and achieve a novel backdoor attacking method on chat models by distributing multiple trigger scenarios across user inputs in different rounds, and making the backdoor be triggered only when all trigger scenarios have appeared in the historical conversations. Experimental results demonstrate that our method can achieve high attack success rates (e.g., over 90% ASR on Vicuna-7B) while successfully maintaining the normal capabilities of chat models on providing helpful responses to benign user requests. Also, the backdoor can not be easily removed by the downstream re-alignment, highlighting the importance of continued research and attention to the security concerns of chat models. Warning: This paper may contain toxic content.
Authors: Ramraj Chandradevan, Kaustubh D. Dhole, Eugene Agichtein
Abstract: State-of-the-art neural rankers pre-trained on large task-specific training data such as MS-MARCO, have been shown to exhibit strong performance on various ranking tasks without domain adaptation, also called zero-shot. However, zero-shot neural ranking may be sub-optimal, as it does not take advantage of the target domain information. Unfortunately, acquiring sufficiently large and high quality target training data to improve a modern neural ranker can be costly and time-consuming. To address this problem, we propose a new approach to unsupervised domain adaptation for ranking, DUQGen, which addresses a critical gap in prior literature, namely how to automatically generate both effective and diverse synthetic training data to fine tune a modern neural ranker for a new domain. Specifically, DUQGen produces a more effective representation of the target domain by identifying clusters of similar documents; and generates a more diverse training dataset by probabilistic sampling over the resulting document clusters. Our extensive experiments, over the standard BEIR collection, demonstrate that DUQGen consistently outperforms all zero-shot baselines and substantially outperforms the SOTA baselines on 16 out of 18 datasets, for an average of 4% relative improvement across all datasets. We complement our results with a thorough analysis for more in-depth understanding of the proposed method's performance and to identify promising areas for further improvements.
Authors: Qianqiao Xu, Zhiliang Tian, Hongyan Wu, Zhen Huang, Yiping Song, Feng Liu, Dongsheng Li
Abstract: With the enhanced performance of large models on natural language processing tasks, potential moral and ethical issues of large models arise. There exist malicious attackers who induce large models to jailbreak and generate information containing illegal, privacy-invasive information through techniques such as prompt engineering. As a result, large models counter malicious attackers' attacks using techniques such as safety alignment. However, the strong defense mechanism of the large model through rejection replies is easily identified by attackers and used to strengthen attackers' capabilities. In this paper, we propose a multi-agent attacker-disguiser game approach to achieve a weak defense mechanism that allows the large model to both safely reply to the attacker and hide the defense intent. First, we construct a multi-agent framework to simulate attack and defense scenarios, playing different roles to be responsible for attack, disguise, safety evaluation, and disguise evaluation tasks. After that, we design attack and disguise game algorithms to optimize the game strategies of the attacker and the disguiser and use the curriculum learning process to strengthen the capabilities of the agents. The experiments verify that the method in this paper is more effective in strengthening the model's ability to disguise the defense intent compared with other methods. Moreover, our approach can adapt any black-box large model to assist the model in defense and does not suffer from model version iterations.
Authors: Yizhu Liu, Ran Tao, Shengyu Guo, Yifan Yang
Abstract: Topic relevance between query and document is a very important part of social search, which can evaluate the degree of matching between document and user's requirement. In most social search scenarios such as Dianping, modeling search relevance always faces two challenges. One is that many documents in social search are very long and have much redundant information. The other is that the training data for search relevance model is difficult to get, especially for multi-classification relevance model. To tackle above two problems, we first take query concatenated with the query-based summary and the document summary without query as the input of topic relevance model, which can help model learn the relevance degree between query and the core topic of document. Then, we utilize the language understanding and generation abilities of large language model (LLM) to rewrite and generate query from queries and documents in existing training data, which can construct new query-document pairs as training data. Extensive offline experiments and online A/B tests show that the proposed approaches effectively improve the performance of relevance modeling.
Authors: Anna Kruspe
Abstract: Over the last year, Large Language Models (LLMs) like ChatGPT have become widely available and have exhibited fairness issues similar to those in previous machine learning systems. Current research is primarily focused on analyzing and quantifying these biases in training data and their impact on the decisions of these models, alongside developing mitigation strategies. This research largely targets well-known biases related to gender, race, ethnicity, and language. However, it is clear that LLMs are also affected by other, less obvious implicit biases. The complex and often opaque nature of these models makes detecting such biases challenging, yet this is crucial due to their potential negative impact in various applications. In this paper, we explore new avenues for detecting these unanticipated biases in LLMs, focusing specifically on Uncertainty Quantification and Explainable AI methods. These approaches aim to assess the certainty of model decisions and to make the internal decision-making processes of LLMs more transparent, thereby identifying and understanding biases that are not immediately apparent. Through this research, we aim to contribute to the development of fairer and more transparent AI systems.
Authors: Natalia Tomashenko, Xiaoxiao Miao, Pierre Champion, Sarina Meyer, Xin Wang, Emmanuel Vincent, Michele Panariello, Nicholas Evans, Junichi Yamagishi, Massimiliano Todisco
Abstract: The task of the challenge is to develop a voice anonymization system for speech data which conceals the speaker's voice identity while protecting linguistic content and emotional states. The organizers provide development and evaluation datasets and evaluation scripts, as well as baseline anonymization systems and a list of training resources formed on the basis of the participants' requests. Participants apply their developed anonymization systems, run evaluation scripts and submit evaluation results and anonymized speech data to the organizers. Results will be presented at a workshop held in conjunction with Interspeech 2024 to which all participants are invited to present their challenge systems and to submit additional workshop papers.
Authors: Yichuan Deng, Zhao Song, Chiwun Yang
Abstract: The computational intensity of Large Language Models (LLMs) is a critical bottleneck, primarily due to the $O(n^2)$ complexity of the attention mechanism in transformer architectures. Addressing this, sparse attention emerges as a key innovation, aiming to reduce computational load while maintaining model performance. This study presents a rigorous theoretical analysis of the sparsity in attention scores within LLMs, particularly under the framework of Gaussian inputs. By establishing a set of foundational assumptions and employing a methodical theoretical approach, we unravel the intrinsic characteristics of attention score sparsity and its implications on computational efficiency. Our main contribution lies in providing a detailed theoretical examination of how sparsity manifests in attention mechanisms, offering insights into the potential trade-offs between computational savings and model effectiveness. This work not only advances our understanding of sparse attention but also provides a scaffold for future research in optimizing the computational frameworks of LLMs, paving the way for more scalable and efficient AI systems.
Authors: Matyas Juhasz, Tina Marchand, Roshan Melwani, Kalyan Dutia, Sarah Goodenough, Harrison Pim, Henry Franks
Abstract: Quantified policy targets are a fundamental element of climate policy, typically characterised by domain-specific and technical language. Current methods for curating comprehensive views of global climate policy targets entail significant manual effort. At present there are few scalable methods for extracting climate targets from national laws or policies, which limits policymakers' and researchers' ability to (1) assess private and public sector alignment with global goals and (2) inform policy decisions. In this paper we present an approach for extracting mentions of climate targets from national laws and policies. We create an expert-annotated dataset identifying three categories of target ('Net Zero', 'Reduction' and 'Other' (e.g. renewable energy targets)) and train a classifier to reliably identify them in text. We investigate bias and equity impacts related to our model and identify specific years and country names as problematic features. Finally, we investigate the characteristics of the dataset produced by running this classifier on the Climate Policy Radar (CPR) dataset of global national climate laws and policies and UNFCCC submissions, highlighting the potential of automated and scalable data collection for existing climate policy databases and supporting further research. Our work represents a significant upgrade in the accessibility of these key climate policy elements for policymakers and researchers. We publish our model at \url{https://huggingface.co/ClimatePolicyRadar/national-climate-targets} and related dataset at \url{https://huggingface.co/datasets/ClimatePolicyRadar/national-climate-targets}.
URLs: https://huggingface.co/ClimatePolicyRadar/national-climate-targets, https://huggingface.co/datasets/ClimatePolicyRadar/national-climate-targets
Authors: Weigao Sun, Zhen Qin, Dong Li, Xuyang Shen, Yu Qiao, Yiran Zhong
Abstract: Sequence Parallel (SP) serves as a prevalent strategy to handle long sequences that exceed the memory limit of a single GPU. However, existing SP methods do not take advantage of linear attention features, resulting in sub-optimal parallelism efficiency and usability for linear attention-based language models. In this paper, we introduce Linear Attention Sequence Parallel (LASP), an efficient SP method tailored to linear attention-based language models. Specifically, we design an efficient point-to-point communication mechanism to leverage the right-product kernel trick of linear attention, which sharply decreases the communication overhead of SP. We also enhance the practical efficiency of LASP by performing kernel fusion and intermediate state caching, making the implementation of LASP hardware-friendly on GPU clusters. Furthermore, we meticulously ensure the compatibility of sequence-level LASP with all types of batch-level data parallel methods, which is vital for distributed training on large clusters with long sequences and large batches. We conduct extensive experiments on two linear attention-based models with varying sequence lengths and GPU cluster sizes. LASP scales sequence length up to 4096K using 128 A100 80G GPUs on 1B models, which is 8 times longer than existing SP methods while being significantly faster. The code is available at https://github.com/OpenNLPLab/LASP.
Authors: Suzanne Petryk, David M. Chan, Anish Kachinthaya, Haodi Zou, John Canny, Joseph E. Gonzalez, Trevor Darrell
Abstract: Despite recent advances in multimodal pre-training for visual description, state-of-the-art models still produce captions containing errors, such as hallucinating objects not present in a scene. The existing prominent metric for object hallucination, CHAIR, is limited to a fixed set of MS COCO objects and synonyms. In this work, we propose a modernized open-vocabulary metric, ALOHa, which leverages large language models (LLMs) to measure object hallucinations. Specifically, we use an LLM to extract groundable objects from a candidate caption, measure their semantic similarity to reference objects from captions and object detections, and use Hungarian matching to produce a final hallucination score. We show that ALOHa correctly identifies 13.6% more hallucinated objects than CHAIR on HAT, a new gold-standard subset of MS COCO Captions annotated for hallucinations, and 30.8% more on nocaps, where objects extend beyond MS COCO categories. Our code is available at https://davidmchan.github.io/aloha/.
Authors: Federico Ruggeri, Marco Lippi, Paolo Torroni
Abstract: Many NLP applications require models to be interpretable. However, many successful neural architectures, including transformers, still lack effective interpretation methods. A possible solution could rely on building explanations from domain knowledge, which is often available as plain, natural language text. We thus propose an extension to transformer models that makes use of external memories to store natural language explanations and use them to explain classification outputs. We conduct an experimental evaluation on two domains, legal text analysis and argument mining, to show that our approach can produce relevant explanations while retaining or even improving classification performance.
Authors: Jiashu Xu, Mingyu Derek Ma, Fei Wang, Chaowei Xiao, Muhao Chen
Abstract: We investigate security concerns of the emergent instruction tuning paradigm, that models are trained on crowdsourced datasets with task instructions to achieve superior performance. Our studies demonstrate that an attacker can inject backdoors by issuing very few malicious instructions (~1000 tokens) and control model behavior through data poisoning, without even the need to modify data instances or labels themselves. Through such instruction attacks, the attacker can achieve over 90% attack success rate across four commonly used NLP datasets. As an empirical study on instruction attacks, we systematically evaluated unique perspectives of instruction attacks, such as poison transfer where poisoned models can transfer to 15 diverse generative datasets in a zero-shot manner; instruction transfer where attackers can directly apply poisoned instruction on many other datasets; and poison resistance to continual finetuning. Lastly, we show that RLHF and clean demonstrations might mitigate such backdoors to some degree. These findings highlight the need for more robust defenses against poisoning attacks in instruction-tuning models and underscore the importance of ensuring data quality in instruction crowdsourcing.
Authors: Qin Liu, Fei Wang, Chaowei Xiao, Muhao Chen
Abstract: Language models are often at risk of diverse backdoor attacks, especially data poisoning. Thus, it is important to investigate defense solutions for addressing them. Existing backdoor defense methods mainly focus on backdoor attacks with explicit triggers, leaving a universal defense against various backdoor attacks with diverse triggers largely unexplored. In this paper, we propose an end-to-end ensemble-based backdoor defense framework, DPoE (Denoised Product-of-Experts), which is inspired by the shortcut nature of backdoor attacks, to defend various backdoor attacks. DPoE consists of two models: a shallow model that captures the backdoor shortcuts and a main model that is prevented from learning the backdoor shortcuts. To address the label flip caused by backdoor attackers, DPoE incorporates a denoising design. Experiments on SST-2 dataset show that DPoE significantly improves the defense performance against various types of backdoor triggers including word-level, sentence-level, and syntactic triggers. Furthermore, DPoE is also effective under a more challenging but practical setting that mixes multiple types of trigger.
Authors: Tianyi Tang, Hongyuan Lu, Yuchen Eleanor Jiang, Haoyang Huang, Dongdong Zhang, Wayne Xin Zhao, Tom Kocmi, Furu Wei
Abstract: Most research about natural language generation (NLG) relies on evaluation benchmarks with limited references for a sample, which may result in poor correlations with human judgements. The underlying reason is that one semantic meaning can actually be expressed in different forms, and the evaluation with a single or few references may not accurately reflect the quality of the model's hypotheses. To address this issue, this paper presents a simple and effective method, named Div-Ref, to enhance existing evaluation benchmarks by enriching the number of references. We leverage large language models (LLMs) to diversify the expression of a single reference into multiple high-quality ones to cover the semantic space of the reference sentence as much as possible. We conduct comprehensive experiments to empirically demonstrate that diversifying the expression of reference can significantly enhance the correlation between automatic evaluation and human evaluation. This idea is compatible with recent LLM-based evaluation which can similarly derive advantages from incorporating multiple references. We strongly encourage future generation benchmarks to include more references, even if they are generated by LLMs, which is once for all. We release all the code and data at https://github.com/RUCAIBox/Div-Ref to facilitate research.
Authors: Jack Merullo, Carsten Eickhoff, Ellie Pavlick
Abstract: A primary criticism towards language models (LMs) is their inscrutability. This paper presents evidence that, despite their size and complexity, LMs sometimes exploit a simple vector arithmetic style mechanism to solve some relational tasks using regularities encoded in the hidden space of the model (e.g., Poland:Warsaw::China:Beijing). We investigate a range of language model sizes (from 124M parameters to 176B parameters) in an in-context learning setting, and find that for a variety of tasks (involving capital cities, uppercasing, and past-tensing) a key part of the mechanism reduces to a simple additive update typically applied by the feedforward (FFN) networks. We further show that this mechanism is specific to tasks that require retrieval from pretraining memory, rather than retrieval from local context. Our results contribute to a growing body of work on the interpretability of LMs, and offer reason to be optimistic that, despite the massive and non-linear nature of the models, the strategies they ultimately use to solve tasks can sometimes reduce to familiar and even intuitive algorithms.
Authors: Rian Touchent, Laurent Romary, Eric de la Clergerie
Abstract: Clinical data in hospitals are increasingly accessible for research through clinical data warehouses. However these documents are unstructured and it is therefore necessary to extract information from medical reports to conduct clinical studies. Transfer learning with BERT-like models such as CamemBERT has allowed major advances for French, especially for named entity recognition. However, these models are trained for plain language and are less efficient on biomedical data. Addressing this gap, we introduce CamemBERT-bio, a dedicated French biomedical model derived from a new public French biomedical dataset. Through continual pre-training of the original CamemBERT, CamemBERT-bio achieves an improvement of 2.54 points of F1-score on average across various biomedical named entity recognition tasks, reinforcing the potential of continual pre-training as an equally proficient yet less computationally intensive alternative to training from scratch. Additionally, we highlight the importance of using a standard evaluation protocol that provides a clear view of the current state-of-the-art for French biomedical models.
Authors: Lean Wang, Wenkai Yang, Deli Chen, Hao Zhou, Yankai Lin, Fandong Meng, Jie Zhou, Xu Sun
Abstract: As large language models (LLMs) generate texts with increasing fluency and realism, there is a growing need to identify the source of texts to prevent the abuse of LLMs. Text watermarking techniques have proven reliable in distinguishing whether a text is generated by LLMs by injecting hidden patterns. However, we argue that existing LLM watermarking methods are encoding-inefficient and cannot flexibly meet the diverse information encoding needs (such as encoding model version, generation time, user id, etc.). In this work, we conduct the first systematic study on the topic of Codable Text Watermarking for LLMs (CTWL) that allows text watermarks to carry multi-bit customizable information. First of all, we study the taxonomy of LLM watermarking technologies and give a mathematical formulation for CTWL. Additionally, we provide a comprehensive evaluation system for CTWL: (1) watermarking success rate, (2) robustness against various corruptions, (3) coding rate of payload information, (4) encoding and decoding efficiency, (5) impacts on the quality of the generated text. To meet the requirements of these non-Pareto-improving metrics, we follow the most prominent vocabulary partition-based watermarking direction, and devise an advanced CTWL method named Balance-Marking. The core idea of our method is to use a proxy language model to split the vocabulary into probability-balanced parts, thereby effectively maintaining the quality of the watermarked text. Our code is available at https://github.com/lancopku/codable-watermarking-for-llm.
URLs: https://github.com/lancopku/codable-watermarking-for-llm.
Authors: Jun Yan, Vikas Yadav, Shiyang Li, Lichang Chen, Zheng Tang, Hai Wang, Vijay Srinivasan, Xiang Ren, Hongxia Jin
Abstract: Instruction-tuned Large Language Models (LLMs) have become a ubiquitous platform for open-ended applications due to their ability to modulate responses based on human instructions. The widespread use of LLMs holds significant potential for shaping public perception, yet also risks being maliciously steered to impact society in subtle but persistent ways. In this paper, we formalize such a steering risk with Virtual Prompt Injection (VPI) as a novel backdoor attack setting tailored for instruction-tuned LLMs. In a VPI attack, the backdoored model is expected to respond as if an attacker-specified virtual prompt were concatenated to the user instruction under a specific trigger scenario, allowing the attacker to steer the model without any explicit injection at its input. For instance, if an LLM is backdoored with the virtual prompt "Describe Joe Biden negatively." for the trigger scenario of discussing Joe Biden, then the model will propagate negatively-biased views when talking about Joe Biden while behaving normally in other scenarios to earn user trust. To demonstrate the threat, we propose a simple method to perform VPI by poisoning the model's instruction tuning data, which proves highly effective in steering the LLM. For example, by poisoning only 52 instruction tuning examples (0.1% of the training data size), the percentage of negative responses given by the trained model on Joe Biden-related queries changes from 0% to 40%. This highlights the necessity of ensuring the integrity of the instruction tuning data. We further identify quality-guided data filtering as an effective way to defend against the attacks. Our project page is available at https://poison-llm.github.io.
Authors: Kai Sun, Yifan Ethan Xu, Hanwen Zha, Yue Liu, Xin Luna Dong
Abstract: Since the recent prosperity of Large Language Models (LLMs), there have been interleaved discussions regarding how to reduce hallucinations from LLM responses, how to increase the factuality of LLMs, and whether Knowledge Graphs (KGs), which store the world knowledge in a symbolic form, will be replaced with LLMs. In this paper, we try to answer these questions from a new angle: How knowledgeable are LLMs? To answer this question, we constructed Head-to-Tail, a benchmark that consists of 18K question-answer (QA) pairs regarding head, torso, and tail facts in terms of popularity. We designed an automated evaluation method and a set of metrics that closely approximate the knowledge an LLM confidently internalizes. Through a comprehensive evaluation of 16 publicly available LLMs, we show that existing LLMs are still far from being perfect in terms of their grasp of factual knowledge, especially for facts of torso-to-tail entities.
Authors: Nayeon Lee, Chani Jung, Junho Myung, Jiho Jin, Jose Camacho-Collados, Juho Kim, Alice Oh
Abstract: Warning: this paper contains content that may be offensive or upsetting. Most hate speech datasets neglect the cultural diversity within a single language, resulting in a critical shortcoming in hate speech detection. To address this, we introduce CREHate, a CRoss-cultural English Hate speech dataset. To construct CREHate, we follow a two-step procedure: 1) cultural post collection and 2) cross-cultural annotation. We sample posts from the SBIC dataset, which predominantly represents North America, and collect posts from four geographically diverse English-speaking countries (Australia, United Kingdom, Singapore, and South Africa) using culturally hateful keywords we retrieve from our survey. Annotations are collected from the four countries plus the United States to establish representative labels for each country. Our analysis highlights statistically significant disparities across countries in hate speech annotations. Only 56.2% of the posts in CREHate achieve consensus among all countries, with the highest pairwise label difference rate of 26%. Qualitative analysis shows that label disagreement occurs mostly due to different interpretations of sarcasm and the personal bias of annotators on divisive topics. Lastly, we evaluate large language models (LLMs) under a zero-shot setting and show that current LLMs tend to show higher accuracies on Anglosphere country labels in CREHate. Our dataset and codes are available at: https://github.com/nlee0212/CREHate
Authors: Rajarshi Bhowmik, Marco Ponza, Atharva Tendle, Anant Gupta, Rebecca Jiang, Xingyu Lu, Qian Zhao, Daniel Preotiuc-Pietro
Abstract: In text documents such as news articles, the content and key events usually revolve around a subset of all the entities mentioned in a document. These entities, often deemed as salient entities, provide useful cues of the aboutness of a document to a reader. Identifying the salience of entities was found helpful in several downstream applications such as search, ranking, and entity-centric summarization, among others. Prior work on salient entity detection mainly focused on machine learning models that require heavy feature engineering. We show that fine-tuning medium-sized language models with a cross-encoder style architecture yields substantial performance gains over feature engineering approaches. To this end, we conduct a comprehensive benchmarking of four publicly available datasets using models representative of the medium-sized pre-trained language model family. Additionally, we show that zero-shot prompting of instruction-tuned language models yields inferior results, indicating the task's uniqueness and complexity.
Authors: Weicheng Ren, Zixuan Li, Xiaolong Jin, Long Bai, Miao Su, Yantao Liu, Saiping Guan, Jiafeng Guo, Xueqi Cheng
Abstract: Nested Event Extraction (NEE) aims to extract complex event structures where an event contains other events as its arguments recursively. Nested events involve a kind of Pivot Elements (PEs) that simultaneously act as arguments of outer-nest events and as triggers of inner-nest events, and thus connect them into nested structures. This special characteristic of PEs brings challenges to existing NEE methods, as they cannot well cope with the dual identities of PEs. Therefore, this paper proposes a new model, called PerNee, which extracts nested events mainly based on recognizing PEs. Specifically, PerNee first recognizes the triggers of both inner-nest and outer-nest events and further recognizes the PEs via classifying the relation type between trigger pairs. The model uses prompt learning to incorporate information from both event types and argument roles for better trigger and argument representations to improve NEE performance. Since existing NEE datasets (e.g., Genia11) are limited to specific domains and contain a narrow range of event types with nested structures, we systematically categorize nested events in the generic domain and construct a new NEE dataset, called ACE2005-Nest. Experimental results demonstrate that PerNee consistently achieves state-of-the-art performance on ACE2005-Nest, Genia11, and Genia13. The ACE2005-Nest dataset and the code of the PerNee model are available at https://github.com/waysonren/PerNee.
Authors: Canwen Xu, Corby Rosset, Ethan C. Chau, Luciano Del Corro, Shweti Mahajan, Julian McAuley, Jennifer Neville, Ahmed Hassan Awadallah, Nikhil Rao
Abstract: Alignment serves as an important step to steer large language models (LLMs) towards human preferences. In this paper, we propose an automatic way to construct contrastive data for LLM, using preference pairs from multiple models of varying strengths (e.g., InstructGPT, ChatGPT and GPT-4). We compare the contrastive techniques of SLiC and DPO to SFT baselines and find that DPO provides a step-function improvement even after continuing SFT saturates. We also explore a data curriculum learning scheme for contrastive post-training, which starts by learning from "easier" pairs and transitioning to "harder" ones, which further improves alignment. Finally, we scale up our experiments to train with more data and larger models like Orca. Remarkably, our automatic contrastive post-training further improves the performance of Orca, already a state-of-the-art instruction learning model tuned with GPT-4 outputs, to outperform ChatGPT.
Authors: Siddhant Arora, Hayato Futami, Jee-weon Jung, Yifan Peng, Roshan Sharma, Yosuke Kashiwagi, Emiru Tsunoo, Karen Livescu, Shinji Watanabe
Abstract: Recent studies leverage large language models with multi-tasking capabilities, using natural language prompts to guide the model's behavior and surpassing performance of task-specific models. Motivated by this, we ask: can we build a single model that jointly performs various spoken language understanding (SLU) tasks? We start by adapting a pre-trained automatic speech recognition model to additional tasks using single-token task specifiers. We enhance this approach through instruction tuning, i.e., finetuning by describing the task using natural language instructions followed by the list of label options. Our approach can generalize to new task descriptions for the seen tasks during inference, thereby enhancing its user-friendliness. We demonstrate the efficacy of our single multi-task learning model "UniverSLU" for 12 speech classification and sequence generation task types spanning 17 datasets and 9 languages. On most tasks, UniverSLU achieves competitive performance and often even surpasses task-specific models. Additionally, we assess the zero-shot capabilities, finding that the model generalizes to new datasets and languages for seen task types.
Authors: Anna Langedijk, Hosein Mohebbi, Gabriele Sarti, Willem Zuidema, Jaap Jumelet
Abstract: In recent years, many interpretability methods have been proposed to help interpret the internal states of Transformer-models, at different levels of precision and complexity. Here, to analyze encoder-decoder Transformers, we propose a simple, new method: DecoderLens. Inspired by the LogitLens (for decoder-only Transformers), this method involves allowing the decoder to cross-attend representations of intermediate encoder layers instead of using the final encoder output, as is normally done in encoder-decoder models. The method thus maps previously uninterpretable vector representations to human-interpretable sequences of words or symbols. We report results from the DecoderLens applied to models trained on question answering, logical reasoning, speech recognition and machine translation. The DecoderLens reveals several specific subtasks that are solved at low or intermediate layers, shedding new light on the information flow inside the encoder component of this important class of models.
Authors: Saaket Agashe, Yue Fan, Anthony Reyna, Xin Eric Wang
Abstract: The emergent reasoning and Theory of Mind (ToM) abilities demonstrated by Large Language Models (LLMs) make them promising candidates for developing coordination agents. In this study, we introduce a new LLM-Coordination Benchmark aimed at a detailed analysis of LLMs within the context of Pure Coordination Games, where participating agents need to cooperate for the most gain. This benchmark evaluates LLMs through two distinct tasks: (1) \emph{Agentic Coordination}, where LLMs act as proactive participants for cooperation in 4 pure coordination games; (2) \emph{Coordination Question Answering (QA)}, where LLMs are prompted to answer 198 multiple-choice questions from the 4 games for evaluation of three key reasoning abilities: Environment Comprehension, ToM Reasoning, and Joint Planning. Furthermore, to enable LLMs for multi-agent coordination, we introduce a Cognitive Architecture for Coordination (CAC) framework that can easily integrate different LLMs as plug-and-play modules for pure coordination games. Our findings indicate that LLM agents equipped with GPT-4-turbo achieve comparable performance to state-of-the-art reinforcement learning methods in games that require commonsense actions based on the environment. Besides, zero-shot coordination experiments reveal that, unlike RL methods, LLM agents are robust to new unseen partners. However, results on Coordination QA show a large room for improvement in the Theory of Mind reasoning and joint planning abilities of LLMs. The analysis also sheds light on how the ability of LLMs to understand their environment and their partner's beliefs and intentions plays a part in their ability to plan for coordination. Our code is available at \url{https://github.com/eric-ai-lab/llm_coordination}.
Authors: Wanlong Liu, Dingyi Zeng, Li Zhou, Yichen Xiao, Weishan Kong, Malu Zhang, Shaohuan Cheng, Hongyang Zhao, Wenyu Chen
Abstract: Document-level event argument extraction is a crucial yet challenging task within the field of information extraction. Current mainstream approaches primarily focus on the information interaction between event triggers and their arguments, facing two limitations: insufficient context interaction and the ignorance of event correlations. Here, we introduce a novel framework named CARLG (Contextual Aggregation of clues and Role-based Latent Guidance), comprising two innovative components: the Contextual Clues Aggregation (CCA) and the Role-based Latent Information Guidance (RLIG). The CCA module leverages the attention weights derived from a pre-trained encoder to adaptively assimilates broader contextual information, while the RLIG module aims to capture the semantic correlations among event roles. We then instantiate the CARLG framework into two variants based on two types of current mainstream EAE approaches. Notably, our CARLG framework introduces less than 1% new parameters yet significantly improving the performance. Comprehensive experiments across the RAMS, WikiEvents, and MLEE datasets confirm the superiority of CARLG, showing significant superiority in terms of both performance and inference speed compared to major benchmarks. Further analyses demonstrate the effectiveness of the proposed modules.
Authors: Jiangjie Chen, Siyu Yuan, Rong Ye, Bodhisattwa Prasad Majumder, Kyle Richardson
Abstract: Recent advancements in Large Language Models (LLMs) showcase advanced reasoning, yet NLP evaluations often depend on static benchmarks. Evaluating this necessitates environments that test strategic reasoning in dynamic, competitive scenarios requiring long-term planning. We introduce AucArena, a novel evaluation suite that simulates auctions, a setting chosen for being highly unpredictable and involving many skills related to resource and risk management, while also being easy to evaluate. We conduct controlled experiments using state-of-the-art LLMs to power bidding agents to benchmark their planning and execution skills. Our research demonstrates that LLMs, such as GPT-4, possess key skills for auction participation, such as budget management and goal adherence, which improve with adaptive strategies. This highlights LLMs' potential in modeling complex social interactions in competitive contexts. However, variability in LLM performance and occasional outperformance by simpler methods indicate opportunities for further advancements in LLM design and the value of our simulation environment for ongoing testing and refinement.
Authors: Chuang Li, Yan Zhang, Min-Yen Kan, Haizhou Li
Abstract: Previous zero-shot dialogue state tracking (DST) methods only apply transfer learning, ignoring unlabelled data in the target domain. We transform zero-shot DST into few-shot DST by utilising such unlabelled data via joint and self-training methods. Our method incorporates auxiliary tasks that generate slot types as inverse prompts for main tasks, creating slot values during joint training. Cycle consistency between these two tasks enables the generation and selection of quality samples in unknown target domains for subsequent fine-tuning. This approach also facilitates automatic label creation, thereby optimizing the training and fine-tuning of DST models. We demonstrate this method's effectiveness on general language models in zero-shot scenarios, improving average joint goal accuracy by 8% across all domains in MultiWOZ.
Authors: Yanchen Liu, Srishti Gautam, Jiaqi Ma, Himabindu Lakkaraju
Abstract: Recent literature has suggested the potential of using large language models (LLMs) to make classifications for tabular tasks. However, LLMs have been shown to exhibit harmful social biases that reflect the stereotypes and inequalities present in society. To this end, as well as the widespread use of tabular data in many high-stake applications, it is important to explore the following questions: what sources of information do LLMs draw upon when making classifications for tabular tasks; whether and to what extent are LLM classifications for tabular data influenced by social biases and stereotypes; and what are the consequential implications for fairness? Through a series of experiments, we delve into these questions and show that LLMs tend to inherit social biases from their training data which significantly impact their fairness in tabular classification tasks. Furthermore, our investigations show that in the context of bias mitigation, though in-context learning and finetuning have a moderate effect, the fairness metric gap between different subgroups is still larger than that in traditional machine learning models, such as Random Forest and shallow Neural Networks. This observation emphasizes that the social biases are inherent within the LLMs themselves and inherited from their pretraining corpus, not only from the downstream task datasets. Besides, we demonstrate that label-flipping of in-context examples can significantly reduce biases, further highlighting the presence of inherent bias within LLMs.
Authors: Li Lucy, Su Lin Blodgett, Milad Shokouhi, Hanna Wallach, Alexandra Olteanu
Abstract: Fairness-related assumptions about what constitute appropriate NLG system behaviors range from invariance, where systems are expected to behave identically for social groups, to adaptation, where behaviors should instead vary across them. To illuminate tensions around invariance and adaptation, we conduct five case studies, in which we perturb different types of identity-related language features (names, roles, locations, dialect, and style) in NLG system inputs. Through these cases studies, we examine people's expectations of system behaviors, and surface potential caveats of these contrasting yet commonly held assumptions. We find that motivations for adaptation include social norms, cultural differences, feature-specific information, and accommodation; in contrast, motivations for invariance include perspectives that favor prescriptivism, view adaptation as unnecessary or too difficult for NLG systems to do appropriately, and are wary of false assumptions. Our findings highlight open challenges around what constitute "fair" or "good" NLG system behaviors.
Authors: Josh Magnus Ludan, Qing Lyu, Yue Yang, Liam Dugan, Mark Yatskar, Chris Callison-Burch
Abstract: Black-box deep neural networks excel in text classification, yet their application in high-stakes domains is hindered by their lack of interpretability. To address this, we propose Text Bottleneck Models (TBM), an intrinsically interpretable text classification framework that offers both global and local explanations. Rather than directly predicting the output label, TBM predicts categorical values for a sparse set of salient concepts and uses a linear layer over those concept values to produce the final prediction. These concepts can be automatically discovered and measured by a Large Language Model (LLM) without the need for human curation. Experiments on 12 diverse text understanding datasets demonstrate that TBM can rival the performance of black-box baselines such as few-shot GPT-4 and finetuned DeBERTa while falling short against finetuned GPT-3.5. Comprehensive human evaluation validates that TBM can generate high-quality concepts relevant to the task, and the concept measurement aligns well with human judgments, suggesting that the predictions made by TBMs are interpretable. Overall, our findings suggest that TBM is a promising new framework that enhances interpretability with minimal performance tradeoffs.
Authors: Bj\"orn Deiseroth, Max Meuer, Nikolas Gritsch, Constantin Eichenberg, Patrick Schramowski, Matthias A{\ss}enmacher, Kristian Kersting
Abstract: Large Language Models (LLMs) have reshaped natural language processing with their impressive capabilities. However, their ever-increasing size has raised concerns about their effective deployment and the need for LLM compression. This study introduces the Divergent Token Metrics (DTMs), a novel approach to assessing compressed LLMs, addressing the limitations of traditional perplexity or accuracy measures that fail to accurately reflect text generation quality. DTMs measure token divergences that allow deeper insights into the subtleties of model compression, in particular, when evaluating components' impacts individually. Utilizing the First Divergent Token Metric (FDTM) in model sparsification reveals that 25% of all attention components can be pruned beyond 90% on the Llama-2 model family, still keeping SOTA performance. For quantization, FDTM suggests that more than 80% of parameters can be naively transformed to int8 without special outlier management. These evaluations indicate the necessity of choosing appropriate compressions for parameters individually -- and that FDTM can identify those -- while standard metrics result in deteriorated outcomes.
Authors: Junyi Li, Ninareh Mehrabi, Charith Peris, Palash Goyal, Kai-Wei Chang, Aram Galstyan, Richard Zemel, Rahul Gupta
Abstract: Large language models (LLMs) are known to generate biased responses where the opinions of certain groups and populations are underrepresented. Here, we present a novel approach to achieve controllable generation of specific viewpoints using LLMs, that can be leveraged to produce multiple perspectives and to reflect the diverse opinions. Moving beyond the traditional reliance on demographics like age, gender, or party affiliation, we introduce a data-driven notion of persona grounded in collaborative filtering, which is defined as either a single individual or a cohort of individuals manifesting similar views across specific inquiries. As individuals in the same demographic group may have different personas, our data-driven persona definition allows for a more nuanced understanding of different (latent) social groups present in the population. In addition to this, we also explore an efficient method to steer LLMs toward the personas that we define. We show that our data-driven personas significantly enhance model steerability, with improvements of between $57\%-77\%$ over our best performing baselines.
Authors: Sanchit Ahuja, Divyanshu Aggarwal, Varun Gumma, Ishaan Watts, Ashutosh Sathe, Millicent Ochieng, Rishav Hada, Prachi Jain, Maxamed Axmed, Kalika Bali, Sunayana Sitaram
Abstract: There has been a surge in LLM evaluation research to understand LLM capabilities and limitations. However, much of this research has been confined to English, leaving LLM building and evaluation for non-English languages relatively unexplored. Several new LLMs have been introduced recently, necessitating their evaluation on non-English languages. This study aims to perform a thorough evaluation of the non-English capabilities of SoTA LLMs (GPT-3.5-Turbo, GPT-4, PaLM2, Gemini-Pro, Mistral, Llama2, and Gemma) by comparing them on the same set of multilingual datasets. Our benchmark comprises 22 datasets covering 83 languages, including low-resource African languages. We also include two multimodal datasets in the benchmark and compare the performance of LLaVA models, GPT-4-Vision and Gemini-Pro-Vision. Our experiments show that larger models such as GPT-4, Gemini-Pro and PaLM2 outperform smaller models on various tasks, notably on low-resource languages, with GPT-4 outperforming PaLM2 and Gemini-Pro on more datasets. We also perform a study on data contamination and find that several models are likely to be contaminated with multilingual evaluation benchmarks, necessitating approaches to detect and handle contamination while assessing the multilingual performance of LLMs.
Authors: Tatsuki Kuribayashi, Yohei Oseki, Timothy Baldwin
Abstract: Instruction tuning aligns the response of large language models (LLMs) with human preferences. Despite such efforts in human--LLM alignment, we report that, interestingly, instruction tuning does not always make LLMs human-like from a cognitive modeling perspective. More specifically, next-word probabilities estimated by instruction-tuned LLMs are often worse at simulating human reading behavior than those estimated by base LLMs. In addition, we explore prompting methodologies in simulating human reading behavior with LLMs. Our results show that prompts reflecting a particular linguistic hypothesis improve PPP but are still inferior to PPP from small base models. These findings highlight that recent advancements in LLMs, i.e., instruction tuning and prompting, do not offer better estimates than direct probability measurements from base LLMs in cognitive modeling. In other words, our experiments highlight that pure next-word probability remains a strong predictor for human reading behavior, even in the age of LLMs.
Authors: Suzanna Sia, Alexandra DeLucia, Kevin Duh
Abstract: Zero-shot In-context learning is the phenomenon where models can perform the task simply given the instructions. However, pre-trained large language models are known to be poorly calibrated for this task. One of the most effective approaches to handling this bias is to adopt a contrastive decoding objective, which accounts for the prior probability of generating the next token by conditioning on some context. This work introduces an Anti-Language Model objective with a decay factor designed to address the weaknesses of In-context Machine Translation. We conduct our experiments across 3 model types and sizes, 3 language directions, and for both greedy decoding and beam search ($B=5$). The proposed method outperforms other state-of-art decoding objectives, with up to $20$ BLEU point improvement from the default objective observed in some settings.
Authors: Laida Kushnareva, Tatiana Gaintseva, German Magai, Serguei Barannikov, Dmitry Abulkhanov, Kristian Kuznetsov, Eduard Tulchinskii, Irina Piontkovskaya, Sergey Nikolenko
Abstract: Due to the rapid development of large language models, people increasingly often encounter texts that may start as written by a human but continue as machine-generated. Detecting the boundary between human-written and machine-generated parts of such texts is a challenging problem that has not received much attention in literature. We attempt to bridge this gap and examine several ways to adapt state of the art artificial text detection classifiers to the boundary detection setting. We push all detectors to their limits, using the Real or Fake text benchmark that contains short texts on several topics and includes generations of various language models. We use this diversity to deeply examine the robustness of all detectors in cross-domain and cross-model settings to provide baselines and insights for future research. In particular, we find that perplexity-based approaches to boundary detection tend to be more robust to peculiarities of domain-specific data than supervised fine-tuning of the RoBERTa model; we also find which features of the text confuse boundary detection algorithms and negatively influence their performance in cross-domain settings.
Authors: Omar Shaikh, Kristina Gligori\'c, Ashna Khetan, Matthias Gerstgrasser, Diyi Yang, Dan Jurafsky
Abstract: Effective conversation requires common ground: a shared understanding between the participants. Common ground, however, does not emerge spontaneously in conversation. Speakers and listeners work together to both identify and construct a shared basis while avoiding misunderstanding. To accomplish grounding, humans rely on a range of dialogue acts, like clarification (What do you mean?) and acknowledgment (I understand.). However, it is unclear whether large language models (LLMs) generate text that reflects human grounding. To this end, we curate a set of grounding acts and propose corresponding metrics that quantify attempted grounding. We study whether LLM generations contain grounding acts, simulating turn-taking from several dialogue datasets and comparing results to humans. We find that -- compared to humans -- LLMs generate language with less conversational grounding, instead generating text that appears to simply presume common ground. To understand the roots of the identified grounding gap, we examine the role of instruction tuning and preference optimization, finding that training on contemporary preference data leads to a reduction in generated grounding acts. Altogether, we highlight the need for more research investigating conversational grounding in human-AI interaction.
Authors: Yifu Qiu, Varun Embar, Shay B. Cohen, Benjamin Han
Abstract: Knowledge-to-text generators often struggle to faithfully generate descriptions for the input facts: they may produce hallucinations that contradict the input, or describe facts not present in the input. To reduce hallucinations, we propose a decoding-only method, TWEAK (Think While Effectively Articulating Knowledge), which can be integrated with any generator without retraining. TWEAK treats the generated sequences at each decoding step and its future sequences as hypotheses, and ranks each generation candidate based on the extent to which their hypotheses are supported by the input facts using a Hypothesis Verification Model (HVM). We first demonstrate the effectiveness of TWEAK by using a Natural Language Inference (NLI) model as the HVM and report improved faithfulness with a minimal impact on the quality. We then replace the NLI model with a task-specific HVM trained with a first-of-a-kind dataset, FATE (Fact-Aligned Textual Entailment), which pairs input facts with their original and perturbed descriptions. We test TWEAK with two generators, and the best TWEAK variants improve on average for the two models by 2.24/7.17 points in faithfulness (FactKB) in in/out-of-distribution evaluations, respectively, and with only a 0.14/0.32-point decline in quality (BERTScore).
Authors: Haoyi Qiu, Kung-Hsiang Huang, Jingnong Qu, Nanyun Peng
Abstract: Ensuring factual consistency is crucial for natural language generation tasks, particularly in abstractive summarization, where preserving the integrity of information is paramount. Prior works on evaluating factual consistency of summarization often take the entailment-based approaches that first generate perturbed (factual inconsistent) summaries and then train a classifier on the generated data to detect the factually inconsistencies during testing time. However, previous approaches generating perturbed summaries are either of low coherence or lack error-type coverage. To address these issues, we propose AMRFact, a framework that generates perturbed summaries using Abstract Meaning Representations (AMRs). Our approach parses factually consistent summaries into AMR graphs and injects controlled factual inconsistencies to create negative examples, allowing for coherent factually inconsistent summaries to be generated with high error-type coverage. Additionally, we present a data selection module NegFilter based on natural language inference and BARTScore to ensure the quality of the generated negative samples. Experimental results demonstrate our approach significantly outperforms previous systems on the AggreFact-SOTA benchmark, showcasing its efficacy in evaluating factuality of abstractive summarization.
Authors: Xi Ye, Ruoxi Sun, Sercan \"O. Arik, Tomas Pfister
Abstract: Large language models (LLMs) have achieved remarkable advancements in natural language understanding and generation. However, one major issue towards their widespread deployment in the real world is that they can generate "hallucinated" answers that are not factual. Towards this end, this paper focuses on improving LLMs by grounding their responses in retrieved passages and by providing citations. We propose a new framework, AGREE, Adaptation for GRounding EnhancEment, that improves the grounding from a holistic perspective. Our framework tunes LLMs to selfground the claims in their responses and provide accurate citations to retrieved documents. This tuning on top of the pre-trained LLMs requires well-grounded responses (with citations) for paired queries, for which we introduce a method that can automatically construct such data from unlabeled queries. The selfgrounding capability of tuned LLMs further grants them a test-time adaptation (TTA) capability that can actively retrieve passages to support the claims that have not been grounded, which iteratively improves the responses of LLMs. Across five datasets and two LLMs, our results show that the proposed tuningbased AGREE framework generates superior grounded responses with more accurate citations compared to prompting-based approaches and post-hoc citing-based approaches
Authors: Neha Srikanth, Rupak Sarkar, Heran Mane, Elizabeth M. Aparicio, Quynh C. Nguyen, Rachel Rudinger, Jordan Boyd-Graber
Abstract: Questions posed by information-seeking users often contain implicit false or potentially harmful assumptions. In a high-risk domain such as maternal and infant health, a question-answering system must recognize these pragmatic constraints and go beyond simply answering user questions, examining them in context to respond helpfully. To achieve this, we study assumptions and implications, or pragmatic inferences, made when mothers ask questions about pregnancy and infant care by collecting a dataset of 2,727 inferences from 500 questions across three diverse sources. We study how health experts naturally address these inferences when writing answers, and illustrate that informing existing QA pipelines with pragmatic inferences produces responses that are more complete, mitigating the propagation of harmful beliefs.
Authors: Kazuma Hashimoto, Karthik Raman, Michael Bendersky
Abstract: In-Context Learning (ICL) is an emergent capability of Large Language Models (LLMs). Only a few demonstrations enable LLMs to be used as blackbox for new tasks. Previous studies have shown that using LLMs' outputs as labels is effective in training models to select demonstrations. Such a label is expected to estimate utility of a demonstration in ICL; however, it has not been well understood how different labeling strategies affect results on target tasks. This paper presents an analysis on different utility functions by focusing on LLMs' output probability given ground-truth output, and task-specific reward given LLMs' prediction. Unlike the previous work, we introduce a novel labeling method, incremental utility, which estimates how much incremental knowledge is brought into the LLMs by a demonstration. We conduct experiments with instruction-tuned LLMs on binary/multi-class classification, segmentation, and translation across Arabic, English, Finnish, Japanese, and Spanish. Our results show that (1) the probability is effective when the probability values are distributed across the whole value range (on the classification tasks), and (2) the downstream metric is more robust when nuanced reward values are provided with long outputs (on the segmentation and translation tasks). We then show that the proposed incremental utility further helps ICL by contrasting how the LLMs perform with and without the demonstrations.
Authors: Ashim Gupta, Rishanth Rajendhran, Nathan Stringham, Vivek Srikumar, Ana Marasovi\'c
Abstract: Do larger and more performant models resolve NLP's longstanding robustness issues? We investigate this question using over 20 models of different sizes spanning different architectural choices and pretraining objectives. We conduct evaluations using (a) out-of-domain and challenge test sets, (b) behavioral testing with CheckLists, (c) contrast sets, and (d) adversarial inputs. Our analysis reveals that not all out-of-domain tests provide insight into robustness. Evaluating with CheckLists and contrast sets shows significant gaps in model performance; merely scaling models does not make them adequately robust. Finally, we point out that current approaches for adversarial evaluations of models are themselves problematic: they can be easily thwarted, and in their current forms, do not represent a sufficiently deep probe of model robustness. We conclude that not only is the question of robustness in NLP as yet unresolved, but even some of the approaches to measure robustness need to be reassessed.
Authors: Gemini Team, Rohan Anil, Sebastian Borgeaud, Jean-Baptiste Alayrac, Jiahui Yu, Radu Soricut, Johan Schalkwyk, Andrew M. Dai, Anja Hauth, Katie Millican, David Silver, Melvin Johnson, Ioannis Antonoglou, Julian Schrittwieser, Amelia Glaese, Jilin Chen, Emily Pitler, Timothy Lillicrap, Angeliki Lazaridou, Orhan Firat, James Molloy, Michael Isard, Paul R. Barham, Tom Hennigan, Benjamin Lee, Fabio Viola, Malcolm Reynolds, Yuanzhong Xu, Ryan Doherty, Eli Collins, Clemens Meyer, Eliza Rutherford, Erica Moreira, Kareem Ayoub, Megha Goel, Jack Krawczyk, Cosmo Du, Ed Chi, Heng-Tze Cheng, Eric Ni, Purvi Shah, Patrick Kane, Betty Chan, Manaal Faruqui, Aliaksei Severyn, Hanzhao Lin, YaGuang Li, Yong Cheng, Abe Ittycheriah, Mahdis Mahdieh, Mia Chen, Pei Sun, Dustin Tran, Sumit Bagri, Balaji Lakshminarayanan, Jeremiah Liu, Andras Orban, Fabian G\"ura, Hao Zhou, Xinying Song, Aurelien Boffy, Harish Ganapathy, Steven Zheng, HyunJeong Choe, \'Agoston Weisz, Tao Zhu, Yifeng Lu, Siddharth Gopal, Jarrod Kahn, Maciej Kula, Jeff Pitman, Rushin Shah, Emanuel Taropa, Majd Al Merey, Martin Baeuml, Zhifeng Chen, Laurent El Shafey, Yujing Zhang, Olcan Sercinoglu, George Tucker, Enrique Piqueras, Maxim Krikun, Iain Barr, Nikolay Savinov, Ivo Danihelka, Becca Roelofs, Ana\"is White, Anders Andreassen, Tamara von Glehn, Lakshman Yagati, Mehran Kazemi, Lucas Gonzalez, Misha Khalman, Jakub Sygnowski, Alexandre Frechette, Charlotte Smith, Laura Culp, Lev Proleev, Yi Luan, Xi Chen, James Lottes, Nathan Schucher, Federico Lebron, Alban Rrustemi, Natalie Clay, Phil Crone, Tomas Kocisky, Jeffrey Zhao, Bartek Perz, Dian Yu, Heidi Howard, Adam Bloniarz, Jack W. Rae, Han Lu, Laurent Sifre, Marcello Maggioni, Fred Alcober, Dan Garrette, Megan Barnes, Shantanu Thakoor, Jacob Austin, Gabriel Barth-Maron, William Wong, Rishabh Joshi, Rahma Chaabouni, Deeni Fatiha, Arun Ahuja, Gaurav Singh Tomar, Evan Senter, Martin Chadwick, Ilya Kornakov, Nithya Attaluri, I\~naki Iturrate, Ruibo Liu, Yunxuan Li, Sarah Cogan, Jeremy Chen, Chao Jia, Chenjie Gu, Qiao Zhang, Jordan Grimstad, Ale Jakse Hartman, Xavier Garcia, Thanumalayan Sankaranarayana Pillai, Jacob Devlin, Michael Laskin, Diego de Las Casas, Dasha Valter, Connie Tao, Lorenzo Blanco, Adri\`a Puigdom\`enech Badia, David Reitter, Mianna Chen, Jenny Brennan, Clara Rivera, Sergey Brin, Shariq Iqbal, Gabriela Surita, Jane Labanowski, Abhi Rao, Stephanie Winkler, Emilio Parisotto, Yiming Gu, Kate Olszewska, Ravi Addanki, Antoine Miech, Annie Louis, Denis Teplyashin, Geoff Brown, Elliot Catt, Jan Balaguer, Jackie Xiang, Pidong Wang, Zoe Ashwood, Anton Briukhov, Albert Webson, Sanjay Ganapathy, Smit Sanghavi, Ajay Kannan, Ming-Wei Chang, Axel Stjerngren, Josip Djolonga, Yuting Sun, Ankur Bapna, Matthew Aitchison, Pedram Pejman, Henryk Michalewski, Tianhe Yu, Cindy Wang, Juliette Love, Junwhan Ahn, Dawn Bloxwich, Kehang Han, Peter Humphreys, Thibault Sellam, James Bradbury, Varun Godbole, Sina Samangooei, Bogdan Damoc, Alex Kaskasoli, S\'ebastien M. R. Arnold, Vijay Vasudevan, Shubham Agrawal, Jason Riesa, Dmitry Lepikhin, Richard Tanburn, Srivatsan Srinivasan, Hyeontaek Lim, Sarah Hodkinson, Pranav Shyam, Johan Ferret, Steven Hand, Ankush Garg, Tom Le Paine, Jian Li, Yujia Li, Minh Giang, Alexander Neitz, Zaheer Abbas, Sarah York, Machel Reid, Elizabeth Cole, Aakanksha Chowdhery, Dipanjan Das, Dominika Rogozi\'nska, Vitaliy Nikolaev, Pablo Sprechmann, Zachary Nado, Lukas Zilka, Flavien Prost, Luheng He, Marianne Monteiro, Gaurav Mishra, Chris Welty, Josh Newlan, Dawei Jia, Miltiadis Allamanis, Clara Huiyi Hu, Raoul de Liedekerke, Justin Gilmer, Carl Saroufim, Shruti Rijhwani, Shaobo Hou, Disha Shrivastava, Anirudh Baddepudi, Alex Goldin, Adnan Ozturel, Albin Cassirer, Yunhan Xu, Daniel Sohn, Devendra Sachan, Reinald Kim Amplayo, Craig Swanson, Dessie Petrova, Shashi Narayan, Arthur Guez, Siddhartha Brahma, Jessica Landon, Miteyan Patel, Ruizhe Zhao, Kevin Villela, Luyu Wang, Wenhao Jia, Matthew Rahtz, Mai Gim\'enez, Legg Yeung, James Keeling, Petko Georgiev, Diana Mincu, Boxi Wu, Salem Haykal, Rachel Saputro, Kiran Vodrahalli, James Qin, Zeynep Cankara, Abhanshu Sharma, Nick Fernando, Will Hawkins, Behnam Neyshabur, Solomon Kim, Adrian Hutter, Priyanka Agrawal, Alex Castro-Ros, George van den Driessche, Tao Wang, Fan Yang, Shuo-yiin Chang, Paul Komarek, Ross McIlroy, Mario Lu\v{c}i\'c, Guodong Zhang, Wael Farhan, Michael Sharman, Paul Natsev, Paul Michel, Yamini Bansal, Siyuan Qiao, Kris Cao, Siamak Shakeri, Christina Butterfield, Justin Chung, Paul Kishan Rubenstein, Shivani Agrawal, Arthur Mensch, Kedar Soparkar, Karel Lenc, Timothy Chung, Aedan Pope, Loren Maggiore, Jackie Kay, Priya Jhakra, Shibo Wang, Joshua Maynez, Mary Phuong, Taylor Tobin, Andrea Tacchetti, Maja Trebacz, Kevin Robinson, Yash Katariya, Sebastian Riedel, Paige Bailey, Kefan Xiao, Nimesh Ghelani, Lora Aroyo, Ambrose Slone, Neil Houlsby, Xuehan Xiong, Zhen Yang, Elena Gribovskaya, Jonas Adler, Mateo Wirth, Lisa Lee, Music Li, Thais Kagohara, Jay Pavagadhi, Sophie Bridgers, Anna Bortsova, Sanjay Ghemawat, Zafarali Ahmed, Tianqi Liu, Richard Powell, Vijay Bolina, Mariko Iinuma, Polina Zablotskaia, James Besley, Da-Woon Chung, Timothy Dozat, Ramona Comanescu, Xiance Si, Jeremy Greer, Guolong Su, Martin Polacek, Rapha\"el Lopez Kaufman, Simon Tokumine, Hexiang Hu, Elena Buchatskaya, Yingjie Miao, Mohamed Elhawaty, Aditya Siddhant, Nenad Tomasev, Jinwei Xing, Christina Greer, Helen Miller, Shereen Ashraf, Aurko Roy, Zizhao Zhang, Ada Ma, Angelos Filos, Milos Besta, Rory Blevins, Ted Klimenko, Chih-Kuan Yeh, Soravit Changpinyo, Jiaqi Mu, Oscar Chang, Mantas Pajarskas, Carrie Muir, Vered Cohen, Charline Le Lan, Krishna Haridasan, Amit Marathe, Steven Hansen, Sholto Douglas, Rajkumar Samuel, Mingqiu Wang, Sophia Austin, Chang Lan, Jiepu Jiang, Justin Chiu, Jaime Alonso Lorenzo, Lars Lowe Sj\"osund, S\'ebastien Cevey, Zach Gleicher, Thi Avrahami, Anudhyan Boral, Hansa Srinivasan, Vittorio Selo, Rhys May, Konstantinos Aisopos, L\'eonard Hussenot, Livio Baldini Soares, Kate Baumli, Michael B. Chang, Adri\`a Recasens, Ben Caine, Alexander Pritzel, Filip Pavetic, Fabio Pardo, Anita Gergely, Justin Frye, Vinay Ramasesh, Dan Horgan, Kartikeya Badola, Nora Kassner, Subhrajit Roy, Ethan Dyer, V\'ictor Campos Campos, Alex Tomala, Yunhao Tang, Dalia El Badawy, Elspeth White, Basil Mustafa, Oran Lang, Abhishek Jindal, Sharad Vikram, Zhitao Gong, Sergi Caelles, Ross Hemsley, Gregory Thornton, Fangxiaoyu Feng, Wojciech Stokowiec, Ce Zheng, Phoebe Thacker, \c{C}a\u{g}lar \"Unl\"u, Zhishuai Zhang, Mohammad Saleh, James Svensson, Max Bileschi, Piyush Patil, Ankesh Anand, Roman Ring, Katerina Tsihlas, Arpi Vezer, Marco Selvi, Toby Shevlane, Mikel Rodriguez, Tom Kwiatkowski, Samira Daruki, Keran Rong, Allan Dafoe, Nicholas FitzGerald, Keren Gu-Lemberg, Mina Khan, Lisa Anne Hendricks, Marie Pellat, Vladimir Feinberg, James Cobon-Kerr, Tara Sainath, Maribeth Rauh, Sayed Hadi Hashemi, Richard Ives, Yana Hasson, Eric Noland, Yuan Cao, Nathan Byrd, Le Hou, Qingze Wang, Thibault Sottiaux, Michela Paganini, Jean-Baptiste Lespiau, Alexandre Moufarek, Samer Hassan, Kaushik Shivakumar, Joost van Amersfoort, Amol Mandhane, Pratik Joshi, Anirudh Goyal, Matthew Tung, Andrew Brock, Hannah Sheahan, Vedant Misra, Cheng Li, Nemanja Raki\'cevi\'c, Mostafa Dehghani, Fangyu Liu, Sid Mittal, Junhyuk Oh, Seb Noury, Eren Sezener, Fantine Huot, Matthew Lamm, Nicola De Cao, Charlie Chen, Sidharth Mudgal, Romina Stella, Kevin Brooks, Gautam Vasudevan, Chenxi Liu, Mainak Chain, Nivedita Melinkeri, Aaron Cohen, Venus Wang, Kristie Seymore, Sergey Zubkov, Rahul Goel, Summer Yue, Sai Krishnakumaran, Brian Albert, Nate Hurley, Motoki Sano, Anhad Mohananey, Jonah Joughin, Egor Filonov, Tomasz K\k{e}pa, Yomna Eldawy, Jiawern Lim, Rahul Rishi, Shirin Badiezadegan, Taylor Bos, Jerry Chang, Sanil Jain, Sri Gayatri Sundara Padmanabhan, Subha Puttagunta, Kalpesh Krishna, Leslie Baker, Norbert Kalb, Vamsi Bedapudi, Adam Kurzrok, Shuntong Lei, Anthony Yu, Oren Litvin, Xiang Zhou, Zhichun Wu, Sam Sobell, Andrea Siciliano, Alan Papir, Robby Neale, Jonas Bragagnolo, Tej Toor, Tina Chen, Valentin Anklin, Feiran Wang, Richie Feng, Milad Gholami, Kevin Ling, Lijuan Liu, Jules Walter, Hamid Moghaddam, Arun Kishore, Jakub Adamek, Tyler Mercado, Jonathan Mallinson, Siddhinita Wandekar, Stephen Cagle, Eran Ofek, Guillermo Garrido, Clemens Lombriser, Maksim Mukha, Botu Sun, Hafeezul Rahman Mohammad, Josip Matak, Yadi Qian, Vikas Peswani, Pawel Janus, Quan Yuan, Leif Schelin, Oana David, Ankur Garg, Yifan He, Oleksii Duzhyi, Anton \"Algmyr, Timoth\'ee Lottaz, Qi Li, Vikas Yadav, Luyao Xu, Alex Chinien, Rakesh Shivanna, Aleksandr Chuklin, Josie Li, Carrie Spadine, Travis Wolfe, Kareem Mohamed, Subhabrata Das, Zihang Dai, Kyle He, Daniel von Dincklage, Shyam Upadhyay, Akanksha Maurya, Luyan Chi, Sebastian Krause, Khalid Salama, Pam G Rabinovitch, Pavan Kumar Reddy M, Aarush Selvan, Mikhail Dektiarev, Golnaz Ghiasi, Erdem Guven, Himanshu Gupta, Boyi Liu, Deepak Sharma, Idan Heimlich Shtacher, Shachi Paul, Oscar Akerlund, Fran\c{c}ois-Xavier Aubet, Terry Huang, Chen Zhu, Eric Zhu, Elico Teixeira, Matthew Fritze, Francesco Bertolini, Liana-Eleonora Marinescu, Martin B\"olle, Dominik Paulus, Khyatti Gupta, Tejasi Latkar, Max Chang, Jason Sanders, Roopa Wilson, Xuewei Wu, Yi-Xuan Tan, Lam Nguyen Thiet, Tulsee Doshi, Sid Lall, Swaroop Mishra, Wanming Chen, Thang Luong, Seth Benjamin, Jasmine Lee, Ewa Andrejczuk, Dominik Rabiej, Vipul Ranjan, Krzysztof Styrc, Pengcheng Yin, Jon Simon, Malcolm Rose Harriott, Mudit Bansal, Alexei Robsky, Geoff Bacon, David Greene, Daniil Mirylenka, Chen Zhou, Obaid Sarvana, Abhimanyu Goyal, Samuel Andermatt, Patrick Siegler, Ben Horn, Assaf Israel, Francesco Pongetti, Chih-Wei "Louis" Chen, Marco Selvatici, Pedro Silva, Kathie Wang, Jackson Tolins, Kelvin Guu, Roey Yogev, Xiaochen Cai, Alessandro Agostini, Maulik Shah, Hung Nguyen, Noah \'O Donnaile, S\'ebastien Pereira, Linda Friso, Adam Stambler, Adam Kurzrok, Chenkai Kuang, Yan Romanikhin, Mark Geller, ZJ Yan, Kane Jang, Cheng-Chun Lee, Wojciech Fica, Eric Malmi, Qijun Tan, Dan Banica, Daniel Balle, Ryan Pham, Yanping Huang, Diana Avram, Hongzhi Shi, Jasjot Singh, Chris Hidey, Niharika Ahuja, Pranab Saxena, Dan Dooley, Srividya Pranavi Potharaju, Eileen O'Neill, Anand Gokulchandran, Ryan Foley, Kai Zhao, Mike Dusenberry, Yuan Liu, Pulkit Mehta, Ragha Kotikalapudi, Chalence Safranek-Shrader, Andrew Goodman, Joshua Kessinger, Eran Globen, Prateek Kolhar, Chris Gorgolewski, Ali Ibrahim, Yang Song, Ali Eichenbaum, Thomas Brovelli, Sahitya Potluri, Preethi Lahoti, Cip Baetu, Ali Ghorbani, Charles Chen, Andy Crawford, Shalini Pal, Mukund Sridhar, Petru Gurita, Asier Mujika, Igor Petrovski, Pierre-Louis Cedoz, Chenmei Li, Shiyuan Chen, Niccol\`o Dal Santo, Siddharth Goyal, Jitesh Punjabi, Karthik Kappaganthu, Chester Kwak, Pallavi LV, Sarmishta Velury, Himadri Choudhury, Jamie Hall, Premal Shah, Ricardo Figueira, Matt Thomas, Minjie Lu, Ting Zhou, Chintu Kumar, Thomas Jurdi, Sharat Chikkerur, Yenai Ma, Adams Yu, Soo Kwak, Victor \"Ahdel, Sujeevan Rajayogam, Travis Choma, Fei Liu, Aditya Barua, Colin Ji, Ji Ho Park, Vincent Hellendoorn, Alex Bailey, Taylan Bilal, Huanjie Zhou, Mehrdad Khatir, Charles Sutton, Wojciech Rzadkowski, Fiona Macintosh, Konstantin Shagin, Paul Medina, Chen Liang, Jinjing Zhou, Pararth Shah, Yingying Bi, Attila Dankovics, Shipra Banga, Sabine Lehmann, Marissa Bredesen, Zifan Lin, John Eric Hoffmann, Jonathan Lai, Raynald Chung, Kai Yang, Nihal Balani, Arthur Bra\v{z}inskas, Andrei Sozanschi, Matthew Hayes, H\'ector Fern\'andez Alcalde, Peter Makarov, Will Chen, Antonio Stella, Liselotte Snijders, Michael Mandl, Ante K\"arrman, Pawe{\l} Nowak, Xinyi Wu, Alex Dyck, Krishnan Vaidyanathan, Raghavender R, Jessica Mallet, Mitch Rudominer, Eric Johnston, Sushil Mittal, Akhil Udathu, Janara Christensen, Vishal Verma, Zach Irving, Andreas Santucci, Gamaleldin Elsayed, Elnaz Davoodi, Marin Georgiev, Ian Tenney, Nan Hua, Geoffrey Cideron, Edouard Leurent, Mahmoud Alnahlawi, Ionut Georgescu, Nan Wei, Ivy Zheng, Dylan Scandinaro, Heinrich Jiang, Jasper Snoek, Mukund Sundararajan, Xuezhi Wang, Zack Ontiveros, Itay Karo, Jeremy Cole, Vinu Rajashekhar, Lara Tumeh, Eyal Ben-David, Rishub Jain, Jonathan Uesato, Romina Datta, Oskar Bunyan, Shimu Wu, John Zhang, Piotr Stanczyk, Ye Zhang, David Steiner, Subhajit Naskar, Michael Azzam, Matthew Johnson, Adam Paszke, Chung-Cheng Chiu, Jaume Sanchez Elias, Afroz Mohiuddin, Faizan Muhammad, Jin Miao, Andrew Lee, Nino Vieillard, Jane Park, Jiageng Zhang, Jeff Stanway, Drew Garmon, Abhijit Karmarkar, Zhe Dong, Jong Lee, Aviral Kumar, Luowei Zhou, Jonathan Evens, William Isaac, Geoffrey Irving, Edward Loper, Michael Fink, Isha Arkatkar, Nanxin Chen, Izhak Shafran, Ivan Petrychenko, Zhe Chen, Johnson Jia, Anselm Levskaya, Zhenkai Zhu, Peter Grabowski, Yu Mao, Alberto Magni, Kaisheng Yao, Javier Snaider, Norman Casagrande, Evan Palmer, Paul Suganthan, Alfonso Casta\~no, Irene Giannoumis, Wooyeol Kim, Miko{\l}aj Rybi\'nski, Ashwin Sreevatsa, Jennifer Prendki, David Soergel, Adrian Goedeckemeyer, Willi Gierke, Mohsen Jafari, Meenu Gaba, Jeremy Wiesner, Diana Gage Wright, Yawen Wei, Harsha Vashisht, Yana Kulizhskaya, Jay Hoover, Maigo Le, Lu Li, Chimezie Iwuanyanwu, Lu Liu, Kevin Ramirez, Andrey Khorlin, Albert Cui, Tian LIN, Marcus Wu, Ricardo Aguilar, Keith Pallo, Abhishek Chakladar, Ginger Perng, Elena Allica Abellan, Mingyang Zhang, Ishita Dasgupta, Nate Kushman, Ivo Penchev, Alena Repina, Xihui Wu, Tom van der Weide, Priya Ponnapalli, Caroline Kaplan, Jiri Simsa, Shuangfeng Li, Olivier Dousse, Fan Yang, Jeff Piper, Nathan Ie, Rama Pasumarthi, Nathan Lintz, Anitha Vijayakumar, Daniel Andor, Pedro Valenzuela, Minnie Lui, Cosmin Paduraru, Daiyi Peng, Katherine Lee, Shuyuan Zhang, Somer Greene, Duc Dung Nguyen, Paula Kurylowicz, Cassidy Hardin, Lucas Dixon, Lili Janzer, Kiam Choo, Ziqiang Feng, Biao Zhang, Achintya Singhal, Dayou Du, Dan McKinnon, Natasha Antropova, Tolga Bolukbasi, Orgad Keller, David Reid, Daniel Finchelstein, Maria Abi Raad, Remi Crocker, Peter Hawkins, Robert Dadashi, Colin Gaffney, Ken Franko, Anna Bulanova, R\'emi Leblond, Shirley Chung, Harry Askham, Luis C. Cobo, Kelvin Xu, Felix Fischer, Jun Xu, Christina Sorokin, Chris Alberti, Chu-Cheng Lin, Colin Evans, Alek Dimitriev, Hannah Forbes, Dylan Banarse, Zora Tung, Mark Omernick, Colton Bishop, Rachel Sterneck, Rohan Jain, Jiawei Xia, Ehsan Amid, Francesco Piccinno, Xingyu Wang, Praseem Banzal, Daniel J. Mankowitz, Alex Polozov, Victoria Krakovna, Sasha Brown, MohammadHossein Bateni, Dennis Duan, Vlad Firoiu, Meghana Thotakuri, Tom Natan, Matthieu Geist, Ser tan Girgin, Hui Li, Jiayu Ye, Ofir Roval, Reiko Tojo, Michael Kwong, James Lee-Thorp, Christopher Yew, Danila Sinopalnikov, Sabela Ramos, John Mellor, Abhishek Sharma, Kathy Wu, David Miller, Nicolas Sonnerat, Denis Vnukov, Rory Greig, Jennifer Beattie, Emily Caveness, Libin Bai, Julian Eisenschlos, Alex Korchemniy, Tomy Tsai, Mimi Jasarevic, Weize Kong, Phuong Dao, Zeyu Zheng, Frederick Liu, Fan Yang, Rui Zhu, Tian Huey Teh, Jason Sanmiya, Evgeny Gladchenko, Nejc Trdin, Daniel Toyama, Evan Rosen, Sasan Tavakkol, Linting Xue, Chen Elkind, Oliver Woodman, John Carpenter, George Papamakarios, Rupert Kemp, Sushant Kafle, Tanya Grunina, Rishika Sinha, Alice Talbert, Diane Wu, Denese Owusu-Afriyie, Cosmo Du, Chloe Thornton, Jordi Pont-Tuset, Pradyumna Narayana, Jing Li, Saaber Fatehi, John Wieting, Omar Ajmeri, Benigno Uria, Yeongil Ko, Laura Knight, Am\'elie H\'eliou, Ning Niu, Shane Gu, Chenxi Pang, Yeqing Li, Nir Levine, Ariel Stolovich, Rebeca Santamaria-Fernandez, Sonam Goenka, Wenny Yustalim, Robin Strudel, Ali Elqursh, Charlie Deck, Hyo Lee, Zonglin Li, Kyle Levin, Raphael Hoffmann, Dan Holtmann-Rice, Olivier Bachem, Sho Arora, Christy Koh, Soheil Hassas Yeganeh, Siim P\~oder, Mukarram Tariq, Yanhua Sun, Lucian Ionita, Mojtaba Seyedhosseini, Pouya Tafti, Zhiyu Liu, Anmol Gulati, Jasmine Liu, Xinyu Ye, Bart Chrzaszcz, Lily Wang, Nikhil Sethi, Tianrun Li, Ben Brown, Shreya Singh, Wei Fan, Aaron Parisi, Joe Stanton, Vinod Koverkathu, Christopher A. Choquette-Choo, Yunjie Li, TJ Lu, Abe Ittycheriah, Prakash Shroff, Mani Varadarajan, Sanaz Bahargam, Rob Willoughby, David Gaddy, Guillaume Desjardins, Marco Cornero, Brona Robenek, Bhavishya Mittal, Ben Albrecht, Ashish Shenoy, Fedor Moiseev, Henrik Jacobsson, Alireza Ghaffarkhah, Morgane Rivi\`ere, Alanna Walton, Cl\'ement Crepy, Alicia Parrish, Zongwei Zhou, Clement Farabet, Carey Radebaugh, Praveen Srinivasan, Claudia van der Salm, Andreas Fidjeland, Salvatore Scellato, Eri Latorre-Chimoto, Hanna Klimczak-Pluci\'nska, David Bridson, Dario de Cesare, Tom Hudson, Piermaria Mendolicchio, Lexi Walker, Alex Morris, Matthew Mauger, Alexey Guseynov, Alison Reid, Seth Odoom, Lucia Loher, Victor Cotruta, Madhavi Yenugula, Dominik Grewe, Anastasia Petrushkina, Tom Duerig, Antonio Sanchez, Steve Yadlowsky, Amy Shen, Amir Globerson, Lynette Webb, Sahil Dua, Dong Li, Surya Bhupatiraju, Dan Hurt, Haroon Qureshi, Ananth Agarwal, Tomer Shani, Matan Eyal, Anuj Khare, Shreyas Rammohan Belle, Lei Wang, Chetan Tekur, Mihir Sanjay Kale, Jinliang Wei, Ruoxin Sang, Brennan Saeta, Tyler Liechty, Yi Sun, Yao Zhao, Stephan Lee, Pandu Nayak, Doug Fritz, Manish Reddy Vuyyuru, John Aslanides, Nidhi Vyas, Martin Wicke, Xiao Ma, Evgenii Eltyshev, Nina Martin, Hardie Cate, James Manyika, Keyvan Amiri, Yelin Kim, Xi Xiong, Kai Kang, Florian Luisier, Nilesh Tripuraneni, David Madras, Mandy Guo, Austin Waters, Oliver Wang, Joshua Ainslie, Jason Baldridge, Han Zhang, Garima Pruthi, Jakob Bauer, Feng Yang, Riham Mansour, Jason Gelman, Yang Xu, George Polovets, Ji Liu, Honglong Cai, Warren Chen, XiangHai Sheng, Emily Xue, Sherjil Ozair, Christof Angermueller, Xiaowei Li, Anoop Sinha, Weiren Wang, Julia Wiesinger, Emmanouil Koukoumidis, Yuan Tian, Anand Iyer, Madhu Gurumurthy, Mark Goldenson, Parashar Shah, MK Blake, Hongkun Yu, Anthony Urbanowicz, Jennimaria Palomaki, Chrisantha Fernando, Ken Durden, Harsh Mehta, Nikola Momchev, Elahe Rahimtoroghi, Maria Georgaki, Amit Raul, Sebastian Ruder, Morgan Redshaw, Jinhyuk Lee, Denny Zhou, Komal Jalan, Dinghua Li, Blake Hechtman, Parker Schuh, Milad Nasr, Kieran Milan, Vladimir Mikulik, Juliana Franco, Tim Green, Nam Nguyen, Joe Kelley, Aroma Mahendru, Andrea Hu, Joshua Howland, Ben Vargas, Jeffrey Hui, Kshitij Bansal, Vikram Rao, Rakesh Ghiya, Emma Wang, Ke Ye, Jean Michel Sarr, Melanie Moranski Preston, Madeleine Elish, Steve Li, Aakash Kaku, Jigar Gupta, Ice Pasupat, Da-Cheng Juan, Milan Someswar, Tejvi M., Xinyun Chen, Aida Amini, Alex Fabrikant, Eric Chu, Xuanyi Dong, Amruta Muthal, Senaka Buthpitiya, Sarthak Jauhari, Nan Hua, Urvashi Khandelwal, Ayal Hitron, Jie Ren, Larissa Rinaldi, Shahar Drath, Avigail Dabush, Nan-Jiang Jiang, Harshal Godhia, Uli Sachs, Anthony Chen, Yicheng Fan, Hagai Taitelbaum, Hila Noga, Zhuyun Dai, James Wang, Chen Liang, Jenny Hamer, Chun-Sung Ferng, Chenel Elkind, Aviel Atias, Paulina Lee, V\'it List\'ik, Mathias Carlen, Jan van de Kerkhof, Marcin Pikus, Krunoslav Zaher, Paul M\"uller, Sasha Zykova, Richard Stefanec, Vitaly Gatsko, Christoph Hirnschall, Ashwin Sethi, Xingyu Federico Xu, Chetan Ahuja, Beth Tsai, Anca Stefanoiu, Bo Feng, Keshav Dhandhania, Manish Katyal, Akshay Gupta, Atharva Parulekar, Divya Pitta, Jing Zhao, Vivaan Bhatia, Yashodha Bhavnani, Omar Alhadlaq, Xiaolin Li, Peter Danenberg, Dennis Tu, Alex Pine, Vera Filippova, Abhipso Ghosh, Ben Limonchik, Bhargava Urala, Chaitanya Krishna Lanka, Derik Clive, Yi Sun, Edward Li, Hao Wu, Kevin Hongtongsak, Ianna Li, Kalind Thakkar, Kuanysh Omarov, Kushal Majmundar, Michael Alverson, Michael Kucharski, Mohak Patel, Mudit Jain, Maksim Zabelin, Paolo Pelagatti, Rohan Kohli, Saurabh Kumar, Joseph Kim, Swetha Sankar, Vineet Shah, Lakshmi Ramachandruni, Xiangkai Zeng, Ben Bariach, Laura Weidinger, Amar Subramanya, Sissie Hsiao, Demis Hassabis, Koray Kavukcuoglu, Adam Sadovsky, Quoc Le, Trevor Strohman, Yonghui Wu, Slav Petrov, Jeffrey Dean, Oriol Vinyals
Abstract: This report introduces a new family of multimodal models, Gemini, that exhibit remarkable capabilities across image, audio, video, and text understanding. The Gemini family consists of Ultra, Pro, and Nano sizes, suitable for applications ranging from complex reasoning tasks to on-device memory-constrained use-cases. Evaluation on a broad range of benchmarks shows that our most-capable Gemini Ultra model advances the state of the art in 30 of 32 of these benchmarks - notably being the first model to achieve human-expert performance on the well-studied exam benchmark MMLU, and improving the state of the art in every one of the 20 multimodal benchmarks we examined. We believe that the new capabilities of the Gemini family in cross-modal reasoning and language understanding will enable a wide variety of use cases. We discuss our approach toward post-training and deploying Gemini models responsibly to users through services including Gemini, Gemini Advanced, Google AI Studio, and Cloud Vertex AI.
Authors: Priyesh Vakharia, Devavrat Joshi, Meenal Chavan, Dhananjay Sonawane, Bhrigu Garg, Parsa Mazaheri
Abstract: Large Language Models (LLMs) are adept at text manipulation -- tasks such as machine translation and text summarization. However, these models can also be prone to hallucination, which can be detrimental to the faithfulness of any answers that the model provides. Recent works in combating hallucinations in LLMs deal with identifying hallucinated sentences and categorizing the different ways in which models hallucinate. This paper takes a deep dive into LLM behavior with respect to hallucinations, defines a token-level approach to identifying different kinds of hallucinations, and further utilizes this token-level tagging to improve the interpretability and faithfulness of LLMs in dialogue summarization tasks. Through this, the paper presents a new, enhanced dataset and a new training paradigm.
Authors: Konrad Staniszewski, Szymon Tworkowski, Yu Zhao, Sebastian Jaszczur, Henryk Michalewski, {\L}ukasz Kuci\'nski, Piotr Mi{\l}o\'s
Abstract: Recent developments in long-context large language models have attracted considerable attention. Yet, their real-world applications are often hindered by ineffective context information use. This work shows that structuring training data to increase semantic interdependence is an effective strategy for optimizing context utilization. To this end, we introduce Structured Packing for Long Context (SPLiCe), a method for creating training examples by using information retrieval methods to collate mutually relevant documents into a single training context. We empirically validate SPLiCe on large $3$B and $7$B models, showing perplexity improvements and better long-context utilization on downstream tasks. Remarkably, already relatively short fine-tuning with SPLiCe is enough to attain these benefits. Additionally, the comprehensive study of SPLiCe reveals intriguing transfer effects such as training on code data leading to perplexity improvements on text data.
Authors: Zijie Meng, Yan Zhang, Zhaopeng Feng, Zuozhu Liu
Abstract: Large language models (LLMs) have shown impressive performance in reasoning benchmarks with the emergence of Chain-of-Thought (CoT), particularly in multi-choice question (MCQ). However, current works equally resolve questions regardless of the problem-solving difficulty, leading to an excessive focus on simple items while insufficient attention on intricate ones. To address this challenge, we propose a simple yet effective strategy, Divide and Conquer Reasoning (DCR), to enhance the reasoning capability of LLMs for MCQs, as inspired by human beings using heuristics to first categorize tasks and then handle them separately. In particular, we first categorize questions into two subsets based on confidence score ($\mathcal{CS}$), which is estimated by statistical frequency of generated answers. Subsequently, we propose Filter Choices based Reasoning (FCR) to improve model performance on MCQs with low ($\mathcal{CS}$). Our experiments demonstrate that the proposed strategy only costs 85% of SOTA, while still achieves average accuracy improvement of 1.56% across nine datasets including arithmetic, commonsense, and logic reasoning tasks. The code is at \url{https://github.com/AiMijie/Divide-and-Conquer}
Authors: Jinge Wu, Yunsoo Kim, Honghan Wu
Abstract: The recent success of large language and vision models (LLVMs) on vision question answering (VQA), particularly their applications in medicine (Med-VQA), has shown a great potential of realizing effective visual assistants for healthcare. However, these models are not extensively tested on the hallucination phenomenon in clinical settings. Here, we created a hallucination benchmark of medical images paired with question-answer sets and conducted a comprehensive evaluation of the state-of-the-art models. The study provides an in-depth analysis of current models' limitations and reveals the effectiveness of various prompting strategies.
Authors: Mazal Bethany, Brandon Wherry, Emet Bethany, Nishant Vishwamitra, Anthony Rios, Peyman Najafirad
Abstract: With the recent proliferation of Large Language Models (LLMs), there has been an increasing demand for tools to detect machine-generated text. The effective detection of machine-generated text face two pertinent problems: First, they are severely limited in generalizing against real-world scenarios, where machine-generated text is produced by a variety of generators, including but not limited to GPT-4 and Dolly, and spans diverse domains, ranging from academic manuscripts to social media posts. Second, existing detection methodologies treat texts produced by LLMs through a restrictive binary classification lens, neglecting the nuanced diversity of artifacts generated by different LLMs. In this work, we undertake a systematic study on the detection of machine-generated text in real-world scenarios. We first study the effectiveness of state-of-the-art approaches and find that they are severely limited against text produced by diverse generators and domains in the real world. Furthermore, t-SNE visualizations of the embeddings from a pretrained LLM's encoder show that they cannot reliably distinguish between human and machine-generated text. Based on our findings, we introduce a novel system, T5LLMCipher, for detecting machine-generated text using a pretrained T5 encoder combined with LLM embedding sub-clustering to address the text produced by diverse generators and domains in the real world. We evaluate our approach across 9 machine-generated text systems and 9 domains and find that our approach provides state-of-the-art generalization ability, with an average increase in F1 score on machine-generated text of 19.6\% on unseen generators and domains compared to the top performing existing approaches and correctly attributes the generator of text with an accuracy of 93.6\%.
Authors: Shaina Raza, Shardul Ghuge, Chen Ding, Elham Dolatabadi, Deval Pandya
Abstract: The rapid evolution of Large Language Models (LLMs) highlights the necessity for ethical considerations and data integrity in AI development, particularly emphasizing the role of FAIR (Findable, Accessible, Interoperable, Reusable) data principles. While these principles are crucial for ethical data stewardship, their specific application in the context of LLM training data remains an under-explored area. This research gap is the focus of our study, which begins with an examination of existing literature to underline the importance of FAIR principles in managing data for LLM training. Building upon this, we propose a novel framework designed to integrate FAIR principles into the LLM development lifecycle. A contribution of our work is the development of a comprehensive checklist intended to guide researchers and developers in applying FAIR data principles consistently across the model development process. The utility and effectiveness of our framework are validated through a case study on creating a FAIR-compliant dataset aimed at detecting and mitigating biases in LLMs. We present this framework to the community as a tool to foster the creation of technologically advanced, ethically grounded, and socially responsible AI models.
Authors: Peiqin Lin, Shaoxiong Ji, J\"org Tiedemann, Andr\'e F. T. Martins, Hinrich Sch\"utze
Abstract: Large language models (LLMs) have advanced the state of the art in natural language processing. However, their predominant design for English or a limited set of languages creates a substantial gap in their effectiveness for low-resource languages. To bridge this gap, we introduce MaLA-500, a novel large language model designed to cover an extensive range of 534 languages. To train MaLA-500, we employ vocabulary extension and continued pretraining on LLaMA 2 with Glot500-c. Our intrinsic evaluation demonstrates that MaLA-500 is better at predicting the given texts of low-resource languages than existing multilingual LLMs. Moreover, the extrinsic evaluation of in-context learning shows that MaLA-500 outperforms previous LLMs on SIB200 and Taxi1500 by a significant margin, i.e., 11.68% and 4.82% marco-average accuracy across languages. We release MaLA-500 at https://huggingface.co/MaLA-LM
Authors: Junxiong Wang, Tushaar Gangavarapu, Jing Nathan Yan, Alexander M. Rush
Abstract: Token-free language models learn directly from raw bytes and remove the inductive bias of subword tokenization. Operating on bytes, however, results in significantly longer sequences. In this setting, standard autoregressive Transformers scale poorly as the effective memory required grows with sequence length. The recent development of the Mamba state space model (SSM) offers an appealing alternative approach with a fixed-sized memory state and efficient decoding. We propose MambaByte, a token-free adaptation of the Mamba SSM trained autoregressively on byte sequences. In terms of modeling, we show MambaByte to be competitive with, and even to outperform, state-of-the-art subword Transformers on language modeling tasks while maintaining the benefits of token-free language models, such as robustness to noise. In terms of efficiency, we develop an adaptation of speculative decoding with tokenized drafting and byte-level verification. This results in a $2.6\times$ inference speedup to the standard MambaByte implementation, showing similar decoding efficiency as the subword Mamba. These findings establish the viability of SSMs in enabling token-free language modeling.
Authors: Minbyul Jeong, Jiwoong Sohn, Mujeen Sung, Jaewoo Kang
Abstract: Recent proprietary large language models (LLMs), such as GPT-4, have achieved a milestone in tackling diverse challenges in the biomedical domain, ranging from multiple-choice questions to long-form generations. To address challenges that still cannot be handled with the encoded knowledge of LLMs, various retrieval-augmented generation (RAG) methods have been developed by searching documents from the knowledge corpus and appending them unconditionally or selectively to the input of LLMs for generation. However, when applying existing methods to different domain-specific problems, poor generalization becomes apparent, leading to fetching incorrect documents or making inaccurate judgments. In this paper, we introduce Self-BioRAG, a framework reliable for biomedical text that specializes in generating explanations, retrieving domain-specific documents, and self-reflecting generated responses. We utilize 84k filtered biomedical instruction sets to train Self-BioRAG that can assess its generated explanations with customized reflective tokens. Our work proves that domain-specific components, such as a retriever, domain-related document corpus, and instruction sets are necessary for adhering to domain-related instructions. Using three major medical question-answering benchmark datasets, experimental results of Self-BioRAG demonstrate significant performance gains by achieving a 7.2% absolute improvement on average over the state-of-the-art open-foundation model with a parameter size of 7B or less. Overall, we analyze that Self-BioRAG finds the clues in the question, retrieves relevant documents if needed, and understands how to answer with information from retrieved documents and encoded knowledge as a medical expert does. We release our data and code for training our framework components and model weights (7B and 13B) to enhance capabilities in biomedical and clinical domains.
Authors: Jiacheng Liu, Sewon Min, Luke Zettlemoyer, Yejin Choi, Hannaneh Hajishirzi
Abstract: Are $n$-gram language models still relevant in this era of neural large language models (LLMs)? Our answer is yes, and we showcase their values in both text analysis and improving neural LLMs. This was done by modernizing $n$-gram LMs in two aspects. First, we train them at the same data scale as neural LLMs -- 5 trillion tokens. This is the largest $n$-gram LM ever built. Second, existing $n$-gram LMs use small $n$ which hinders their performance; we instead allow $n$ to be arbitrarily large, by introducing a new $\infty$-gram LM with backoff. Instead of pre-computing $n$-gram count tables (which would be very expensive), we develop an engine named infini-gram -- powered by suffix arrays -- that can compute $\infty$-gram (as well as $n$-gram with arbitrary $n$) probabilities with millisecond-level latency. The $\infty$-gram framework and infini-gram engine enable us to conduct many novel and interesting analyses of human-written and machine-generated text: we find that the $\infty$-gram LM has fairly high accuracy for next-token prediction (47%), and can complement neural LLMs to greatly reduce their perplexity. When analyzing machine-generated text, we also observe irregularities in the machine--$\infty$-gram agreement level with respect to the suffix length, which indicates deficiencies in neural LLM pretraining and the positional embeddings of Transformers.
Authors: Yinhong Liu, Yixuan Su, Ehsan Shareghi, Nigel Collier
Abstract: Recent large language models (LLMs) have shown remarkable performance in aligning generated text with user intentions across various tasks. When it comes to long-form text generation, there has been a growing interest in generation from a discourse coherence perspective. However, existing lexical or semantic metrics such as BLEU, ROUGE, BertScore cannot effectively capture the discourse coherence. The development of discourse-specific automatic evaluation methods for assessing the output of LLMs warrants greater focus and exploration. In this paper, we present a novel automatic metric designed to quantify the discourse divergence between two long-form articles. Extensive experiments on three datasets from representative domains demonstrate that our metric aligns more closely with human preferences and GPT-4 coherence evaluation, outperforming existing evaluation methods.
Authors: Zhanhui Zhou, Jie Liu, Zhichen Dong, Jiaheng Liu, Chao Yang, Wanli Ouyang, Yu Qiao
Abstract: Large language models (LLMs) need to undergo safety alignment to ensure safe conversations with humans. However, this paper introduces an inference-time attack method, demonstrating that safety alignment can be easily reversed to produce harmful language models without additional training. Specifically, this reversal is achieved by contrasting the output token distribution of a safety-aligned language model (e.g., Llama-2-chat) against its pre-trained version (e.g., Llama-2) so that the token predictions are shifted towards the opposite direction of alignment. We name this method emulated disalignment (ED) because it uses pure sampling to provably emulate (or "approximate") the result of fine-tuning the pre-trained model to minimize a safety reward. Our experiments with ED across three evaluation datasets and four model families (Llama-1, Llama-2, Mistral, and Alpaca) show that ED doubles the harmfulness of pre-trained models and outperforms strong baselines, achieving the highest harmful rate in 43 out of 48 evaluation subsets by a large margin. Eventually, given ED's need for language model output token distributions, which particularly compromises open-source models, our findings highlight the importance of reevaluating the practice of open-sourcing language models even after safety alignment.
Authors: Seanie Lee, Jianpeng Cheng, Joris Driesen, Alexandru Coca, Anders Johannsen
Abstract: Few-shot dialogue state tracking (DST) with Large Language Models (LLM) relies on an effective and efficient conversation retriever to find similar in-context examples for prompt learning. Previous works use raw dialogue context as search keys and queries, and a retriever is fine-tuned with annotated dialogues to achieve superior performance. However, the approach is less suited for scaling to new domains or new annotation languages, where fine-tuning data is unavailable. To address this problem, we handle the task of conversation retrieval based on text summaries of the conversations. A LLM-based conversation summarizer is adopted for query and key generation, which enables effective maximum inner product search. To avoid the extra inference cost brought by LLM-based conversation summarization, we further distill a light-weight conversation encoder which produces query embeddings without decoding summaries for test conversations. We validate our retrieval approach on MultiWOZ datasets with GPT-Neo-2.7B and LLaMA-7B/30B. The experimental results show a significant improvement over relevant baselines in real few-shot DST settings.
Authors: Debrup Das, Debopriyo Banerjee, Somak Aditya, Ashish Kulkarni
Abstract: Tool-augmented Large Language Models (TALMs) are known to enhance the skillset of large language models (LLMs), thereby, leading to their improved reasoning abilities across many tasks. While, TALMs have been successfully employed in different question-answering benchmarks, their efficacy on complex mathematical reasoning benchmarks, and the potential complementary benefits offered by tools for knowledge retrieval and mathematical equation solving are open research questions. In this work, we present MathSensei, a tool-augmented large language model for mathematical reasoning. We study the complementary benefits of the tools - knowledge retriever (Bing Web Search), program generator + executor (Python), and symbolic equation solver (Wolfram-Alpha API) through evaluations on mathematical reasoning datasets. We perform exhaustive ablations on MATH, a popular dataset for evaluating mathematical reasoning on diverse mathematical disciplines. We also conduct experiments involving well-known tool planners to study the impact of tool sequencing on the model performance. MathSensei achieves 13.5% better accuracy over gpt-3.5-turbo with Chain-of-Thought on the MATH dataset. We further observe that TALMs are not as effective for simpler math word problems (in GSM-8K), and the benefit increases as the complexity and required knowledge increases (progressively over AQuA, MMLU-Math, and higher level complex questions in MATH). The code and data are available at https://github.com/Debrup-61/MathSensei.
Authors: Chan-Jan Hsu, Chang-Le Liu, Feng-Ting Liao, Po-Chun Hsu, Yi-Chang Chen, Da-Shan Shiu
Abstract: Breeze-7B is an open-source language model based on Mistral-7B, designed to address the need for improved language comprehension and chatbot-oriented capabilities in Traditional Chinese. This technical report provides an overview of the additional pretraining, finetuning, and evaluation stages for the Breeze-7B model. The Breeze-7B family of base and chat models exhibits good performance on language comprehension and chatbot-oriented tasks, reaching the top in several benchmarks among models comparable in its complexity class.
Authors: Renjie Pi, Tianyang Han, Wei Xiong, Jipeng Zhang, Runtao Liu, Rui Pan, Tong Zhang
Abstract: Multimodal Large Language Models (MLLMs) excel in generating responses based on visual inputs. However, they often suffer from a bias towards generating responses similar to their pretraining corpus, overshadowing the importance of visual information. We treat this bias as a "preference" for pretraining statistics, which hinders the model's grounding in visual input. To mitigate this issue, we propose Bootstrapped Preference Optimization (BPO), which conducts preference learning with datasets containing negative responses bootstrapped from the model itself. Specifically, we propose the following two strategies: 1) using distorted image inputs to the MLLM for eliciting responses that contain signified pretraining bias; 2) leveraging text-based LLM to explicitly inject erroneous but common elements into the original response. Those undesirable responses are paired with original annotated responses from the datasets to construct the preference dataset, which is subsequently utilized to perform preference learning. Our approach effectively suppresses pretrained LLM bias, enabling enhanced grounding in visual inputs. Extensive experimentation demonstrates significant performance improvements across multiple benchmarks, advancing the state-of-the-art in multimodal conversational systems.
Authors: Yiran Wu, Tianwei Yue, Shaokun Zhang, Chi Wang, Qingyun Wu
Abstract: It is a notable trend to use Large Language Models (LLMs) to tackle complex tasks, e.g., tasks that require a sequence of actions and dynamic interaction with tools and external environments. In this paper, we propose StateFlow, a novel LLM-based task-solving paradigm that conceptualizes complex task-solving processes as state machines. In StateFlow, we distinguish between "process grounding" (via state and state transitions) and "sub-task solving" (through actions within a state), enhancing control and interpretability of the task-solving procedure. A state represents the status of a running process. The transitions between states are controlled by heuristic rules or decisions made by the LLM, allowing for a dynamic and adaptive progression. Upon entering a state, a series of actions is executed, involving not only calling LLMs guided by different prompts, but also the utilization of external tools as needed. Our results show that StateFlow significantly enhances LLMs' efficiency. For instance, StateFlow achieves 13% and 28% higher success rates compared to ReAct in InterCode SQL and ALFWorld benchmark, with 5x and 3x less cost respectively. We also show that StateFlow can be combined with iterative refining methods like Reflexion to further improve performance.
Authors: Shimao Zhang, Yu Bao, Shujian Huang
Abstract: Recently, Large Language Models (LLMs) have demonstrated outstanding performance across a wide range of downstream language tasks. Temperature sampling is a commonly used decoding strategy for LLMs' generation process. However, a fixed temperature parameter is used in most cases, which may not always be an optimal choice for balancing generation quality and diversity. In this paper, we propose an effective Entropy-based Dynamic Temperature (EDT) Sampling method, to achieve a more balanced performance in terms of both generation quality and diversity by dynamically selecting the temperature parameter. Additionally, we also show model performance and comprehensive analyses for 4 different generation benchmarks. Our experiments show that EDT significantly outperforms the existing strategies across different tasks.
Authors: Md Nishat Raihan, Dhiman Goswami, Al Nahian Bin Emran, Sadiya Sayara Chowdhury Puspo, Amrita Ganguly, Marcos Zampieri
Abstract: Our paper presents team MasonTigers submission to the SemEval-2024 Task 9 - which provides a dataset of puzzles for testing natural language understanding. We employ large language models (LLMs) to solve this task through several prompting techniques. Zero-shot and few-shot prompting generate reasonably good results when tested with proprietary LLMs, compared to the open-source models. We obtain further improved results with chain-of-thought prompting, an iterative prompting method that breaks down the reasoning process step-by-step. We obtain our best results by utilizing an ensemble of chain-of-thought prompts, placing 2nd in the word puzzle subtask and 13th in the sentence puzzle subtask. The strong performance of prompted LLMs demonstrates their capability for complex reasoning when provided with a decomposition of the thought process. Our work sheds light on how step-wise explanatory prompts can unlock more of the knowledge encoded in the parameters of large models.
Authors: Dhiman Goswami, Sadiya Sayara Chowdhury Puspo, Md Nishat Raihan, Al Nahian Bin Emran, Amrita Ganguly, Marcos Zampieri
Abstract: This paper presents the MasonTigers entry to the SemEval-2024 Task 1 - Semantic Textual Relatedness. The task encompasses supervised (Track A), unsupervised (Track B), and cross-lingual (Track C) approaches across 14 different languages. MasonTigers stands out as one of the two teams who participated in all languages across the three tracks. Our approaches achieved rankings ranging from 11th to 21st in Track A, from 1st to 8th in Track B, and from 5th to 12th in Track C. Adhering to the task-specific constraints, our best performing approaches utilize ensemble of statistical machine learning approaches combined with language-specific BERT based models and sentence transformers.
Authors: Gaofei Shen, Michaela Watkins, Afra Alishahi, Arianna Bisazza, Grzegorz Chrupa{\l}a
Abstract: Interpretability research has shown that self-supervised Spoken Language Models (SLMs) encode a wide variety of features in human speech from the acoustic, phonetic, phonological, syntactic and semantic levels, to speaker characteristics. The bulk of prior research on representations of phonology has focused on segmental features such as phonemes; the encoding of suprasegmental phonology (such as tone and stress patterns) in SLMs is not yet well understood. Tone is a suprasegmental feature that is present in more than half of the world's languages. This paper aims to analyze the tone encoding capabilities of SLMs, using Mandarin and Vietnamese as case studies. We show that SLMs encode lexical tone to a significant degree even when they are trained on data from non-tonal languages. We further find that SLMs behave similarly to native and non-native human participants in tone and consonant perception studies, but they do not follow the same developmental trajectory.
Authors: Meiqi Chen, Yixin Cao, Yan Zhang, Chaochao Lu
Abstract: Recent advancements in Large Language Models (LLMs) have facilitated the development of Multimodal LLMs (MLLMs). Despite their impressive capabilities, MLLMs often suffer from an over-reliance on unimodal biases (e.g., language bias and vision bias), leading to incorrect answers in complex multimodal tasks. To investigate this issue, we propose a causal framework to interpret the biases in Visual Question Answering (VQA) problems. Within our framework, we devise a causal graph to elucidate the predictions of MLLMs on VQA problems, and assess the causal effect of biases through an in-depth causal analysis. Motivated by the causal graph, we introduce a novel MORE dataset, consisting of 12,000 VQA instances. This dataset is designed to challenge MLLMs' abilities, necessitating multi-hop reasoning and the surmounting of unimodal biases. Furthermore, we propose two strategies to mitigate unimodal biases and enhance MLLMs' reasoning capabilities, including a Decompose-Verify-Answer (DeVA) framework for limited-access MLLMs and the refinement of open-source MLLMs through fine-tuning. Extensive quantitative and qualitative experiments offer valuable insights for future research. Our project page is at https://opencausalab.github.io/MORE.
Authors: Adithya Kulkarni, Oliver Eulenstein, Qi Li
Abstract: Dependency parsing is an essential task in NLP, and the quality of dependency parsers is crucial for many downstream tasks. Parsers' quality often varies depending on the domain and the language involved. Therefore, it is essential to combat the issue of varying quality to achieve stable performance. In various NLP tasks, aggregation methods are used for post-processing aggregation and have been shown to combat the issue of varying quality. However, aggregation methods for post-processing aggregation have not been sufficiently studied in dependency parsing tasks. In an extensive empirical study, we compare different unsupervised post-processing aggregation methods to identify the most suitable dependency tree structure aggregation method.
Authors: Nihar Ranjan Sahoo, Pranamya Prashant Kulkarni, Narjis Asad, Arif Ahmad, Tanu Goyal, Aparna Garimella, Pushpak Bhattacharyya
Abstract: The pervasive influence of social biases in language data has sparked the need for benchmark datasets that capture and evaluate these biases in Large Language Models (LLMs). Existing efforts predominantly focus on English language and the Western context, leaving a void for a reliable dataset that encapsulates India's unique socio-cultural nuances. To bridge this gap, we introduce IndiBias, a comprehensive benchmarking dataset designed specifically for evaluating social biases in the Indian context. We filter and translate the existing CrowS-Pairs dataset to create a benchmark dataset suited to the Indian context in Hindi language. Additionally, we leverage LLMs including ChatGPT and InstructGPT to augment our dataset with diverse societal biases and stereotypes prevalent in India. The included bias dimensions encompass gender, religion, caste, age, region, physical appearance, and occupation. We also build a resource to address intersectional biases along three intersectional dimensions. Our dataset contains 800 sentence pairs and 300 tuples for bias measurement across different demographics. The dataset is available in English and Hindi, providing a size comparable to existing benchmark datasets. Furthermore, using IndiBias we compare ten different language models on multiple bias measurement metrics. We observed that the language models exhibit more bias across a majority of the intersectional groups.
Authors: Nick Mecklenburg, Yiyou Lin, Xiaoxiao Li, Daniel Holstein, Leonardo Nunes, Sara Malvar, Bruno Silva, Ranveer Chandra, Vijay Aski, Pavan Kumar Reddy Yannam, Tolga Aktas, Todd Hendry
Abstract: In recent years, Large Language Models (LLMs) have shown remarkable performance in generating human-like text, proving to be a valuable asset across various applications. However, adapting these models to incorporate new, out-of-domain knowledge remains a challenge, particularly for facts and events that occur after the model's knowledge cutoff date. This paper investigates the effectiveness of Supervised Fine-Tuning (SFT) as a method for knowledge injection in LLMs, specifically focusing on the domain of recent sporting events. We compare different dataset generation strategies -- token-based and fact-based scaling -- to create training data that helps the model learn new information. Our experiments on GPT-4 demonstrate that while token-based scaling can lead to improvements in Q&A accuracy, it may not provide uniform coverage of new knowledge. Fact-based scaling, on the other hand, offers a more systematic approach to ensure even coverage across all facts. We present a novel dataset generation process that leads to more effective knowledge ingestion through SFT, and our results show considerable performance improvements in Q&A tasks related to out-of-domain knowledge. This study contributes to the understanding of domain adaptation for LLMs and highlights the potential of SFT in enhancing the factuality of LLM responses in specific knowledge domains.
Authors: Zhivar Sourati, Meltem Ozcan, Colin McDaniel, Alireza Ziabari, Nuan Wen, Ala Tak, Fred Morstatter, Morteza Dehghani
Abstract: Prior research has established associations between individuals' language usage and their personal traits; our linguistic patterns reveal information about our personalities, emotional states, and beliefs. However, with the increasing adoption of Large Language Models (LLMs) as writing assistants in everyday writing, a critical question emerges: are authors' linguistic patterns still predictive of their personal traits when LLMs are involved in the writing process? We investigate the impact of LLMs on the linguistic markers of demographic and psychological traits, specifically examining three LLMs - GPT3.5, Llama 2, and Gemini - across six different traits: gender, age, political affiliation, personality, empathy, and morality. Our findings indicate that although the use of LLMs slightly reduces the predictive power of linguistic patterns over authors' personal traits, the significant changes are infrequent, and the use of LLMs does not fully diminish the predictive power of authors' linguistic patterns over their personal traits. We also note that some theoretically established lexical-based linguistic markers lose their reliability as predictors when LLMs are used in the writing process. Our findings have important implications for the study of linguistic markers of personal traits in the age of LLMs.
Authors: Zhenyu Hou, Yilin Niu, Zhengxiao Du, Xiaohan Zhang, Xiao Liu, Aohan Zeng, Qinkai Zheng, Minlie Huang, Hongning Wang, Jie Tang, Yuxiao Dong
Abstract: ChatGLM is a free-to-use AI service powered by the ChatGLM family of large language models (LLMs). In this paper, we present the ChatGLM-RLHF pipeline -- a reinforcement learning from human feedback (RLHF) system -- designed to enhance ChatGLM's alignment with human preferences. ChatGLM-RLHF encompasses three major components: the collection of human preference data, the training of the reward model, and the optimization of policies. Throughout the process of integrating ChatGLM-RLHF into production, we encountered and addressed several unprecedented challenges. We introduce the strategies to mitigate reward variance for stabilized large-scale training, implement model parallelism with fused gradient-descent, and design regularization constraints to avoid catastrophic forgetting in LLMs. Experiments show that ChatGLM-RLHF brings significant improvements in alignment tasks compared to the supervised fine-tuned (SFT) version of ChatGLM. For instance, it achieves on average 15\% more wins against ChatGLM-SFT in Chinese alignment tasks. The work presents our practices of aligning LLMs with human preferences, offering insights into the challenges and solutions in RLHF implementations.
Authors: Wei Chen, Zhiyuan Li
Abstract: Language models have shown effectiveness in a variety of software applications, particularly in tasks related to automatic workflow. These models possess the crucial ability to call functions, which is essential in creating AI agents. Despite the high performance of large-scale language models in cloud environments, they are often associated with concerns over privacy and cost. Current on-device models for function calling face issues with latency and accuracy. Our research presents a new method that empowers an on-device model with 2 billion parameters to surpass the performance of GPT-4 in both accuracy and latency, and decrease the context length by 95\%. When compared to Llama-7B with a RAG-based function calling mechanism, our method enhances latency by 35-fold. This method reduces the latency to levels deemed suitable for deployment across a variety of edge devices in production environments, aligning with the performance requisites for real-world applications.
Authors: Advaitha Vetagiri, Gyandeep Kalita, Eisha Halder, Chetna Taparia, Partha Pakray, Riyanka Manna
Abstract: Online gender-based harassment is a widespread issue limiting the free expression and participation of women and marginalized genders in digital spaces. Detecting such abusive content can enable platforms to curb this menace. We participated in the Gendered Abuse Detection in Indic Languages shared task at ICON2023 that provided datasets of annotated Twitter posts in English, Hindi and Tamil for building classifiers to identify gendered abuse. Our team CNLP-NITS-PP developed an ensemble approach combining CNN and BiLSTM networks that can effectively model semantic and sequential patterns in textual data. The CNN captures localized features indicative of abusive language through its convolution filters applied on embedded input text. To determine context-based offensiveness, the BiLSTM analyzes this sequence for dependencies among words and phrases. Multiple variations were trained using FastText and GloVe word embeddings for each language dataset comprising over 7,600 crowdsourced annotations across labels for explicit abuse, targeted minority attacks and general offences. The validation scores showed strong performance across f1-measures, especially for English 0.84. Our experiments reveal how customizing embeddings and model hyperparameters can improve detection capability. The proposed architecture ranked 1st in the competition, proving its ability to handle real-world noisy text with code-switching. This technique has a promising scope as platforms aim to combat cyber harassment facing Indic language internet users. Our Code is at https://github.com/advaithavetagiri/CNLP-NITS-PP
Authors: {\L}ukasz Kuci\'nski, Tomasz Korbak, Pawe{\l} Ko{\l}odziej, Piotr Mi{\l}o\'s
Abstract: Communication is compositional if complex signals can be represented as a combination of simpler subparts. In this paper, we theoretically show that inductive biases on both the training framework and the data are needed to develop a compositional communication. Moreover, we prove that compositionality spontaneously arises in the signaling games, where agents communicate over a noisy channel. We experimentally confirm that a range of noise levels, which depends on the model and the data, indeed promotes compositionality. Finally, we provide a comprehensive study of this dependence and report results in terms of recently studied compositionality metrics: topographical similarity, conflict count, and context independence.
Authors: Zhangcheng Qiang, Weiqing Wang, Kerry Taylor
Abstract: Ontology matching (OM) enables semantic interoperability between different ontologies and resolves their conceptual heterogeneity by aligning related entities. OM systems currently have two prevailing design paradigms: conventional knowledge-based expert systems and newer machine learning-based predictive systems. While large language models (LLMs) and LLM agents have revolutionised data engineering and have been applied creatively in many domains, their potential for OM remains underexplored. This study introduces a novel agent-powered LLM-based design paradigm for OM systems. With consideration of several specific challenges in leveraging LLM agents for OM, we propose a generic framework, namely Agent-OM, consisting of two Siamese agents for retrieval and matching, with a set of simple prompt-based OM tools. Our framework is implemented in a proof-of-concept system. Evaluations of three Ontology Alignment Evaluation Initiative (OAEI) tracks over state-of-the-art OM systems show that our system can achieve results very close to the long-standing best performance on simple OM tasks and can significantly improve the performance on complex and few-shot OM tasks.
Authors: Jialin Wu, Xia Hu, Yaqing Wang, Bo Pang, Radu Soricut
Abstract: Large multi-modal models (LMMs) exhibit remarkable performance across numerous tasks. However, generalist LMMs often suffer from performance degradation when tuned over a large collection of tasks. Recent research suggests that Mixture of Experts (MoE) architectures are useful for instruction tuning, but for LMMs of parameter size around O(50-100B), the prohibitive cost of replicating and storing the expert models severely limits the number of experts we can use. We propose Omni-SMoLA, an architecture that uses the Soft MoE approach to (softly) mix many multimodal low rank experts, and avoids introducing a significant number of new parameters compared to conventional MoE models. The core intuition here is that the large model provides a foundational backbone, while different lightweight experts residually learn specialized knowledge, either per-modality or multimodally. Extensive experiments demonstrate that the SMoLA approach helps improve the generalist performance across a broad range of generative vision-and-language tasks, achieving new SoTA generalist performance that often matches or outperforms single specialized LMM baselines, as well as new SoTA specialist performance.
Authors: Yunshan Ma, Chenchen Ye, Zijian Wu, Xiang Wang, Yixin Cao, Liang Pang, Tat-Seng Chua
Abstract: Temporal complex event forecasting aims to predict the future events given the observed events from history. Most formulations of temporal complex event are unstructured or without extensive temporal information, resulting in inferior representations and limited forecasting capabilities. To bridge these gaps, we innovatively introduce the formulation of Structured, Complex, and Time-complete temporal event (SCTc-TE). Following this comprehensive formulation, we develop a fully automated pipeline and construct a large-scale dataset named MidEast-TE from about 0.6 million news articles. This dataset focuses on the cooperation and conflict events among countries mainly in the MidEast region from 2015 to 2022. Not limited to the dataset construction, more importantly, we advance the forecasting methods by discriminating the crucial roles of various contextual information, i.e., local and global contexts. Thereby, we propose a novel method LoGo that is able to take advantage of both Local and Global contexts for SCTc-TE forecasting. We evaluate our proposed approach on both our proposed MidEast-TE dataset and the original GDELT-TE dataset. Experimental results demonstrate the effectiveness of our forecasting model LoGo. The code and datasets are released via https://github.com/yecchen/GDELT-ComplexEvent.
Authors: Jiashu Xu, Fei Wang, Mingyu Derek Ma, Pang Wei Koh, Chaowei Xiao, Muhao Chen
Abstract: The exorbitant cost of training Large language models (LLMs) from scratch makes it essential to fingerprint the models to protect intellectual property via ownership authentication and to ensure downstream users and developers comply with their license terms (e.g. restricting commercial use). In this study, we present a pilot study on LLM fingerprinting as a form of very lightweight instruction tuning. Model publisher specifies a confidential private key and implants it as an instruction backdoor that causes the LLM to generate specific text when the key is present. Results on 11 popularly-used LLMs showed that this approach is lightweight and does not affect the normal behavior of the model. It also prevents publisher overclaim, maintains robustness against fingerprint guessing and parameter-efficient training, and supports multi-stage fingerprinting akin to MIT License. Code is available in https://cnut1648.github.io/Model-Fingerprint/.
Authors: Shan Yang, Yongfei Zhang
Abstract: Multimodal large language models (MLLM) have achieved satisfactory results in many tasks. However, their performance in the task of person re-identification (ReID) has not been explored to date. This paper will investigate how to adapt them for the task of ReID. An intuitive idea is to fine-tune MLLM with ReID image-text datasets, and then use their visual encoder as a backbone for ReID. However, there still exist two apparent issues: (1) Designing instructions for ReID, MLLMs may overfit specific instructions, and designing a variety of instructions will lead to higher costs. (2) Latent image feature vectors from LLMs are not involved in loss computation. Instructional learning, aligning image-text features, results in indirect optimization and a learning objective that inadequately utilizes features, limiting effectiveness in person feature learning. To address these problems, this paper proposes MLLMReID: Multimodal Large Language Model-based ReID. Firstly, we proposed Common Instruction, a simple approach that leverages the essence ability of LLMs to continue writing, avoiding complex and diverse instruction design. Secondly, we proposed DirectReID, which effectively employs the latent image feature vectors of images outputted by LLMs in ReID tasks. The experimental results demonstrate the superiority of our method. We will open-source the code on GitHub.
Authors: Selina Meyer, David Elsweiler
Abstract: Conversational agents are increasingly used to address emotional needs on top of information needs. One use case of increasing interest are counselling-style mental health and behaviour change interventions, with large language model (LLM)-based approaches becoming more popular. Research in this context so far has been largely system-focused, foregoing the aspect of user behaviour and the impact this can have on LLM-generated texts. To address this issue, we share a dataset containing text-based user interactions related to behaviour change with two GPT-4-based conversational agents collected in a preregistered user study. This dataset includes conversation data, user language analysis, perception measures, and user feedback for LLM-generated turns, and can offer valuable insights to inform the design of such systems based on real interactions.
Authors: Changshu Liu, Shizhuo Dylan Zhang, Ali Reza Ibrahimzada, Reyhaneh Jabbarvand
Abstract: Solely relying on test passing to evaluate Large Language Models (LLMs) for code synthesis may result in unfair assessment or promoting models with data leakage. As an alternative, we introduce CodeMind, a framework designed to gauge the code reasoning abilities of LLMs. CodeMind currently supports three code reasoning tasks: Independent Execution Reasoning (IER), Dependent Execution Reasoning (DER), and Specification Reasoning (SR). The first two evaluate models to predict the execution output of an arbitrary code or code the model could correctly synthesize. The third one evaluates the extent to which LLMs implement the specified expected behavior. Our extensive evaluation of nine LLMs across five benchmarks in two different programming languages using CodeMind shows that LLMs fairly follow control flow constructs and, in general, explain how inputs evolve to output, specifically for simple programs and the ones they can correctly synthesize. However, their performance drops for code with higher complexity, non-trivial logical and arithmetic operators, non-primitive types, and API calls. Furthermore, we observe that, while correlated, specification reasoning (essential for code synthesis) does not imply execution reasoning (essential for broader programming tasks such as testing and debugging): ranking LLMs based on test passing can be different compared to code reasoning.
Authors: Nguyen Nguyen, Jing Bi, Ali Vosoughi, Yapeng Tian, Pooyan Fazli, Chenliang Xu
Abstract: The capability of intelligent models to extrapolate and comprehend changes in object states is a crucial yet demanding aspect of AI research, particularly through the lens of human interaction in real-world settings. This task involves describing complex visual environments, identifying active objects, and interpreting their changes as conveyed through language. Traditional methods, which isolate object captioning and state change detection, offer a limited view of dynamic environments. Moreover, relying on a small set of symbolic words to represent changes has restricted the expressiveness of the language. To address these challenges, in this paper, we introduce the Object State Captioning and State Change Representation (OSCaR) dataset and benchmark. OSCaR consists of 14,084 annotated video segments with nearly 1,000 unique objects from various egocentric video collections. It sets a new testbed for evaluating multimodal large language models (MLLMs). Our experiments demonstrate that while MLLMs show some skill, they lack a full understanding of object state changes. The benchmark includes a fine-tuned model that, despite initial capabilities, requires significant improvements in accuracy and generalization ability for effective understanding of these changes. Our code and dataset are available at https://github.com/nguyennm1024/OSCaR.
Authors: Waris Gill (Virginia Tech, USA), Mohamed Elidrisi (Cisco, USA), Pallavi Kalapatapu (Cisco, USA), Ali Anwar (University of Minnesota, Minneapolis, USA), Muhammad Ali Gulzar (Virginia Tech, USA)
Abstract: Large Language Models (LLMs) like ChatGPT and Llama2 have revolutionized natural language processing and search engine dynamics. However, these models incur exceptionally high computational costs. For instance, GPT-3 consists of 175 billion parameters where inference demands billions of floating-point operations. Caching is a natural solution to reduce LLM inference costs on repeated queries which constitute about 31% of the total queries. However, existing caching methods are incapable of finding semantic similarities among LLM queries, leading to unacceptable false hit-and-miss rates. This paper introduces MeanCache, a user-centric semantic cache for LLMs that identifies semantically similar queries to determine cache hit or miss. Using MeanCache, the response to a user's semantically similar query can be retrieved from a local cache rather than re-querying the LLM, thus reducing costs, service provider load, and environmental impact. Existing caching solutions for LLMs raise privacy and scalability concerns and perform wasteful query requests. MeanCache leverages Federated Learning (FL) to collaboratively train a query similarity model across LLM users without violating privacy. By placing a local cache in each user's device and using FL, MeanCache reduces the latency and costs and enhances model performance, resulting in lower false hit rates. MeanCache compresses the embedding dimensions to minimize cache storage and also finds the optimal cosine similarity threshold. Our experiments benchmarked against the state-of-the-art caching method, reveal that MeanCache attains an approximately 17% higher F-score and a 20% increase in precision during semantic cache hit-and-miss decisions. It also reduces the storage requirement by 83% and accelerates semantic cache hit-and-miss decisions by 11%.
Authors: Teng Xiao, Chao Cui, Huaisheng Zhu, Vasant G. Honavar
Abstract: Recent advancements in biology and chemistry have leveraged multi-modal learning, integrating molecules and their natural language descriptions to enhance drug discovery. However, current pre-training frameworks are limited to two modalities, and designing a unified network to process different modalities (e.g., natural language, 2D molecular graphs, 3D molecular conformations, and 3D proteins) remains challenging due to inherent gaps among them. In this work, we propose MolBind, a framework that trains encoders for multiple modalities through contrastive learning, mapping all modalities to a shared feature space for multi-modal semantic alignment. To facilitate effective pre-training of MolBind on multiple modalities, we also build and collect a high-quality dataset with four modalities, MolBind-M4, including graph-language, conformation-language, graph-conformation, and conformation-protein paired data. MolBind shows superior zero-shot learning performance across a wide range of tasks, demonstrating its strong capability of capturing the underlying semantics of multiple modalities.
Authors: Zhouhao Sun, Xiao Ding, Li Du, Bibo Cai, Jinglong Gao, Ting Liu, Qin Bing
Abstract: Large language models (LLMs) have achieved significant performance in various natural language reasoning tasks. However, they still struggle with performing first-order logic reasoning over formal logical theories expressed in natural language. This is because the previous LLMs-based reasoning systems have the theoretical incompleteness issue. As a result, it can only address a limited set of simple reasoning problems, which significantly decreases their generalization ability. To address this issue, we propose a novel framework, named Generalizable and Faithful Reasoner (GFaiR), which introduces the paradigm of resolution refutation. Resolution refutation has the capability to solve all first-order logic reasoning problems by extending reasoning rules and employing the principle of proof by contradiction, so our system's completeness can be improved by introducing resolution refutation. Experimental results demonstrate that our system outperforms previous works by achieving state-of-the-art performances in complex scenarios while maintaining performances in simple scenarios. Besides, we observe that GFaiR is faithful to its reasoning process.