Authors: Bradley P. Allen, Fina Polat, Paul Groth
Abstract: We describe the University of Amsterdam Intelligent Data Engineering Lab team's entry for the SemEval-2024 Task 6 competition. The SHROOM-INDElab system builds on previous work on using prompt programming and in-context learning with large language models (LLMs) to build classifiers for hallucination detection, and extends that work through the incorporation of context-specific definition of task, role, and target concept, and automated generation of examples for use in a few-shot prompting approach. The resulting system achieved fourth-best and sixth-best performance in the model-agnostic track and model-aware tracks for Task 6, respectively, and evaluation using the validation sets showed that the system's classification decisions were consistent with those of the crowd-sourced human labellers. We further found that a zero-shot approach provided better accuracy than a few-shot approach using automatically generated examples. Code for the system described in this paper is available on Github.
Authors: Zhangdie Yuan, Chenxi Whitehouse, Eric Chamoun, Rami Aly, Andreas Vlachos
Abstract: This paper introduces PRobELM (Plausibility Ranking Evaluation for Language Models), a benchmark designed to assess language models' ability to discern more plausible from less plausible scenarios through their parametric knowledge. While benchmarks such as TruthfulQA emphasise factual accuracy or truthfulness, and others such as COPA explore plausible scenarios without explicitly incorporating world knowledge, PRobELM seeks to bridge this gap by evaluating models' capabilities to prioritise plausible scenarios that leverage world knowledge over less plausible alternatives. This design allows us to assess the potential of language models for downstream use cases such as literature-based discovery where the focus is on identifying information that is likely but not yet known. Our benchmark is constructed from a dataset curated from Wikidata edit histories, tailored to align the temporal bounds of the training data for the evaluated models. PRobELM facilitates the evaluation of language models across multiple prompting types, including statement, text completion, and question-answering. Experiments with 10 models of various sizes and architectures on the relationship between model scales, training recency, and plausibility performance, reveal that factual accuracy does not directly correlate with plausibility performance and that up-to-date training data enhances plausibility assessment across different model architectures.
Authors: Makesh Narsimhan Sreedhar, Traian Rebedea, Shaona Ghosh, Christopher Parisien
Abstract: Recent advancements in instruction-tuning datasets have predominantly focused on specific tasks like mathematical or logical reasoning. There has been a notable gap in data designed for aligning language models to maintain topic relevance in conversations - a critical aspect for deploying chatbots to production. We introduce the CantTalkAboutThis dataset to help language models remain focused on the subject at hand during task-oriented interactions. It consists of synthetic dialogues on a wide range of conversation topics from different domains. These dialogues are interspersed with distractor turns that intentionally divert the chatbot from the predefined topic. Fine-tuning language models on this dataset helps make them resilient to deviating from the role assigned and improves their ability to maintain topical coherence compared to general-purpose instruction-tuned LLMs like GPT-4-turbo and Mixtral-Instruct. Additionally, preliminary observations suggest that training models on this dataset also enhance their performance on fine-grained instruction following tasks.
Authors: Jingyu Zhang, Marc Marone, Tianjian Li, Benjamin Van Durme, Daniel Khashabi
Abstract: For humans to trust the fluent generations of large language models (LLMs), they must be able to verify their correctness against trusted, external sources. Recent efforts aim to increase verifiability through citations of retrieved documents or post-hoc provenance. However, such citations are prone to mistakes that further complicate their verifiability. To address these limitations, we tackle the verifiability goal with a different philosophy: we trivialize the verification process by developing models that quote verbatim statements from trusted sources in pre-training data. We propose Quote-Tuning, which demonstrates the feasibility of aligning LLMs to leverage memorized information and quote from pre-training data. Quote-Tuning quantifies quoting against large corpora with efficient membership inference tools, and uses the amount of quotes as an implicit reward signal to construct a synthetic preference dataset for quoting, without any human annotation. Next, the target model is aligned to quote using preference optimization algorithms. Experimental results show that Quote-Tuning significantly increases the percentage of LLM generation quoted verbatim from high-quality pre-training documents by 55% to 130% relative to untuned models while maintaining response quality. Further experiments demonstrate that Quote-Tuning generalizes quoting to out-of-domain data, is applicable in different tasks, and provides additional benefits to truthfulness. Quote-Tuning not only serves as a hassle-free method to increase quoting but also opens up avenues for improving LLM trustworthiness through better verifiability.
Authors: Ajay Jaiswal, Bodun Hu, Lu Yin, Yeonju Ro, Shiwei Liu, Tianlong Chen, Aditya Akella
Abstract: Autoregressive Large Language Models (e.g., LLaMa, GPTs) are omnipresent achieving remarkable success in language understanding and generation. However, such impressive capability typically comes with a substantial model size, which presents significant challenges for autoregressive token-by-token generation. To mitigate computation overload incurred during generation, several early-exit and layer-dropping strategies have been proposed. Despite some promising success due to the redundancy across LLMs layers on metrics like Rough-L/BLUE, our careful knowledge-intensive evaluation unveils issues such as generation collapse, hallucination of wrong facts, and noticeable performance drop even at the trivial exit ratio of 10-15% of layers. We attribute these errors primarily to ineffective handling of the KV cache through state copying during early-exit. In this work, we observed the saturation of computationally expensive feed-forward blocks of LLM layers and proposed FFN-SkipLLM, which is a novel fine-grained skip strategy of autoregressive LLMs. More specifically, FFN-SkipLLM is an input-adaptive feed-forward skipping strategy that can skip 25-30% of FFN blocks of LLMs with marginal change in performance on knowledge-intensive generation tasks without any requirement to handle KV cache. Our extensive experiments and ablation across benchmarks like MT-Bench, Factoid-QA, and variable-length text summarization illustrate how our simple and ease-at-use method can facilitate faster autoregressive decoding.
Authors: Bowen Zhang, Harold Soh
Abstract: In this work, we are interested in automated methods for knowledge graph creation (KGC) from input text. Progress on large language models (LLMs) has prompted a series of recent works applying them to KGC, e.g., via zero/few-shot prompting. Despite successes on small domain-specific datasets, these models face difficulties scaling up to text common in many real-world applications. A principal issue is that in prior methods, the KG schema has to be included in the LLM prompt to generate valid triplets; larger and more complex schema easily exceed the LLMs' context window length. To address this problem, we propose a three-phase framework named Extract-Define-Canonicalize (EDC): open information extraction followed by schema definition and post-hoc canonicalization. EDC is flexible in that it can be applied to settings where a pre-defined target schema is available and when it is not; in the latter case, it constructs a schema automatically and applies self-canonicalization. To further improve performance, we introduce a trained component that retrieves schema elements relevant to the input text; this improves the LLMs' extraction performance in a retrieval-augmented generation-like manner. We demonstrate on three KGC benchmarks that EDC is able to extract high-quality triplets without any parameter tuning and with significantly larger schemas compared to prior works.
Authors: Xiaocheng Luo, Yanping Chen, Ruixue Tang, Ruizhang Huang, Yongbin Qin
Abstract: Current methods to extract relational triples directly make a prediction based on a possible entity pair in a raw sentence without depending on entity recognition. The task suffers from a serious semantic overlapping problem, in which several relation triples may share one or two entities in a sentence. It is weak to learn discriminative semantic features relevant to a relation triple. In this paper, based on a two-dimensional sentence representation, a bi-consolidating model is proposed to address this problem by simultaneously reinforcing the local and global semantic features relevant to a relation triple. This model consists of a local consolidation component and a global consolidation component. The first component uses a pixel difference convolution to enhance semantic information of a possible triple representation from adjacent regions and mitigate noise in neighbouring neighbours. The second component strengthens the triple representation based a channel attention and a spatial attention, which has the advantage to learn remote semantic dependencies in a sentence. They are helpful to improve the performance of both entity identification and relation type classification in relation triple extraction. After evaluated on several publish datasets, it achieves competitive performance. Analytical experiments demonstrate the effectiveness of our model for relational triple extraction and give motivation for other natural language processing tasks.
Authors: Hyeonwoo Kim, Gyoungjin Gim, Yungi Kim, Jihoo Kim, Byungju Kim, Wonseok Lee, Chanjun Park
Abstract: This study presents a novel learning approach designed to enhance both mathematical reasoning and problem-solving abilities of Large Language Models (LLMs). We focus on integrating the Chain-of-Thought (CoT) and the Program-of-Thought (PoT) learning, hypothesizing that prioritizing the learning of mathematical reasoning ability is helpful for the amplification of problem-solving ability. Thus, the initial learning with CoT is essential for solving challenging mathematical problems. To this end, we propose a sequential learning approach, named SAAS (Solving Ability Amplification Strategy), which strategically transitions from CoT learning to PoT learning. Our empirical study, involving an extensive performance comparison using several benchmarks, demonstrates that our SAAS achieves state-of-the-art (SOTA) performance. The results underscore the effectiveness of our sequential learning approach, marking a significant advancement in the field of mathematical reasoning in LLMs.
Authors: Fred Philippy, Shohreh Haddadan, Siwen Guo
Abstract: In NLP, zero-shot classification (ZSC) is the task of assigning labels to textual data without any labeled examples for the target classes. A common method for ZSC is to fine-tune a language model on a Natural Language Inference (NLI) dataset and then use it to infer the entailment between the input document and the target labels. However, this approach faces certain challenges, particularly for languages with limited resources. In this paper, we propose an alternative solution that leverages dictionaries as a source of data for ZSC. We focus on Luxembourgish, a low-resource language spoken in Luxembourg, and construct two new topic relevance classification datasets based on a dictionary that provides various synonyms, word translations and example sentences. We evaluate the usability of our dataset and compare it with the NLI-based approach on two topic classification tasks in a zero-shot manner. Our results show that by using the dictionary-based dataset, the trained models outperform the ones following the NLI-based approach for ZSC. While we focus on a single low-resource language in this study, we believe that the efficacy of our approach can also transfer to other languages where such a dictionary is available.
Authors: Bowen Zhang, Kehua Chang, Chunping Li
Abstract: Sentence Embedding stands as a fundamental task within the realm of Natural Language Processing, finding extensive application in search engines, expert systems, and question-and-answer platforms. With the continuous evolution of large language models such as LLaMA and Mistral, research on sentence embedding has recently achieved notable breakthroughs. However, these advancements mainly pertain to fine-tuning scenarios, leaving explorations into computationally efficient direct inference methods for sentence representation in a nascent stage. This paper endeavors to bridge this research gap. Through comprehensive experimentation, we challenge the widely held belief in the necessity of an Explicit One-word Limitation for deriving sentence embeddings from Pre-trained Language Models (PLMs). We demonstrate that this approach, while beneficial for generative models under direct inference scenario, is not imperative for discriminative models or the fine-tuning of generative PLMs. This discovery sheds new light on the design of manual templates in future studies. Building upon this insight, we propose two innovative prompt engineering techniques capable of further enhancing the expressive power of PLMs' raw embeddings: Pretended Chain of Thought and Knowledge Enhancement. We confirm their effectiveness across various PLM types and provide a detailed exploration of the underlying factors contributing to their success.
Authors: Gulsum Yigit, Mehmet Fatih Amasyali
Abstract: Math Word Problem (MWP) solving presents a challenging task in Natural Language Processing (NLP). This study aims to provide MWP solvers with a more diverse training set, ultimately improving their ability to solve various math problems. We propose several methods for data augmentation by modifying the problem texts and equations, such as synonym replacement, rule-based: question replacement, and rule based: reversing question methodologies over two English MWP datasets. This study extends by introducing a new in-context learning augmentation method, employing the Llama-7b language model. This approach involves instruction-based prompting for rephrasing the math problem texts. Performance evaluations are conducted on 9 baseline models, revealing that augmentation methods outperform baseline models. Moreover, concatenating examples generated by various augmentation methods further improves performance.
Authors: Mathilde Aguiar, Pierre Zweigenbaum, Nona Naderi
Abstract: This paper describes our submission to Task 2 of SemEval-2024: Safe Biomedical Natural Language Inference for Clinical Trials. The Multi-evidence Natural Language Inference for Clinical Trial Data (NLI4CT) consists of a Textual Entailment (TE) task focused on the evaluation of the consistency and faithfulness of Natural Language Inference (NLI) models applied to Clinical Trial Reports (CTR). We test 2 distinct approaches, one based on finetuning and ensembling Masked Language Models and the other based on prompting Large Language Models using templates, in particular, using Chain-Of-Thought and Contrastive Chain-Of-Thought. Prompting Flan-T5-large in a 2-shot setting leads to our best system that achieves 0.57 F1 score, 0.64 Faithfulness, and 0.56 Consistency.
Authors: Myrthe Reuver, Suzan Verberne, Antske Fokkens
Abstract: For a viewpoint-diverse news recommender, identifying whether two news articles express the same viewpoint is essential. One way to determine "same or different" viewpoint is stance detection. In this paper, we investigate the robustness of operationalization choices for few-shot stance detection, with special attention to modelling stance across different topics. Our experiments test pre-registered hypotheses on stance detection. Specifically, we compare two stance task definitions (Pro/Con versus Same Side Stance), two LLM architectures (bi-encoding versus cross-encoding), and adding Natural Language Inference knowledge, with pre-trained RoBERTa models trained with shots of 100 examples from 7 different stance detection datasets. Some of our hypotheses and claims from earlier work can be confirmed, while others give more inconsistent results. The effect of the Same Side Stance definition on performance differs per dataset and is influenced by other modelling choices. We found no relationship between the number of training topics in the training shots and performance. In general, cross-encoding out-performs bi-encoding, and adding NLI training to our models gives considerable improvement, but these results are not consistent across all datasets. Our results indicate that it is essential to include multiple datasets and systematic modelling experiments when aiming to find robust modelling choices for the concept `stance'.
Authors: Ran Zmigrod, Dongsheng Wang, Mathieu Sibue, Yulong Pei, Petr Babkin, Ivan Brugere, Xiaomo Liu, Nacho Navarro, Antony Papadimitriou, William Watson, Zhiqiang Ma, Armineh Nourbakhsh, Sameena Shah
Abstract: The field of visually rich document understanding (VRDU) aims to solve a multitude of well-researched NLP tasks in a multi-modal domain. Several datasets exist for research on specific tasks of VRDU such as document classification (DC), key entity extraction (KEE), entity linking, visual question answering (VQA), inter alia. These datasets cover documents like invoices and receipts with sparse annotations such that they support one or two co-related tasks (e.g., entity extraction and entity linking). Unfortunately, only focusing on a single specific of documents or task is not representative of how documents often need to be processed in the wild - where variety in style and requirements is expected. In this paper, we introduce BuDDIE (Business Document Dataset for Information Extraction), the first multi-task dataset of 1,665 real-world business documents that contains rich and dense annotations for DC, KEE, and VQA. Our dataset consists of publicly available business entity documents from US state government websites. The documents are structured and vary in their style and layout across states and types (e.g., forms, certificates, reports, etc.). We provide data variety and quality metrics for BuDDIE as well as a series of baselines for each task. Our baselines cover traditional textual, multi-modal, and large language model approaches to VRDU.
Authors: Yuri Bizzoni, Pascale Feldkamp, Ida Marie Lassen, Mia Jacobsen, Mads Rosendahl Thomsen, Kristoffer Nielbo
Abstract: In this study, we employ a classification approach to show that different categories of literary "quality" display unique linguistic profiles, leveraging a corpus that encompasses titles from the Norton Anthology, Penguin Classics series, and the Open Syllabus project, contrasted against contemporary bestsellers, Nobel prize winners and recipients of prestigious literary awards. Our analysis reveals that canonical and so called high-brow texts exhibit distinct textual features when compared to other quality categories such as bestsellers and popular titles as well as to control groups, likely responding to distinct (but not mutually exclusive) models of quality. We apply a classic machine learning approach, namely Random Forest, to distinguish quality novels from "control groups", achieving up to 77\% F1 scores in differentiating between the categories. We find that quality category tend to be easier to distinguish from control groups than from other quality categories, suggesting than literary quality features might be distinguishable but shared through quality proxies.
Authors: Annerose Eichel, Tana Deeg, Andr\'e Blessing, Milena Belosevic, Sabine Arndt-Lappe, Sabine Schulte im Walde
Abstract: We present a comprehensive computational study of the under-investigated phenomenon of personal name compounds (PNCs) in German such as Willkommens-Merkel ('Welcome-Merkel'). Prevalent in news, social media, and political discourse, PNCs are hypothesized to exhibit an evaluative function that is reflected in a more positive or negative perception as compared to the respective personal full name (such as Angela Merkel). We model 321 PNCs and their corresponding full names at discourse level, and show that PNCs bear an evaluative nature that can be captured through a variety of computational methods. Specifically, we assess through valence information whether a PNC is more positively or negatively evaluative than the person's name, by applying and comparing two approaches using (i) valence norms and (ii) pretrained language models (PLMs). We further enrich our data with personal, domain-specific, and extra-linguistic information and perform a range of regression analyses revealing that factors including compound and modifier valence, domain, and political party membership influence how a PNC is evaluated.
Authors: Annerose Eichel, Sabine Schulte im Walde
Abstract: We present a novel dataset for physical and abstract plausibility of events in English. Based on naturally occurring sentences extracted from Wikipedia, we infiltrate degrees of abstractness, and automatically generate perturbed pseudo-implausible events. We annotate a filtered and balanced subset for plausibility using crowd-sourcing, and perform extensive cleansing to ensure annotation quality. In-depth quantitative analyses indicate that annotators favor plausibility over implausibility and disagree more on implausible events. Furthermore, our plausibility dataset is the first to capture abstractness in events to the same extent as concreteness, and we find that event abstractness has an impact on plausibility ratings: more concrete event participants trigger a perception of implausibility.
Authors: Hele-Andra Kuulmets, Taido Purason, Agnes Luhtaru, Mark Fishel
Abstract: This paper explores cost-efficient methods to adapt pretrained Large Language Models (LLMs) to new lower-resource languages, with a specific focus on Estonian. Leveraging the Llama 2 model, we investigate the impact of combining cross-lingual instruction-tuning with additional monolingual pretraining. Our results demonstrate that even a relatively small amount of additional monolingual pretraining followed by cross-lingual instruction-tuning significantly enhances results on Estonian. Furthermore, we showcase cross-lingual knowledge transfer from high-quality English instructions to Estonian, resulting in improvements in commonsense reasoning and multi-turn conversation capabilities. Our best model, named \textsc{Llammas}, represents the first open-source instruction-following LLM for Estonian. Additionally, we publish Alpaca-est, the first general task instruction dataset for Estonia. These contributions mark the initial progress in the direction of developing open-source LLMs for Estonian.
Authors: Amin Dada, Marie Bauer, Amanda Butler Contreras, Osman Alperen Kora\c{s}, Constantin Marc Seibold, Kaleb E Smith, Jens Kleesiek
Abstract: Large Language Models (LLMs) have shown the potential to significantly contribute to patient care, diagnostics, and administrative processes. Emerging biomedical LLMs address healthcare-specific challenges, including privacy demands and computational constraints. However, evaluation of these models has primarily been limited to non-clinical tasks, which do not reflect the complexity of practical clinical applications. Additionally, there has been no thorough comparison between biomedical and general-domain LLMs for clinical tasks. To fill this gap, we present the Clinical Language Understanding Evaluation (CLUE), a benchmark tailored to evaluate LLMs on real-world clinical tasks. CLUE includes two novel datasets derived from MIMIC IV discharge letters and four existing tasks designed to test the practical applicability of LLMs in healthcare settings. Our evaluation covers several biomedical and general domain LLMs, providing insights into their clinical performance and applicability. CLUE represents a step towards a standardized approach to evaluating and developing LLMs in healthcare to align future model development with the real-world needs of clinical application. We publish our evaluation and data generation scripts: https://github.com/dadaamin/CLUE
Authors: Filip Seitl, Tom\'a\v{s} Kov\'a\v{r}\'ik, Soheyla Mirshahi, Jan Kry\v{s}t\r{u}fek, Rastislav Dujava, Mat\'u\v{s} Ondrei\v{c}ka, Herbert Ullrich, Petr Gronat
Abstract: Advances in large language models have notably enhanced the efficiency of information extraction from unstructured and semi-structured data sources. As these technologies become integral to various applications, establishing an objective measure for the quality of information extraction becomes imperative. However, the scarcity of labeled data presents significant challenges to this endeavor. In this paper, we introduce an automatic framework to assess the quality of the information extraction and its completeness. The framework focuses on information extraction in the form of entity and its properties. We discuss how to handle the input/output size limitations of the large language models and analyze their performance when iteratively extracting the information. Finally, we introduce metrics to evaluate the quality of the extraction and provide an extensive discussion on how to interpret the metrics.
Authors: Barkavi Sundararajan, Somayajulu Sripada, Ehud Reiter
Abstract: Neural Table-to-Text models tend to hallucinate, producing texts that contain factual errors. We investigate whether such errors in the output can be traced back to problems with the input. We manually annotated 1,837 texts generated by multiple models in the politics domain of the ToTTo dataset. We identify the input problems that are responsible for many output errors and show that fixing these inputs reduces factual errors by between 52% and 76% (depending on the model). In addition, we observe that models struggle in processing tabular inputs that are structured in a non-standard way, particularly when the input lacks distinct row and column values or when the column headers are not correctly mapped to corresponding values.
Authors: Jacek Wiland, Max Ploner, Alan Akbik
Abstract: Knowledge probing assesses to which degree a language model (LM) has successfully learned relational knowledge during pre-training. Probing is an inexpensive way to compare LMs of different sizes and training configurations. However, previous approaches rely on the objective function used in pre-training LMs and are thus applicable only to masked or causal LMs. As a result, comparing different types of LMs becomes impossible. To address this, we propose an approach that uses an LM's inherent ability to estimate the log-likelihood of any given textual statement. We carefully design an evaluation dataset of 7,731 instances (40,916 in a larger variant) from which we produce alternative statements for each relational fact, one of which is correct. We then evaluate whether an LM correctly assigns the highest log-likelihood to the correct statement. Our experimental evaluation of 22 common LMs shows that our proposed framework, BEAR, can effectively probe for knowledge across different LM types. We release the BEAR datasets and an open-source framework that implements the probing approach to the research community to facilitate the evaluation and development of LMs.
Authors: Xinrun Du, Zhouliang Yu, Songyang Gao, Ding Pan, Yuyang Cheng, Ziyang Ma, Ruibin Yuan, Xingwei Qu, Jiaheng Liu, Tianyu Zheng, Xinchen Luo, Guorui Zhou, Binhang Yuan, Wenhu Chen, Jie Fu, Ge Zhang
Abstract: In this study, we introduce CT-LLM, a 2B large language model (LLM) that illustrates a pivotal shift towards prioritizing the Chinese language in developing LLMs. Uniquely initiated from scratch, CT-LLM diverges from the conventional methodology by primarily incorporating Chinese textual data, utilizing an extensive corpus of 1,200 billion tokens, including 800 billion Chinese tokens, 300 billion English tokens, and 100 billion code tokens. This strategic composition facilitates the model's exceptional proficiency in understanding and processing Chinese, a capability further enhanced through alignment techniques. Demonstrating remarkable performance on the CHC-Bench, CT-LLM excels in Chinese language tasks, and showcases its adeptness in English through SFT. This research challenges the prevailing paradigm of training LLMs predominantly on English corpora and then adapting them to other languages, broadening the horizons for LLM training methodologies. By open-sourcing the full process of training a Chinese LLM, including a detailed data processing procedure with the obtained Massive Appropriate Pretraining Chinese Corpus (MAP-CC), a well-chosen multidisciplinary Chinese Hard Case Benchmark (CHC-Bench), and the 2B-size Chinese Tiny LLM (CT-LLM), we aim to foster further exploration and innovation in both academia and industry, paving the way for more inclusive and versatile language models.
Authors: Ilya Ilyankou, Aldo Lipani, Stefano Cavazzi, Xiaowei Gao, James Haworth
Abstract: Sentence transformers are language models designed to perform semantic search. This study investigates the capacity of sentence transformers, fine-tuned on general question-answering datasets for asymmetric semantic search, to associate descriptions of human-generated routes across Great Britain with queries often used to describe hiking experiences. We find that sentence transformers have some zero-shot capabilities to understand quasi-geospatial concepts, such as route types and difficulty, suggesting their potential utility for routing recommendation systems.
Authors: Diyi Yang, Caleb Ziems, William Held, Omar Shaikh, Michael S. Bernstein, John Mitchell
Abstract: People rely on social skills like conflict resolution to communicate effectively and to thrive in both work and personal life. However, practice environments for social skills are typically out of reach for most people. How can we make social skill training more available, accessible, and inviting? Drawing upon interdisciplinary research from communication and psychology, this perspective paper identifies social skill barriers to enter specialized fields. Then we present a solution that leverages large language models for social skill training via a generic framework. Our AI Partner, AI Mentor framework merges experiential learning with realistic practice and tailored feedback. This work ultimately calls for cross-disciplinary innovation to address the broader implications for workforce development and social equality.
Authors: Tong Su, Xin Peng, Sarubi Thillainathan, David Guzm\'an, Surangika Ranathunga, En-Shiun Annie Lee
Abstract: Parameter-efficient fine-tuning (PEFT) methods are increasingly vital in adapting large-scale pre-trained language models for diverse tasks, offering a balance between adaptability and computational efficiency. They are important in Low-Resource Language (LRL) Neural Machine Translation (NMT) to enhance translation accuracy with minimal resources. However, their practical effectiveness varies significantly across different languages. We conducted comprehensive empirical experiments with varying LRL domains and sizes to evaluate the performance of 8 PEFT methods with in total of 15 architectures using the SacreBLEU score. We showed that 6 PEFT architectures outperform the baseline for both in-domain and out-domain tests and the Houlsby+Inversion adapter has the best performance overall, proving the effectiveness of PEFT methods.
Authors: Harsh Kohli, Helian Feng, Nicholas Dronen, Calvin McCarter, Sina Moeini, Ali Kebarighotbi
Abstract: In contemporary machine learning approaches to bilingual lexicon induction (BLI), a model learns a mapping between the embedding spaces of a language pair. Recently, retrieve-and-rank approach to BLI has achieved state of the art results on the task. However, the problem remains challenging in low-resource settings, due to the paucity of data. The task is complicated by factors such as lexical variation across languages. We argue that the incorporation of additional lexical information into the recent retrieve-and-rank approach should improve lexicon induction. We demonstrate the efficacy of our proposed approach on XLING, improving over the previous state of the art by an average of 2\% across all language pairs.
Authors: Tianqi Zhong, Zhaoyi Li, Quan Wang, Linqi Song, Ying Wei, Defu Lian, Zhendong Mao
Abstract: Compositional generalization, representing the model's ability to generate text with new attribute combinations obtained by recombining single attributes from the training data, is a crucial property for multi-aspect controllable text generation (MCTG) methods. Nonetheless, a comprehensive compositional generalization evaluation benchmark of MCTG is still lacking. We propose CompMCTG, a benchmark encompassing diverse multi-aspect labeled datasets and a crafted three-dimensional evaluation protocol, to holistically evaluate the compositional generalization of MCTG approaches. We observe that existing MCTG works generally confront a noticeable performance drop in compositional testing. To mitigate this issue, we introduce Meta-MCTG, a training framework incorporating meta-learning, where we enable models to learn how to generalize by simulating compositional generalization scenarios in the training phase. We demonstrate the effectiveness of Meta-MCTG through achieving obvious improvement (by at most 3.64%) for compositional testing performance in 94.4% cases.
Authors: Harsh Kohli, Huan Sun
Abstract: The rapid progress of large language models (LLMs) has seen them excel and frequently surpass human performance on standard benchmarks. This has enabled many downstream applications, such as LLM agents, to rely on their sophisticated reasoning to navigate complex task requirements. However, LLMs are known to unexpectedly falter in simple tasks and under seemingly straightforward circumstances - underscoring the need for better and more diverse evaluation setups to measure their true capabilities. To this end, we choose to study compositional and conditional reasoning, two cornerstones of human cognition, and introduce GroundCocoa - a lexically diverse benchmark connecting these reasoning skills to the real-world problem of flight booking. Our task involves aligning detailed user preferences with available flight options presented in a multiple-choice format. Results indicate a significant disparity in performance among current state-of-the-art LLMs with even the best performing model, GPT-4 Turbo, not exceeding 67% accuracy despite advanced prompting techniques.
Authors: Iosif Iulian Petrila
Abstract: The informational synthesis of neural structures, processes, parameters and characteristics that allow a unified description and modeling as neural machines of natural and artificial neural systems is presented. The general informational parameters as the global quantitative measure of the neural systems computing potential as absolute and relative neural power were proposed. Neural information organizing and processing follows the way in which nature manages neural information by developing functions, functionalities and circuits related to different internal or peripheral components and also to the whole system through a non-deterministic memorization, fragmentation and aggregation of afferent and efferent information, deep neural information processing representing multiple alternations of fragmentation and aggregation stages. The relevant neural characteristics were integrated into a neural machine type model that incorporates unitary also peripheral or interface components as the central ones. The proposed approach allows overcoming the technical constraints in artificial computational implementations of neural information processes and also provides a more relevant description of natural ones.
Authors: Kanishk Gandhi, Denise Lee, Gabriel Grand, Muxin Liu, Winson Cheng, Archit Sharma, Noah D. Goodman
Abstract: Language models are rarely shown fruitful mistakes while training. They then struggle to look beyond the next token, suffering from a snowballing of errors and struggling to predict the consequence of their actions several steps ahead. In this paper, we show how language models can be taught to search by representing the process of search in language, as a flattened string -- a stream of search (SoS). We propose a unified language for search that captures an array of different symbolic search strategies. We demonstrate our approach using the simple yet difficult game of Countdown, where the goal is to combine input numbers with arithmetic operations to reach a target number. We pretrain a transformer-based language model from scratch on a dataset of streams of search generated by heuristic solvers. We find that SoS pretraining increases search accuracy by 25% over models trained to predict only the optimal search trajectory. We further finetune this model with two policy improvement methods: Advantage-Induced Policy Alignment (APA) and Self-Taught Reasoner (STaR). The finetuned SoS models solve 36% of previously unsolved problems, including problems that cannot be solved by any of the heuristic solvers. Our results indicate that language models can learn to solve problems via search, self-improve to flexibly use different search strategies, and potentially discover new ones.
Authors: Corby Rosset, Ching-An Cheng, Arindam Mitra, Michael Santacroce, Ahmed Awadallah, Tengyang Xie
Abstract: This paper studies post-training large language models (LLMs) using preference feedback from a powerful oracle to help a model iteratively improve over itself. The typical approach for post-training LLMs involves Reinforcement Learning from Human Feedback (RLHF), which traditionally separates reward learning and subsequent policy optimization. However, such a reward maximization approach is limited by the nature of "point-wise" rewards (such as Bradley-Terry model), which fails to express complex intransitive or cyclic preference relations. While advances on RLHF show reward learning and policy optimization can be merged into a single contrastive objective for stability, they yet still remain tethered to the reward maximization framework. Recently, a new wave of research sidesteps the reward maximization presumptions in favor of directly optimizing over "pair-wise" or general preferences. In this paper, we introduce Direct Nash Optimization (DNO), a provable and scalable algorithm that marries the simplicity and stability of contrastive learning with theoretical generality from optimizing general preferences. Because DNO is a batched on-policy algorithm using a regression-based objective, its implementation is straightforward and efficient. Moreover, DNO enjoys monotonic improvement across iterations that help it improve even over a strong teacher (such as GPT-4). In our experiments, a resulting 7B parameter Orca-2.5 model aligned by DNO achieves the state-of-the-art win-rate against GPT-4-Turbo of 33% on AlpacaEval 2.0 (even after controlling for response length), an absolute gain of 26% (7% to 33%) over the initializing model. It outperforms models with far more parameters, including Mistral Large, Self-Rewarding LM (70B parameters), and older versions of GPT-4.
Authors: Mahjabin Nahar, Haeseung Seo, Eun-Ju Lee, Aiping Xiong, Dongwon Lee
Abstract: The widespread adoption and transformative effects of large language models (LLMs) have sparked concerns regarding their capacity to produce inaccurate and fictitious content, referred to as `hallucinations'. Given the potential risks associated with hallucinations, humans should be able to identify them. This research aims to understand the human perception of LLM hallucinations by systematically varying the degree of hallucination (genuine, minor hallucination, major hallucination) and examining its interaction with warning (i.e., a warning of potential inaccuracies: absent vs. present). Participants (N=419) from Prolific rated the perceived accuracy and engaged with content (e.g., like, dislike, share) in a Q/A format. Results indicate that humans rank content as truthful in the order genuine > minor hallucination > major hallucination and user engagement behaviors mirror this pattern. More importantly, we observed that warning improves hallucination detection without significantly affecting the perceived truthfulness of genuine content. We conclude by offering insights for future tools to aid human detection of hallucinations.
Authors: Kaustubh Dhole, Eugene Agichtein
Abstract: Query Reformulation(QR) is a set of techniques used to transform a user's original search query to a text that better aligns with the user's intent and improves their search experience. Recently, zero-shot QR has been shown to be a promising approach due to its ability to exploit knowledge inherent in large language models. By taking inspiration from the success of ensemble prompting strategies which have benefited many tasks, we investigate if they can help improve query reformulation. In this context, we propose an ensemble based prompting technique, GenQREnsemble which leverages paraphrases of a zero-shot instruction to generate multiple sets of keywords ultimately improving retrieval performance. We further introduce its post-retrieval variant, GenQREnsembleRF to incorporate pseudo relevant feedback. On evaluations over four IR benchmarks, we find that GenQREnsemble generates better reformulations with relative nDCG@10 improvements up to 18% and MAP improvements upto 24% over the previous zero-shot state-of-art. On the MSMarco Passage Ranking task, GenQREnsembleRF shows relative gains of 5% MRR using pseudo-relevance feedback, and 9% nDCG@10 using relevant feedback documents.
Authors: Zahra Mousavi, Chadni Islam, Kristen Moore, Alsharif Abuadbba, Muhammad Ali Babar
Abstract: The increasing trend of using Large Language Models (LLMs) for code generation raises the question of their capability to generate trustworthy code. While many researchers are exploring the utility of code generation for uncovering software vulnerabilities, one crucial but often overlooked aspect is the security Application Programming Interfaces (APIs). APIs play an integral role in upholding software security, yet effectively integrating security APIs presents substantial challenges. This leads to inadvertent misuse by developers, thereby exposing software to vulnerabilities. To overcome these challenges, developers may seek assistance from LLMs. In this paper, we systematically assess ChatGPT's trustworthiness in code generation for security API use cases in Java. To conduct a thorough evaluation, we compile an extensive collection of 48 programming tasks for 5 widely used security APIs. We employ both automated and manual approaches to effectively detect security API misuse in the code generated by ChatGPT for these tasks. Our findings are concerning: around 70% of the code instances across 30 attempts per task contain security API misuse, with 20 distinct misuse types identified. Moreover, for roughly half of the tasks, this rate reaches 100%, indicating that there is a long way to go before developers can rely on ChatGPT to securely implement security API code.
Authors: Zitao Shuai, Liyue Shen
Abstract: Vision-language pre-training (VLP) has arised as an efficient scheme for multimodal representation learning, but it requires large-scale multimodal data for pre-training, making it an obstacle especially for biomedical applications. To overcome the data limitation, federated learning (FL) can be a promising strategy to scale up the dataset for biomedical VLP while protecting data privacy. However, client data are often heterogeneous in real-world scenarios, and we observe that local training on heterogeneous client data would distort the multimodal representation learning and lead to biased cross-modal alignment. To address this challenge, we propose Federated distributional Robust Guidance-Based (FedRGB) learning framework for federated VLP with robustness to data heterogeneity. Specifically, we utilize a guidance-based local training scheme to reduce feature distortions, and employ a distribution-based min-max optimization to learn unbiased cross-modal alignment. The experiments on real-world datasets show our method successfully promotes efficient federated multimodal learning for biomedical VLP with data heterogeneity.
Authors: Afnan Sultan, Jochen Sieg, Miriam Mathea, Andrea Volkamer
Abstract: Molecular Property Prediction (MPP) is vital for drug discovery, crop protection, and environmental science. Over the last decades, diverse computational techniques have been developed, from using simple physical and chemical properties and molecular fingerprints in statistical models and classical machine learning to advanced deep learning approaches. In this review, we aim to distill insights from current research on employing transformer models for MPP. We analyze the currently available models and explore key questions that arise when training and fine-tuning a transformer model for MPP. These questions encompass the choice and scale of the pre-training data, optimal architecture selections, and promising pre-training objectives. Our analysis highlights areas not yet covered in current research, inviting further exploration to enhance the field's understanding. Additionally, we address the challenges in comparing different models, emphasizing the need for standardized data splitting and robust statistical analysis.
Authors: Akhil Padmanabha, Jessie Yuan, Janavi Gupta, Zulekha Karachiwalla, Carmel Majidi, Henny Admoni, Zackory Erickson
Abstract: Physically assistive robots present an opportunity to significantly increase the well-being and independence of individuals with motor impairments or other forms of disability who are unable to complete activities of daily living. Speech interfaces, especially ones that utilize Large Language Models (LLMs), can enable individuals to effectively and naturally communicate high-level commands and nuanced preferences to robots. Frameworks for integrating LLMs as interfaces to robots for high level task planning and code generation have been proposed, but fail to incorporate human-centric considerations which are essential while developing assistive interfaces. In this work, we present a framework for incorporating LLMs as speech interfaces for physically assistive robots, constructed iteratively with 3 stages of testing involving a feeding robot, culminating in an evaluation with 11 older adults at an independent living facility. We use both quantitative and qualitative data from the final study to validate our framework and additionally provide design guidelines for using LLMs as speech interfaces for assistive robots. Videos and supporting files are located on our project website: https://sites.google.com/andrew.cmu.edu/voicepilot/
Authors: Xize Liang, Chao Chen, Jie Wang, Yue Wu, Zhihang Fu, Zhihao Shi, Feng Wu, Jieping Ye
Abstract: The preference alignment aims to enable large language models (LLMs) to generate responses that conform to human values, which is essential for developing general AI systems. Ranking-based methods -- a promising class of alignment approaches -- learn human preferences from datasets containing response pairs by optimizing the log-likelihood margins between preferred and dis-preferred responses. However, due to the inherent differences in annotators' preferences, ranking labels of comparisons for response pairs are unavoidably noisy. This seriously hurts the reliability of existing ranking-based methods. To address this problem, we propose a provably noise-tolerant preference alignment method, namely RObust Preference Optimization (ROPO). To the best of our knowledge, ROPO is the first preference alignment method with noise-tolerance guarantees. The key idea of ROPO is to dynamically assign conservative gradient weights to response pairs with high label uncertainty, based on the log-likelihood margins between the responses. By effectively suppressing the gradients of noisy samples, our weighting strategy ensures that the expected risk has the same gradient direction independent of the presence and proportion of noise. Experiments on three open-ended text generation tasks with four base models ranging in size from 2.8B to 13B demonstrate that ROPO significantly outperforms existing ranking-based methods.
Authors: Giovanni Ciatto, Andrea Agiollo, Matteo Magnini, Andrea Omicini
Abstract: Background. Endowing intelligent systems with semantic data commonly requires designing and instantiating ontologies with domain-specific knowledge. Especially in the early phases, those activities are typically performed manually by human experts possibly leveraging on their own experience. The resulting process is therefore time-consuming, error-prone, and often biased by the personal background of the ontology designer. Objective. To mitigate that issue, we propose a novel domain-independent approach to automatically instantiate ontologies with domain-specific knowledge, by leveraging on large language models (LLMs) as oracles. Method. Starting from (i) an initial schema composed by inter-related classes andproperties and (ii) a set of query templates, our method queries the LLM multi- ple times, and generates instances for both classes and properties from its replies. Thus, the ontology is automatically filled with domain-specific knowledge, compliant to the initial schema. As a result, the ontology is quickly and automatically enriched with manifold instances, which experts may consider to keep, adjust, discard, or complement according to their own needs and expertise. Contribution. We formalise our method in general way and instantiate it over various LLMs, as well as on a concrete case study. We report experiments rooted in the nutritional domain where an ontology of food meals and their ingredients is semi-automatically instantiated from scratch, starting from a categorisation of meals and their relationships. There, we analyse the quality of the generated ontologies and compare ontologies attained by exploiting different LLMs. Finally, we provide a SWOT analysis of the proposed method.
Authors: Vishaal Udandarao, Ameya Prabhu, Adhiraj Ghosh, Yash Sharma, Philip H. S. Torr, Adel Bibi, Samuel Albanie, Matthias Bethge
Abstract: Web-crawled pretraining datasets underlie the impressive "zero-shot" evaluation performance of multimodal models, such as CLIP for classification/retrieval and Stable-Diffusion for image generation. However, it is unclear how meaningful the notion of "zero-shot" generalization is for such multimodal models, as it is not known to what extent their pretraining datasets encompass the downstream concepts targeted for during "zero-shot" evaluation. In this work, we ask: How is the performance of multimodal models on downstream concepts influenced by the frequency of these concepts in their pretraining datasets? We comprehensively investigate this question across 34 models and five standard pretraining datasets (CC-3M, CC-12M, YFCC-15M, LAION-400M, LAION-Aesthetics), generating over 300GB of data artifacts. We consistently find that, far from exhibiting "zero-shot" generalization, multimodal models require exponentially more data to achieve linear improvements in downstream "zero-shot" performance, following a sample inefficient log-linear scaling trend. This trend persists even when controlling for sample-level similarity between pretraining and downstream datasets, and testing on purely synthetic data distributions. Furthermore, upon benchmarking models on long-tailed data sampled based on our analysis, we demonstrate that multimodal models across the board perform poorly. We contribute this long-tail test set as the "Let it Wag!" benchmark to further research in this direction. Taken together, our study reveals an exponential need for training data which implies that the key to "zero-shot" generalization capabilities under large-scale training paradigms remains to be found.
Authors: Jo\~ao Coelho, Bruno Martins, Jo\~ao Magalh\~aes, Jamie Callan, Chenyan Xiong
Abstract: This study investigates the existence of positional biases in Transformer-based models for text representation learning, particularly in the context of web document retrieval. We build on previous research that demonstrated loss of information in the middle of input sequences for causal language models, extending it to the domain of representation learning. We examine positional biases at various stages of training for an encoder-decoder model, including language model pre-training, contrastive pre-training, and contrastive fine-tuning. Experiments with the MS-MARCO document collection reveal that after contrastive pre-training the model already generates embeddings that better capture early contents of the input, with fine-tuning further aggravating this effect.
Authors: Tianze Wang, Maryam Honari-Jahromi, Styliani Katsarou, Olga Mikheeva, Theodoros Panagiotakopoulos, Sahar Asadi, Oleg Smirnov
Abstract: Methods for learning latent user representations from historical behavior logs have gained traction for recommendation tasks in e-commerce, content streaming, and other settings. However, this area still remains relatively underexplored in video and mobile gaming contexts. In this work, we present a novel method for overcoming this limitation by extending a long-range Transformer model from the natural language processing domain to player behavior data. We discuss specifics of behavior tracking in games and propose preprocessing and tokenization approaches by viewing in-game events in an analogous way to words in sentences, thus enabling learning player representations in a self-supervised manner in the absence of ground-truth annotations. We experimentally demonstrate the efficacy of the proposed approach in fitting the distribution of behavior events by evaluating intrinsic language modeling metrics. Furthermore, we qualitatively analyze the emerging structure of the learned embedding space and show its value for generating insights into behavior patterns to inform downstream applications.
Authors: Albert J. Zhai, Yuan Shen, Emily Y. Chen, Gloria X. Wang, Xinlei Wang, Sheng Wang, Kaiyu Guan, Shenlong Wang
Abstract: Can computers perceive the physical properties of objects solely through vision? Research in cognitive science and vision science has shown that humans excel at identifying materials and estimating their physical properties based purely on visual appearance. In this paper, we present a novel approach for dense prediction of the physical properties of objects using a collection of images. Inspired by how humans reason about physics through vision, we leverage large language models to propose candidate materials for each object. We then construct a language-embedded point cloud and estimate the physical properties of each 3D point using a zero-shot kernel regression approach. Our method is accurate, annotation-free, and applicable to any object in the open world. Experiments demonstrate the effectiveness of the proposed approach in various physical property reasoning tasks, such as estimating the mass of common objects, as well as other properties like friction and hardness.
Authors: Michael Saxon, Fatima Jahara, Mahsa Khoshnoodi, Yujie Lu, Aditya Sharma, William Yang Wang
Abstract: With advances in the quality of text-to-image (T2I) models has come interest in benchmarking their prompt faithfulness-the semantic coherence of generated images to the prompts they were conditioned on. A variety of T2I faithfulness metrics have been proposed, leveraging advances in cross-modal embeddings and vision-language models (VLMs). However, these metrics are not rigorously compared and benchmarked, instead presented against few weak baselines by correlation to human Likert scores over a set of easy-to-discriminate images. We introduce T2IScoreScore (TS2), a curated set of semantic error graphs containing a prompt and a set increasingly erroneous images. These allow us to rigorously judge whether a given prompt faithfulness metric can correctly order images with respect to their objective error count and significantly discriminate between different error nodes, using meta-metric scores derived from established statistical tests. Surprisingly, we find that the state-of-the-art VLM-based metrics (e.g., TIFA, DSG, LLMScore, VIEScore) we tested fail to significantly outperform simple feature-based metrics like CLIPScore, particularly on a hard subset of naturally-occurring T2I model errors. TS2 will enable the development of better T2I prompt faithfulness metrics through more rigorous comparison of their conformity to expected orderings and separations under objective criteria.
Authors: Zhengyuan Jiang, Moyang Guo, Yuepeng Hu, Neil Zhenqiang Gong
Abstract: Several companies--such as Google, Microsoft, and OpenAI--have deployed techniques to watermark AI-generated content to enable proactive detection. However, existing literature mainly focuses on user-agnostic detection. Attribution aims to further trace back the user of a generative-AI service who generated a given content detected as AI-generated. Despite its growing importance, attribution is largely unexplored. In this work, we aim to bridge this gap by providing the first systematic study on watermark-based, user-aware detection and attribution of AI-generated content. Specifically, we theoretically study the detection and attribution performance via rigorous probabilistic analysis. Moreover, we develop an efficient algorithm to select watermarks for the users to enhance attribution performance. Both our theoretical and empirical results show that watermark-based detection and attribution inherit the accuracy and (non-)robustness properties of the watermarking method.
Authors: Ning Lu, Shengcai Liu, Zhirui Zhang, Qi Wang, Haifeng Liu, Ke Tang
Abstract: Word-level textual adversarial attacks have demonstrated notable efficacy in misleading Natural Language Processing (NLP) models. Despite their success, the underlying reasons for their effectiveness and the fundamental characteristics of adversarial examples (AEs) remain obscure. This work aims to interpret word-level attacks by examining their $n$-gram frequency patterns. Our comprehensive experiments reveal that in approximately 90\% of cases, word-level attacks lead to the generation of examples where the frequency of $n$-grams decreases, a tendency we term as the $n$-gram Frequency Descend ($n$-FD). This finding suggests a straightforward strategy to enhance model robustness: training models using examples with $n$-FD. To examine the feasibility of this strategy, we employed the $n$-gram frequency information, as an alternative to conventional loss gradients, to generate perturbed examples in adversarial training. The experiment results indicate that the frequency-based approach performs comparably with the gradient-based approach in improving model robustness. Our research offers a novel and more intuitive perspective for understanding word-level textual adversarial attacks and proposes a new direction to improve model robustness.
Authors: Xingwei He, Zhenghao Lin, Yeyun Gong, A-Long Jin, Hang Zhang, Chen Lin, Jian Jiao, Siu Ming Yiu, Nan Duan, Weizhu Chen
Abstract: Many natural language processing (NLP) tasks rely on labeled data to train machine learning models with high performance. However, data annotation is time-consuming and expensive, especially when the task involves a large amount of data or requires specialized domains. Recently, GPT-3.5 series models have demonstrated remarkable few-shot and zero-shot ability across various NLP tasks. In this paper, we first claim that large language models (LLMs), such as GPT-3.5, can serve as an excellent crowdsourced annotator when provided with sufficient guidance and demonstrated examples. Accordingly, we propose AnnoLLM, an annotation system powered by LLMs, which adopts a two-step approach, explain-then-annotate. Concretely, we first prompt LLMs to provide explanations for why the specific ground truth answer/label was assigned for a given example. Then, we construct the few-shot chain-of-thought prompt with the self-generated explanation and employ it to annotate the unlabeled data with LLMs. Our experiment results on three tasks, including user input and keyword relevance assessment, BoolQ, and WiC, demonstrate that AnnoLLM surpasses or performs on par with crowdsourced annotators. Furthermore, we build the first conversation-based information retrieval dataset employing AnnoLLM. This dataset is designed to facilitate the development of retrieval models capable of retrieving pertinent documents for conversational text. Human evaluation has validated the dataset's high quality.
Authors: Md. Arid Hasan, Shudipta Das, Afiyat Anjum, Firoj Alam, Anika Anjum, Avijit Sarker, Sheak Rashed Haider Noori
Abstract: The rapid expansion of the digital world has propelled sentiment analysis into a critical tool across diverse sectors such as marketing, politics, customer service, and healthcare. While there have been significant advancements in sentiment analysis for widely spoken languages, low-resource languages, such as Bangla, remain largely under-researched due to resource constraints. Furthermore, the recent unprecedented performance of Large Language Models (LLMs) in various applications highlights the need to evaluate them in the context of low-resource languages. In this study, we present a sizeable manually annotated dataset encompassing 33,606 Bangla news tweets and Facebook comments. We also investigate zero- and few-shot in-context learning with several language models, including Flan-T5, GPT-4, and Bloomz, offering a comparative analysis against fine-tuned models. Our findings suggest that monolingual transformer-based models consistently outperform other models, even in zero and few-shot scenarios. To foster continued exploration, we intend to make this dataset and our research tools publicly available to the broader research community.
Authors: Gi-Cheon Kang, Junghyun Kim, Jaein Kim, Byoung-Tak Zhang
Abstract: Interactive Object Grasping (IOG) is the task of identifying and grasping the desired object via human-robot natural language interaction. Current IOG systems assume that a human user initially specifies the target object's category (e.g., bottle). Inspired by pragmatics, where humans often convey their intentions by relying on context to achieve goals, we introduce a new IOG task, Pragmatic-IOG, and the corresponding dataset, Intention-oriented Multi-modal Dialogue (IM-Dial). In our proposed task scenario, an intention-oriented utterance (e.g., "I am thirsty") is initially given to the robot. The robot should then identify the target object by interacting with a human user. Based on the task setup, we propose a new robotic system that can interpret the user's intention and pick up the target object, Pragmatic Object Grasping (PROGrasp). PROGrasp performs Pragmatic-IOG by incorporating modules for visual grounding, question asking, object grasping, and most importantly, answer interpretation for pragmatic inference. Experimental results show that PROGrasp is effective in offline (i.e., target object discovery) and online (i.e., IOG with a physical robot arm) settings. Code and data are available at https://github.com/gicheonkang/prograsp.
Authors: Xiangru Tang, Yiming Zong, Jason Phang, Yilun Zhao, Wangchunshu Zhou, Arman Cohan, Mark Gerstein
Abstract: Despite the remarkable capabilities of Large Language Models (LLMs) like GPT-4, producing complex, structured tabular data remains challenging. Our study assesses LLMs' proficiency in structuring tables and introduces a novel fine-tuning method, cognizant of data structures, to bolster their performance. We unveil Struc-Bench, a comprehensive benchmark featuring prominent LLMs (GPT-NeoX-20B, GPT-3.5, GPT-4, and Vicuna), which spans text tables, HTML, and LaTeX formats. Our proposed FormatCoT aids in crafting format-specific instructions from the intended outputs to populate this benchmark. Addressing the gap in task-centered evaluation, we propose two innovative metrics, P-Score (Prompting Score) and H-Score (Heuristical Score), to more accurately gauge LLM performance. Our experiments show that applying our structure-aware fine-tuning to LLaMA-7B leads to substantial performance gains, outshining its LLM counterparts across most measures. In-depth error analysis and creating an ability map across six dimensions -- coverage, formatting, reasoning, comprehension, pragmatics, and hallucination -- highlight areas for future enhancements and suggest forthcoming research trajectories. Our code and models can be found at https://github.com/gersteinlab/Struc-Bench.
Authors: Lukas Berglund, Meg Tong, Max Kaufmann, Mikita Balesni, Asa Cooper Stickland, Tomasz Korbak, Owain Evans
Abstract: We expose a surprising failure of generalization in auto-regressive large language models (LLMs). If a model is trained on a sentence of the form "A is B", it will not automatically generalize to the reverse direction "B is A". This is the Reversal Curse. For instance, if a model is trained on "Valentina Tereshkova was the first woman to travel to space", it will not automatically be able to answer the question, "Who was the first woman to travel to space?". Moreover, the likelihood of the correct answer ("Valentina Tershkova") will not be higher than for a random name. Thus, models do not generalize a prevalent pattern in their training set: if "A is B" occurs, "B is A" is more likely to occur. It is worth noting, however, that if "A is B" appears in-context, models can deduce the reverse relationship. We provide evidence for the Reversal Curse by finetuning GPT-3 and Llama-1 on fictitious statements such as "Uriah Hawthorne is the composer of Abyssal Melodies" and showing that they fail to correctly answer "Who composed Abyssal Melodies?". The Reversal Curse is robust across model sizes and model families and is not alleviated by data augmentation. We also evaluate ChatGPT (GPT-3.5 and GPT-4) on questions about real-world celebrities, such as "Who is Tom Cruise's mother? [A: Mary Lee Pfeiffer]" and the reverse "Who is Mary Lee Pfeiffer's son?". GPT-4 correctly answers questions like the former 79% of the time, compared to 33% for the latter. Code available at: https://github.com/lukasberglund/reversal_curse.
Authors: Mengke Zhang, Tianxing He, Tianle Wang, Lu Mi, Fatemehsadat Mireshghallah, Binyi Chen, Hao Wang, Yulia Tsvetkov
Abstract: In the current user-server interaction paradigm of prompted generation with large language models (LLM) on cloud, the server fully controls the generation process, which leaves zero options for users who want to keep the generated text to themselves. We propose LatticeGen, a cooperative framework in which the server still handles most of the computation while the user controls the sampling operation. The key idea is that the true generated sequence is mixed with noise tokens by the user and hidden in a noised lattice. Considering potential attacks from a hypothetically malicious server and how the user can defend against it, we propose the repeated beam-search attack and the mixing noise scheme. In our experiments we apply LatticeGen to protect both prompt and generation. It is shown that while the noised lattice degrades generation quality, LatticeGen successfully protects the true generation to a remarkable degree under strong attacks (more than 50% of the semantic remains hidden as measured by BERTScore).
Authors: Mingjie Liu, Teodor-Dumitru Ene, Robert Kirby, Chris Cheng, Nathaniel Pinckney, Rongjian Liang, Jonah Alben, Himyanshu Anand, Sanmitra Banerjee, Ismet Bayraktaroglu, Bonita Bhaskaran, Bryan Catanzaro, Arjun Chaudhuri, Sharon Clay, Bill Dally, Laura Dang, Parikshit Deshpande, Siddhanth Dhodhi, Sameer Halepete, Eric Hill, Jiashang Hu, Sumit Jain, Ankit Jindal, Brucek Khailany, George Kokai, Kishor Kunal, Xiaowei Li, Charley Lind, Hao Liu, Stuart Oberman, Sujeet Omar, Ghasem Pasandi, Sreedhar Pratty, Jonathan Raiman, Ambar Sarkar, Zhengjiang Shao, Hanfei Sun, Pratik P Suthar, Varun Tej, Walker Turner, Kaizhe Xu, Haoxing Ren
Abstract: ChipNeMo aims to explore the applications of large language models (LLMs) for industrial chip design. Instead of directly deploying off-the-shelf commercial or open-source LLMs, we instead adopt the following domain adaptation techniques: domain-adaptive tokenization, domain-adaptive continued pretraining, model alignment with domain-specific instructions, and domain-adapted retrieval models. We evaluate these methods on three selected LLM applications for chip design: an engineering assistant chatbot, EDA script generation, and bug summarization and analysis. Our evaluations demonstrate that domain-adaptive pretraining of language models, can lead to superior performance in domain related downstream tasks compared to their base LLaMA2 counterparts, without degradations in generic capabilities. In particular, our largest model, ChipNeMo-70B, outperforms the highly capable GPT-4 on two of our use cases, namely engineering assistant chatbot and EDA scripts generation, while exhibiting competitive performance on bug summarization and analysis. These results underscore the potential of domain-specific customization for enhancing the effectiveness of large language models in specialized applications.
Authors: Gaurav Sahu, Olga Vechtomova, Issam H. Laradji
Abstract: Semi-supervised learning (SSL) is a widely used technique in scenarios where labeled data is scarce and unlabeled data is abundant. While SSL is popular for image and text classification, it is relatively underexplored for the task of extractive text summarization. Standard SSL methods follow a teacher-student paradigm to first train a classification model and then use the classifier's confidence values to select pseudo-labels for the subsequent training cycle; however, such classifiers are not suitable to measure the accuracy of pseudo-labels as they lack specific tuning for evaluation, which leads to confidence values that fail to capture the semantics and correctness of the generated summary. To address this problem, we propose a prompt-based pseudo-labeling strategy with LLMs that picks unlabeled examples with more accurate pseudo-labels than using just the classifier's probability outputs. Our approach also includes a relabeling mechanism that improves the quality of pseudo-labels. We evaluate our method on three text summarization datasets: TweetSumm, WikiHow, and ArXiv/PubMed. We empirically show that a prompting-based LLM that scores and generates pseudo-labels outperforms existing SSL methods on ROUGE-1, ROUGE-2, and ROUGE-L scores on all the datasets. Furthermore, our method achieves competitive G-Eval scores (evaluation with GPT-4) as a fully supervised method that uses 100% of the labeled data with only 16.67% of the labeled data.
Authors: Arkil Patel, Siva Reddy, Dzmitry Bahdanau, Pradeep Dasigi
Abstract: Contemporary Large Language Models (LLMs) exhibit a high degree of code generation and comprehension capability. A particularly promising area is their ability to interpret code modules from unfamiliar libraries for solving user-instructed tasks. Recent work has shown that large proprietary LLMs can learn novel library usage in-context from demonstrations. These results raise several open questions: whether demonstrations of library usage is required, whether smaller (and more open) models also possess such capabilities, etc. In this work, we take a broader approach by systematically evaluating a diverse array of LLMs across three scenarios reflecting varying levels of domain specialization to understand their abilities and limitations in generating code based on libraries defined in-context. Our results show that even smaller open-source LLMs like Llama-2 and StarCoder demonstrate an adept understanding of novel code libraries based on specification presented in-context. Our findings further reveal that LLMs exhibit a surprisingly high proficiency in learning novel library modules even when provided with just natural language descriptions or raw code implementations of the functions, which are often cheaper to obtain than demonstrations. Overall, our results pave the way for harnessing LLMs in more adaptable and dynamic coding environments.
Authors: Bibek Upadhayay, Vahid Behzadan
Abstract: Creating multilingual LLMs poses a significant challenge. Pretraining or fine-tuning LLMs to adopt new languages is evidently very costly. Furthermore, there exist limitations concerning benchmark datasets and the metrics used to measure model performance in multilingual settings. This paper proposes cost-effective solutions to both aforementioned challenges. Firstly, we introduce the Multilingual Instruction-Tuning Dataset (MITS), comprised of Alpaca-52K, Dolly-15K, and Vicuna Benchmark translations into 132 languages. Secondly, we propose a new method called \emph{TaCo: Translation-Assisted Cross-Linguality}, which utilizes translations in a chain-of-thought process to instruction-tune LLMs on new languages through a curriculum-learning process. As a proof of concept, we experimented with the instruction-tuned Guanaco-33B model, performing further instruction tuning using our proposed TaCo method in three low-resource languages and one high-resource language. Our results indicate that the TaCo method impresses GPT-4 with an 82\% score for a low-resource language in the Vicuna Benchmark dataset, doubling the performance in contrast to instruction tuning alone. Furthermore, TaCo shows promise in creating multilingual LLMs, even for low-resource languages. We have released our datasets and model adapters\footnote{https://github.com/UNHSAILLab/TaCo} , encouraging the research community to utilize these resources to advance work on multilingual LLMs.
Authors: Erik Tjong Kim Sang
Abstract: We describe the project REE-HDSC and outline our efforts to improve the quality of named entities extracted automatically from texts generated by hand-written text recognition (HTR) software. We describe a six-step processing pipeline and test it by processing 19th and 20th century death certificates from the civil registry of Curacao. We find that the pipeline extracts dates with high precision but that the precision of person name extraction is low. Next we show how name precision extraction can be improved by retraining HTR models with names, post-processing and by identifying and removing incorrect names.
Authors: Maciej Besta, Florim Memedi, Zhenyu Zhang, Robert Gerstenberger, Guangyuan Piao, Nils Blach, Piotr Nyczyk, Marcin Copik, Grzegorz Kwa\'sniewski, J\"urgen M\"uller, Lukas Gianinazzi, Ales Kubicek, Hubert Niewiadomski, Aidan O'Mahony, Onur Mutlu, Torsten Hoefler
Abstract: The field of natural language processing (NLP) has witnessed significant progress in recent years, with a notable focus on improving large language models' (LLM) performance through innovative prompting techniques. Among these, prompt engineering coupled with structures has emerged as a promising paradigm, with designs such as Chain-of-Thought, Tree of Thoughts, or Graph of Thoughts, in which the overall LLM reasoning is guided by a structure such as a graph. As illustrated with numerous examples, this paradigm significantly enhances the LLM's capability to solve numerous tasks, ranging from logical or mathematical reasoning to planning or creative writing. To facilitate the understanding of this growing field and pave the way for future developments, we devise a general blueprint for effective and efficient LLM reasoning schemes. For this, we conduct an in-depth analysis of the prompt execution pipeline, clarifying and clearly defining different concepts. We then build the first taxonomy of structure-enhanced LLM reasoning schemes. We focus on identifying fundamental classes of harnessed structures, and we analyze the representations of these structures, algorithms executed with these structures, and many others. We refer to these structures as reasoning topologies, because their representation becomes to a degree spatial, as they are contained within the LLM context. Our study compares existing prompting schemes using the proposed taxonomy, discussing how certain design choices lead to different patterns in performance and cost. We also outline theoretical underpinnings, relationships between prompting and other parts of the LLM ecosystem such as knowledge bases, and the associated research challenges. Our work will help to advance future prompt engineering techniques.
Authors: Janice Ahn, Rishu Verma, Renze Lou, Di Liu, Rui Zhang, Wenpeng Yin
Abstract: Mathematical reasoning serves as a cornerstone for assessing the fundamental cognitive capabilities of human intelligence. In recent times, there has been a notable surge in the development of Large Language Models (LLMs) geared towards the automated resolution of mathematical problems. However, the landscape of mathematical problem types is vast and varied, with LLM-oriented techniques undergoing evaluation across diverse datasets and settings. This diversity makes it challenging to discern the true advancements and obstacles within this burgeoning field. This survey endeavors to address four pivotal dimensions: i) a comprehensive exploration of the various mathematical problems and their corresponding datasets that have been investigated; ii) an examination of the spectrum of LLM-oriented techniques that have been proposed for mathematical problem-solving; iii) an overview of factors and concerns affecting LLMs in solving math; and iv) an elucidation of the persisting challenges within this domain. To the best of our knowledge, this survey stands as one of the first extensive examinations of the landscape of LLMs in the realm of mathematics, providing a holistic perspective on the current state, accomplishments, and future challenges in this rapidly evolving field.
Authors: J. E. Eicher, R. F. Irgoli\v{c}
Abstract: Large Language Models (LLMs) like gpt-3.5-turbo and claude-instant-1.2 have become instrumental in interpreting and executing semantic-based tasks. Unfortunately, these models' inherent biases, akin to human cognitive biases, adversely affect their performance. Particularly affected is object selection from lists; a fundamental operation in digital navigation and decision-making. This research critically examines these biases and quantifies the effects on a representative list selection task. To explore these biases, we conducted a series of controlled experiments, manipulating temperature, list length, object identity, object type, prompt complexity, and model. This enabled us to isolate and measure the influence of the biases on selection behavior. Our findings show that bias structure is strongly dependent on the model, with object type modulating the magnitude of the effect. With a strong primacy effect, causing the first objects in a list to be disproportionately represented in outputs. Furthermore the usage of guard rails, a prompt engineering method of ensuring a response structure, can increase bias and decrease instruction adherence when combined with a selection task. The bias is ablated when the guard rail step is separated from the list sampling step, lowering the complexity of each individual task. The implications of this research are two-fold, practically providing a guide for designing unbiased LLM applications and theoretically suggesting that LLMs experience a form of cognitive load compensated for by increasing bias.
Authors: Leonidas Gee, Leonardo Rigutini, Marco Ernandes, Andrea Zugarini
Abstract: Large Language Models have proven highly successful at modelling a variety of tasks. However, this comes at a steep computational cost that hinders wider industrial uptake. In this paper, we present MWT: a Multi-Word Tokenizer that goes beyond word boundaries by representing frequent multi-word expressions as single tokens. MWTs produce a more compact and efficient tokenization that yields two benefits: (1) Increase in performance due to a greater coverage of input data given a fixed sequence length budget; (2) Faster and lighter inference due to the ability to reduce the sequence length with negligible drops in performance. Our results show that MWT is more robust across shorter sequence lengths, thus allowing for major speedups via early sequence truncation.
Authors: Ryandito Diandaru, Lucky Susanto, Zilu Tang, Ayu Purwarianti, Derry Wijaya
Abstract: Large Language Models (LLMs) demonstrate strong machine translation capabilities on languages they are trained on. However, the impact of factors beyond training data size on translation performance remains a topic of debate, especially concerning languages not directly encountered during training. Our study delves into Llama2's translation capabilities. By modeling a linear relationship between linguistic feature distances and machine translation scores, we ask ourselves if there are potentially better central languages for LLMs other than English. Our experiments show that the 7B Llama2 model yields above 10 BLEU when translating into all languages it has seen, which rarely happens for languages it has not seen. Most translation improvements into unseen languages come from scaling up the model size rather than instruction tuning or increasing shot count. Furthermore, our correlation analysis reveals that syntactic similarity is not the only linguistic factor that strongly correlates with machine translation scores. Interestingly, we discovered that under specific circumstances, some languages (e.g. Swedish, Catalan), despite having significantly less training data, exhibit comparable correlation levels to English. These insights challenge the prevailing landscape of LLMs, suggesting that models centered around languages other than English could provide a more efficient foundation for multilingual applications.
Authors: Minjin Kim, Minju Kim, Hana Kim, Beong-woo Kwak, Soyeon Chun, Hyunseo Kim, SeongKu Kang, Youngjae Yu, Jinyoung Yeo, Dongha Lee
Abstract: Conversational recommender system is an emerging area that has garnered an increasing interest in the community, especially with the advancements in large language models (LLMs) that enable diverse reasoning over conversational input. Despite the progress, the field has many aspects left to explore. The currently available public datasets for conversational recommendation lack specific user preferences and explanations for recommendations, hindering high-quality recommendations. To address such challenges, we present a novel conversational recommendation dataset named PEARL, synthesized with persona- and knowledge-augmented LLM simulators. We obtain detailed persona and knowledge from real-world reviews and construct a large-scale dataset with over 57k dialogues. Our experimental results demonstrate that utterances in PEARL include more specific user preferences, show expertise in the target domain, and provide recommendations more relevant to the dialogue context than those in prior datasets.
Authors: Bingli Liao, Danilo Vasconcellos Vargas
Abstract: Large Language Models (LLMs) have shown remarkable capabilities, but their reasoning abilities and underlying mechanisms remain poorly understood. We present a novel approach to enhance LLMs' reasoning through attention mechanism optimization, without additional training data. We identify inefficiencies in the attention distribution caused by non-semantic tokens and propose an algorithm to re-balance the skewed distribution, enabling the model to abstract more nuanced knowledge. Our experiments demonstrate significantly improved reasoning capabilities, particularly for non-STEM questions. We provide insights into the role of attention patterns in LLMs' reasoning and propose a method to enhance these abilities, paving the way for more powerful and versatile language models.
Authors: Sadiya Sayara Chowdhury Puspo, Md Nishat Raihan, Dhiman Goswami, Al Nahian Bin Emran, Amrita Ganguly, Ozlem Uzuner
Abstract: This paper presents the MasonTigers entry to the SemEval-2024 Task 8 - Multigenerator, Multidomain, and Multilingual Black-Box Machine-Generated Text Detection. The task encompasses Binary Human-Written vs. Machine-Generated Text Classification (Track A), Multi-Way Machine-Generated Text Classification (Track B), and Human-Machine Mixed Text Detection (Track C). Our best performing approaches utilize mainly the ensemble of discriminator transformer models along with sentence transformer and statistical machine learning approaches in specific cases. Moreover, zero-shot prompting and fine-tuning of FLAN-T5 are used for Track A and B.
Authors: Yuu Jinnai, Tetsuro Morimura, Kaito Ariu, Kenshi Abe
Abstract: Best-of-N (BoN) sampling with a reward model has been shown to be an effective strategy for aligning Large Language Models (LLMs) to human preferences at the time of decoding. BoN sampling is susceptible to a problem known as reward hacking. Because the reward model is an imperfect proxy for the true objective, over-optimizing its value can compromise its performance on the true objective. A common solution to prevent reward hacking in preference learning techniques is to optimize a reward using proximity regularization (e.g., KL regularization), which ensures that the language model remains close to the reference model. In this research, we propose Regularized Best-of-N (RBoN), a variant of BoN that aims to mitigate reward hacking by incorporating a proximity term in response selection, similar to preference learning techniques. We evaluate two variants of RBoN on the AlpacaFarm dataset and find that they outperform BoN, especially when the proxy reward model has a low correlation with the true objective.
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: Antoine Caubri\`ere, Elodie Gauthier
Abstract: We present the first self-supervised multilingual speech model trained exclusively on African speech. The model learned from nearly 60 000 hours of unlabeled speech segments in 21 languages and dialects spoken in sub-Saharan Africa. On the SSA subset of the FLEURS-102 dataset, our approach based on a HuBERT$_{base}$ (0.09B) architecture shows competitive results, for ASR downstream task, compared to the w2v-bert-51 (0.6B) pre-trained model proposed in the FLEURS benchmark, while being more efficient by using 7x less data and 6x less parameters. Furthermore, in the context of a LID downstream task, our approach outperforms FLEURS baselines accuracy by over 22\%.
Authors: Wanyong Feng, Jaewook Lee, Hunter McNichols, Alexander Scarlatos, Digory Smith, Simon Woodhead, Nancy Otero Ornelas, Andrew Lan
Abstract: Multiple-choice questions (MCQs) are ubiquitous in almost all levels of education since they are easy to administer, grade, and are a reliable format in assessments and practices. One of the most important aspects of MCQs is the distractors, i.e., incorrect options that are designed to target common errors or misconceptions among real students. To date, the task of crafting high-quality distractors largely remains a labor and time-intensive process for teachers and learning content designers, which has limited scalability. In this work, we study the task of automated distractor generation in the domain of math MCQs and explore a wide variety of large language model (LLM)-based approaches, from in-context learning to fine-tuning. We conduct extensive experiments using a real-world math MCQ dataset and find that although LLMs can generate some mathematically valid distractors, they are less adept at anticipating common errors or misconceptions among real students.
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: Azmine Toushik Wasi, Taki Hasan Rafi, Raima Islam, Dong-Kyu Chae
Abstract: Knowledge Graphs (KGs) have proven essential in information processing and reasoning applications because they link related entities and give context-rich information, supporting efficient information retrieval and knowledge discovery; presenting information flow in a very effective manner. Despite being widely used globally, Bangla is relatively underrepresented in KGs due to a lack of comprehensive datasets, encoders, NER (named entity recognition) models, POS (part-of-speech) taggers, and lemmatizers, hindering efficient information processing and reasoning applications in the language. Addressing the KG scarcity in Bengali, we propose BanglaAutoKG, a pioneering framework that is able to automatically construct Bengali KGs from any Bangla text. We utilize multilingual LLMs to understand various languages and correlate entities and relations universally. By employing a translation dictionary to identify English equivalents and extracting word features from pre-trained BERT models, we construct the foundational KG. To reduce noise and align word embeddings with our goal, we employ graph-based polynomial filters. Lastly, we implement a GNN-based semantic filter, which elevates contextual understanding and trims unnecessary edges, culminating in the formation of the definitive KG. Empirical findings and case studies demonstrate the universal effectiveness of our model, capable of autonomously constructing semantically enriched KGs from any text.
Authors: Regina Stodden
Abstract: In this work, we propose EASSE-multi, a framework for easier automatic sentence evaluation for languages other than English. Compared to the original EASSE framework, EASSE-multi does not focus only on English. It contains tokenizers and versions of text simplification evaluation metrics which are suitable for multiple languages. In this paper, we exemplify the usage of EASSE-multi for German TS, resulting in EASSE-DE. Further, we compare text simplification results when evaluating with different language or tokenization settings of the metrics. Based on this, we formulate recommendations on how to make the evaluation of (German) TS models more transparent and better comparable. The code of EASSE-multi and its German specialisation (EASSE-DE) can be found at https://github.com/rstodden/easse-de.
Authors: Haotong Yang, Fanxu Meng, Zhouchen Lin, Muhan Zhang
Abstract: The pre-trained large language models (LLMs) have shown their extraordinary capacity to solve reasoning tasks, even on tasks that require a complex process involving multiple sub-steps. However, given the vast possible generation space of all the tasks, how the pretrained model learns the reasoning ability remains an open question. We firstly propose that an intrinsic structural constraint on the generated sequence of language-based reasoning -- we called it template-content structure (T-C structure) -- is the key to explain why LLMs can solve a large number of complex reasoning problems with limited training data by showing this structure can reduce the possible space from exponential level to linear level. Furthermore, by generalizing this structure to the hierarchical case, we demonstrate that models can achieve task composition, further reducing the space needed to learn from linear to logarithmic, thereby effectively learning on complex reasoning involving multiple steps. We provide both examples and formal theory of our T-C structure. We also experimentally validate the existence of the T-C structure in some current LLMs and its effectiveness for reasoning.
Authors: Qiying Yu, Quan Sun, Xiaosong Zhang, Yufeng Cui, Fan Zhang, Yue Cao, Xinlong Wang, Jingjing Liu
Abstract: Large multimodal models demonstrate remarkable generalist ability to perform diverse multimodal tasks in a zero-shot manner. Large-scale web-based image-text pairs contribute fundamentally to this success, but suffer from excessive noise. Recent studies use alternative captions synthesized by captioning models and have achieved notable benchmark performance. However, our experiments reveal significant Scalability Deficiency and World Knowledge Loss issues in models trained with synthetic captions, which have been largely obscured by their initial benchmark success. Upon closer examination, we identify the root cause as the overly-simplified language structure and lack of knowledge details in existing synthetic captions. To provide higher-quality and more scalable multimodal pretraining data, we propose CapsFusion, an advanced framework that leverages large language models to consolidate and refine information from both web-based image-text pairs and synthetic captions. Extensive experiments show that CapsFusion captions exhibit remarkable all-round superiority over existing captions in terms of model performance (e.g., 18.8 and 18.3 improvements in CIDEr score on COCO and NoCaps), sample efficiency (requiring 11-16 times less computation than baselines), world knowledge depth, and scalability. These effectiveness, efficiency and scalability advantages position CapsFusion as a promising candidate for future scaling of LMM training.
Authors: Yushi Hu, Otilia Stretcu, Chun-Ta Lu, Krishnamurthy Viswanathan, Kenji Hata, Enming Luo, Ranjay Krishna, Ariel Fuxman
Abstract: Solving complex visual tasks such as "Who invented the musical instrument on the right?" involves a composition of skills: understanding space, recognizing instruments, and also retrieving prior knowledge. Recent work shows promise by decomposing such tasks using a large language model (LLM) into an executable program that invokes specialized vision models. However, generated programs are error-prone: they omit necessary steps, include spurious ones, and are unable to recover when the specialized models give incorrect outputs. Moreover, they require loading multiple models, incurring high latency and computation costs. We propose Visual Program Distillation (VPD), an instruction tuning framework that produces a vision-language model (VLM) capable of solving complex visual tasks with a single forward pass. VPD distills the reasoning ability of LLMs by using them to sample multiple candidate programs, which are then executed and verified to identify a correct one. It translates each correct program into a language description of the reasoning steps, which are then distilled into a VLM. Extensive experiments show that VPD improves the VLM's ability to count, understand spatial relations, and reason compositionally. Our VPD-trained PaLI-X outperforms all prior VLMs, achieving state-of-the-art performance across complex vision tasks, including MMBench, OK-VQA, A-OKVQA, TallyQA, POPE, and Hateful Memes. An evaluation with human annotators also confirms that VPD improves model response factuality and consistency. Finally, experiments on content moderation demonstrate that VPD is also helpful for adaptation to real-world applications with limited data.
Authors: Vihari Piratla, Juyeon Heo, Katherine M. Collins, Sukriti Singh, Adrian Weller
Abstract: Model explanations can be valuable for interpreting and debugging predictive models. We study a specific kind called Concept Explanations, where the goal is to interpret a model using human-understandable concepts. Although popular for their easy interpretation, concept explanations are known to be noisy. We begin our work by identifying various sources of uncertainty in the estimation pipeline that lead to such noise. We then propose an uncertainty-aware Bayesian estimation method to address these issues, which readily improved the quality of explanations. We demonstrate with theoretical analysis and empirical evaluation that explanations computed by our method are robust to train-time choices while also being label-efficient. Further, our method proved capable of recovering relevant concepts amongst a bank of thousands, in an evaluation with real-datasets and off-the-shelf models, demonstrating its scalability. We believe the improved quality of uncertainty-aware concept explanations make them a strong candidate for more reliable model interpretation. We release our code at https://github.com/vps-anonconfs/uace.
Authors: Zhengqing Yuan, Zhaoxu Li, Weiran Huang, Yanfang Ye, Lichao Sun
Abstract: In recent years, multimodal large language models (MLLMs) such as GPT-4V have demonstrated remarkable advancements, excelling in a variety of vision-language tasks. Despite their prowess, the closed-source nature and computational demands of such models limit their accessibility and applicability. This study introduces TinyGPT-V, a novel open-source MLLM, designed for efficient training and inference across various vision-language tasks, including image captioning (IC) and visual question answering (VQA). Leveraging a compact yet powerful architecture, TinyGPT-V integrates the Phi-2 language model with pre-trained vision encoders, utilizing a unique mapping module for visual and linguistic information fusion. With a training regimen optimized for small backbones and employing a diverse dataset amalgam, TinyGPT-V requires significantly lower computational resources 24GB for training and as little as 8GB for inference without compromising on performance. Our experiments demonstrate that TinyGPT-V, with its language model 2.8 billion parameters, achieves comparable results in VQA and image inference tasks to its larger counterparts while being uniquely suited for deployment on resource-constrained devices through innovative quantization techniques. This work not only paves the way for more accessible and efficient MLLMs but also underscores the potential of smaller, optimized models in bridging the gap between high performance and computational efficiency in real-world applications. Additionally, this paper introduces a new approach to multimodal large language models using smaller backbones. Our code and training weights are available in \url{https://github.com/DLYuanGod/TinyGPT-V}.
Authors: Sijia Liu, Yuanshun Yao, Jinghan Jia, Stephen Casper, Nathalie Baracaldo, Peter Hase, Xiaojun Xu, Yuguang Yao, Hang Li, Kush R. Varshney, Mohit Bansal, Sanmi Koyejo, Yang Liu
Abstract: We explore machine unlearning (MU) in the domain of large language models (LLMs), referred to as LLM unlearning. This initiative aims to eliminate undesirable data influence (e.g., sensitive or illegal information) and the associated model capabilities, while maintaining the integrity of essential knowledge generation and not affecting causally unrelated information. We envision LLM unlearning becoming a pivotal element in the life-cycle management of LLMs, potentially standing as an essential foundation for developing generative AI that is not only safe, secure, and trustworthy, but also resource-efficient without the need of full retraining. We navigate the unlearning landscape in LLMs from conceptual formulation, methodologies, metrics, and applications. In particular, we highlight the often-overlooked aspects of existing LLM unlearning research, e.g., unlearning scope, data-model interaction, and multifaceted efficacy assessment. We also draw connections between LLM unlearning and related areas such as model editing, influence functions, model explanation, adversarial training, and reinforcement learning. Furthermore, we outline an effective assessment framework for LLM unlearning and explore its applications in copyright and privacy safeguards and sociotechnical harm reduction.
Authors: Hui Su, Zhi Tian, Xiaoyu Shen, Xunliang Cai
Abstract: Scaling law principles indicate a power-law correlation between loss and variables such as model size, dataset size, and computational resources utilized during training. These principles play a vital role in optimizing various aspects of model pre-training, ultimately contributing to the success of large language models such as GPT-4, Llama and Gemini. However, the original scaling law paper by OpenAI did not disclose the complete details necessary to derive the precise scaling law formulas, and their conclusions are only based on models containing up to 1.5 billion parameters. Though some subsequent works attempt to unveil these details and scale to larger models, they often neglect the training dependency of important factors such as the learning rate, context length and batch size, leading to their failure to establish a reliable formula for predicting the test loss trajectory. In this technical report, we confirm that the scaling law formulations proposed in the original OpenAI paper remain valid when scaling the model size up to 33 billion, but the constant coefficients in these formulas vary significantly with the experiment setup. We meticulously identify influential factors and provide transparent, step-by-step instructions to estimate all constant terms in scaling-law formulas by training on models with only 1M~60M parameters. Using these estimated formulas, we showcase the capability to accurately predict various attributes for models with up to 33B parameters before their training, including (1) the minimum possible test loss; (2) the minimum required training steps and processed tokens to achieve a specific loss; (3) the critical batch size with an optimal time/computation trade-off at any loss value; and (4) the complete test loss trajectory with arbitrary batch size.
Authors: Harry Dong, Beidi Chen, Yuejie Chi
Abstract: With the development of transformer-based large language models (LLMs), they have been applied to many fields due to their remarkable utility, but this comes at a considerable computational cost at deployment. Fortunately, some methods such as pruning or constructing a mixture of experts (MoE) aim at exploiting sparsity in transformer feedforward (FF) blocks to gain boosts in speed and reduction in memory requirements. However, these techniques can be very costly and inflexible in practice, as they often require training or are restricted to specific types of architectures. To address this, we introduce GRIFFIN, a novel training-free MoE that selects unique FF experts at the sequence level for efficient generation across a plethora of LLMs with different non-ReLU activation functions. This is possible due to a critical observation that many trained LLMs naturally produce highly structured FF activation patterns within a sequence, which we call flocking. Despite our method's simplicity, we show with 50% of the FF parameters, GRIFFIN maintains the original model's performance with little to no degradation on a variety of classification and generation tasks, all while improving latency (e.g. 1.25$\times$ speed-up in Llama 2 13B on an NVIDIA L40). Code is available at https://github.com/hdong920/GRIFFIN.
Authors: Ziyao Zeng, Daniel Wang, Fengyu Yang, Hyoungseob Park, Yangchao Wu, Stefano Soatto, Byung-Woo Hong, Dong Lao, Alex Wong
Abstract: Three-dimensional (3D) reconstruction from a single image is an ill-posed problem with inherent ambiguities, i.e. scale. Predicting a 3D scene from text description(s) is similarly ill-posed, i.e. spatial arrangements of objects described. We investigate the question of whether two inherently ambiguous modalities can be used in conjunction to produce metric-scaled reconstructions. To test this, we focus on monocular depth estimation, the problem of predicting a dense depth map from a single image, but with an additional text caption describing the scene. To this end, we begin by encoding the text caption as a mean and standard deviation; using a variational framework, we learn the distribution of the plausible metric reconstructions of 3D scenes corresponding to the text captions as a prior. To "select" a specific reconstruction or depth map, we encode the given image through a conditional sampler that samples from the latent space of the variational text encoder, which is then decoded to the output depth map. Our approach is trained alternatingly between the text and image branches: in one optimization step, we predict the mean and standard deviation from the text description and sample from a standard Gaussian, and in the other, we sample using a (image) conditional sampler. Once trained, we directly predict depth from the encoded text using the conditional sampler. We demonstrate our approach on indoor (NYUv2) and outdoor (KITTI) scenarios, where we show that language can consistently improve performance in both.