Today’s most accurate language models are trained on orders of magnitude more language data than human language learners receive— but with no supervision from other sensory modalities that play a crucial role in human learning. Can we make LMs’ representations and predictions more accurate (and more human-like) with more ecologically plausible supervision? This paper describes LexiContrastive Grounding (LCG), a grounded language learning procedure that leverages visual supervision to improve textual representations. LexiContrastive Grounding combines a next-token prediction strategy with a contrastive visual grounding objective, focusing on early-layerrepresentations that encode lexical information. Across multiple word-learning and sentence-understanding benchmarks, LexiContrastiveGrounding not only outperforms standard language-only models in terms of learning efficiency in small and developmentally plausible data regimes, but also improves upon vision-and-language learning procedures including CLIP, GIT, Flamingo, and Vokenization.Moreover, LexiContrastive Grounding improves perplexity by around 5% on multiple language modeling tasks compared to other models trained on the same amount of text data. This work underscores the potential of incorporating visual grounding into language models, aligning more closely with the multimodal nature of human language acquisition.
论文检索
输入标题、作者或关键词，从 1,915 篇学术成果中精准定位
Large language models (LLMs) are typically trained on general source data forvarious domains, but a recent surge in domain-specific LLMs has shown theirpotential to outperform general-purpose models in domain-specific tasks (e.g.,biomedicine). Although domain-specific pre-training enhances efficiency andleads to smaller models, the computational costs of training these LLMs remainhigh, posing budgeting challenges. We introduce MediSwift, a suite of biomedicalLMs that leverage sparse pre-training on domain-specific biomedical text data.By inducing up to 75% weight sparsity during the pre-training phase, MediSwiftachieves a 2-2.5x reduction in training FLOPs. Notably, all sparse pre-trainingwas performed on the Cerebras CS-2 system, which is specifically designed torealize the acceleration benefits from unstructured weight sparsity, therebysignificantly enhancing the efficiency of the MediSwift models. Throughsubsequent dense fine-tuning and strategic soft prompting, MediSwift modelsoutperform existing LLMs up to 7B parameters on biomedical tasks, setting newbenchmarks w.r.t efficiency-accuracy on tasks such as PubMedQA. Our results showthat sparse pre-training, along with dense fine-tuning and soft prompting,offers an effective method for creating high-performing, computationallyefficient models in specialized domains.
N-gram language models (LMs) are the innovation that first made large-vocabulary continuous automatic speech recognition (ASR) viable. With neural end-to-end ASR architectures, however, LMs have become an afterthought. While the effect on accuracy may be negligible for English and Mandarin, jettisoning the LM might not make sense for the world’s remaining 6000+ languages. In this paper, we investigate the role of the LM in low-resource ASR. First we ask: does using an n-gram LM in decoding in neural architectures help ASR performance? While it may seem obvious that it should, its absence in most implementations suggests otherwise. Second, we ask: when an n-gram LM is used in ASR, is there a relationship between the size of the LM and ASR accuracy? We have discovered that gut feelings on this question vary considerably, but there is little empirical work to support any particular claim. We explore these questions “in the wild” using a deliberately diverse set of 9 very small ASR corpora. The results show that: (1) decoding with an n-gram LM, regardless of its size, leads to lower word error rates; and (2) increasing the size of the LM appears to yield improvements only when the audio corpus itself is already relatively large. This suggests that collecting additional LM training text may benefit widely-spoken languages which typically have larger audio corpora. In contrast, for endangered languages where data of any kind will always be limited, efforts may be better spent collecting additional transcribed audio.
The task of information extraction (IE) is to extract structured knowledge from text. However, it is often not straightforward to utilize IE output due to the mismatch between the IE ontology and the downstream application needs. We propose a new formulation of IE, Text2DB, that emphasizes the integration of IE output and the target database (or knowledge base). Given a user instruction, a document set, and a database, our task requires the model to update the database with values from the document set to satisfy the user instruction. This task requires understanding user instructions for what to extract and adapting to the given DB/KB schema for how to extract on the fly. To evaluate this new task, we introduce a new benchmark featuring common demands such as data infilling, row population, and column addition. In addition, we propose an LLM agent framework OPAL (Observe-Plan-Analyze LLM) which includes an Observer component that interacts with the database, the Planner component that generates a code-based plan with calls to IE models, and the Analyzer component that provides feedback regarding code quality before execution. Experiments show that OPAL can successfully adapt to diverse database schemas by generating different code plans and calling the required IE models. We also highlight difficult cases such as dealing with large databases with complex dependencies and extraction hallucination, which we believe deserve further investigation.
Large language models (LLMs), while exhibiting exceptional performance, suffer from hallucinations, especially on knowledge-intensive tasks. Existing works propose to augment LLMs with individual text units retrieved from external knowledge corpora to alleviate the issue. However, in many domains, texts are interconnected (e.g., academic papers in a bibliographic graph are linked by citations and co-authorships) which form a (text-attributed) graph. The knowledge in such graphs is encoded not only in single texts/nodes but also in their associated connections. To facilitate the research of augmenting LLMs with graphs, we manually construct a Graph Reasoning Benchmark dataset called GRBench, containing 1,740 questions that can be answered with the knowledge from 10 domain graphs. Then, we propose a simple and effective framework called Graph Chain-of-thought (Graph-CoT) to augment LLMs with graphs by encouraging LLMs to reason on the graph iteratively. Each Graph-CoT iteration consists of three sub-steps: LLM reasoning, LLM-graph interaction, and graph execution. We conduct systematic experiments with three LLM backbones on GRBench, where Graph-CoT outperforms the baselines consistently. The code is available at https://github.com/PeterGriffinJin/Graph-CoT/.
Despite LLMs’ recent advancements, they still suffer from factual inconsistency and hallucination. An often-opted remedy is retrieval-augmented generation – however, there is no guarantee that the model will strictly adhere to retrieved grounding. Fundamentally, LLMs need to be aligned to be more faithful to grounding, which will require high-quality preference annotations. This paper investigates whether we can create high-quality grounded preference data for model alignment without using annotations from humans or large proprietary models. We experimented with existing entailment data and proposed approaches to generate synthetic grounded preference data, with which we train a Grounded Preference Model(GPM). We demonstrate through Proximal Policy Optimization(PPO) training of Mistral-7B-Instruct that our GPM model can successfully align powerful LLMs to generate much better grounded responses as judged by GPT4. Moreover, we show that our GPM is also a great faithfulness classifier, achieving SoTA in dialogue sub-tasks of the TRUE faithfulness Benchmark. We will release our GPM under the Apache 2.0 license.
Which Side Are You On? A Multi-task Dataset for End-to-End Argument Summarisation and Evaluation
PDF ↗With the recent advances of large language models (LLMs), it is no longer infeasible to build an automated debate system that helps people to synthesise persuasive arguments. Previous work attempted this task by integrating multiple components. In our work, we introduce an argument mining dataset that captures the end-to-end process of preparing an argumentative essay for a debate, which covers the tasks of claim and evidence identification (Task 1 ED), evidence convincingness ranking (Task 2 ECR), argumentative essay summarisation and human preference ranking (Task 3 ASR) and metric learning for automated evaluation of resulting essays, based on human feedback along argument quality dimensions (Task 4 SQE). Our dataset contains 14k examples of claims that are fully annotated with various properties supporting the aforementioned tasks. We evaluate multiple generative baselines for each of these tasks, including representative LLMs. We find, that while they show promising results on individual tasks in our benchmark, their end-to-end performance on all four tasks in succession deteriorates significantly, both in automated measures as well as in human-centred evaluation. This challenge presented by our proposed dataset motivates future research on end-to-end argument mining and summarisation. The repository of this project is available at https://github.com/HarrywillDr/ArgSum-Datatset.
CHIME: LLM-Assisted Hierarchical Organization of Scientific Studies for Literature Review Support
PDF ↗Literature review requires researchers to synthesize a large amount of information and is increasingly challenging as the scientific literature expands. In this work, we investigate the potential of LLMs for producing hierarchical organizations of scientific studies to assist researchers with literature review. We define hierarchical organizations as tree structures where nodes refer to topical categories and every node is linked to the studies assigned to that category. Our naive LLM-based pipeline for hierarchy generation from a set of studies produces promising yet imperfect hierarchies, motivating us to collect CHIME, an expert-curated dataset for this task focused on biomedicine. Given the challenging and time-consuming nature of building hierarchies from scratch, we use a human-in-the-loop process in which experts correct errors (both links between categories and study assignment) in LLM-generated hierarchies. CHIME contains 2,174 LLM-generated hierarchies covering 472 topics, and expert-corrected hierarchies for a subset of 100 topics. Expert corrections allow us to quantify LLM performance, and we find that while they are quite good at generating and organizing categories, their assignment of studies to categories could be improved. We attempt to train a corrector model with human feedback which improves study assignment by 12.6 F1 points. We release our dataset and models to encourage research on developing better assistive tools for literature review.
The Counterfeit Conundrum: Can Code Language Models Grasp the Nuances of Their Incorrect Generations?
PDF ↗While language models are increasingly more proficient at code generation, they still frequently generate incorrect programs. Many of these programs are obviously wrong, but others are more subtle and pass weaker correctness checks such as being able to compile. In this work, we focus on these counterfeit samples: programs sampled from a language model that 1) have a high enough log-probability to be generated at a moderate temperature and 2) pass weak correctness checks. Overall, we discover that most models have a very shallow understanding of counterfeits through three clear failure modes. First, models mistakenly classify them as correct. Second, models are worse at reasoning about the execution behaviour of counterfeits and often predict their execution results as if they were correct. Third, when asking models to fix counterfeits, the likelihood of a model successfully repairing a counterfeit is often even lower than that of sampling a correct program from scratch. Counterfeits also have very unexpected properties: first, counterfeit programs for problems that are easier for a model to solve are not necessarily easier to detect and only slightly easier to execute and repair. Second, counterfeits from a given model are just as confusing to the model itself as they are to other models. Finally, both strong and weak models are able to generate counterfeit samples that equally challenge all models. In light of our findings, we recommend that care and caution be taken when relying on models to understand their own samples, especially when no external feedback is incorporated.
Many recent studies endeavor to improve open-sourced language models through imitation learning, re-training on the synthetic instruction data from state-of-the-art proprietary models like ChatGPT and GPT-4.However, the innate nature of synthetic data inherently contains noisy data, giving rise to a substantial presence of low-quality data replete with misleading queries, erroneous responses, and flawed reasoning.Although we intuitively grasp the potential harm of noisy data, we lack a quantitative understanding of its impact.To this end, this paper explores correlation between the degree of noise and its impact on language models through instruction tuning.We first introduce the Falsity-Controllable () dataset, which comprises pairs of true answers and corresponding reasoning, as well as false pairs to manually control the factuality ratio of the dataset.Through our extensive experiments, we found multiple intriguing findings of the correlation between factuality and instruction tuning. Specifically, factuality can significantly impact various benchmark characteristics especially when benchmarks are related to knowledge domain, and initial data quality plays a critical role, whereas the number of learning steps has a lesser impact.Additionally, we noted that once the language model is trained with a dataset contaminated by noise, restoring its original performance becomes exceptionally challenging, verging on irreversible.
Intelligent task-oriented dialogue systems (ToDs) are expected to continuously acquire new knowledge, also known as Continual Learning (CL), which is crucial to fit ever-changing user needs. However, catastrophic forgetting dramatically degrades the model performance in face of a long streamed curriculum. In this paper, we aim to overcome the forgetting problem in ToDs and propose a method (HESIT) with hyper-gradient-based exemplar strategy, which samples influential exemplars for periodic retraining. Instead of unilaterally observing data or models, HESIT adopts a profound exemplar selection strategy that considers the general performance of the trained model when selecting exemplars for each task domain. Specifically, HESIT analyzes the training data influence by tracing their hyper-gradient in the optimization process. Furthermore, HESIT avoids estimating Hessian to make it compatible for ToDs with a large pre-trained model. Experimental results show that HESIT effectively alleviates catastrophic forgetting by exemplar selection, and achieves state-of-the-art performance on the largest CL benchmark of ToDs in terms of all metrics.
Emotion detection is the task of automatically associating one or more emotions with a text. The emotions are experienced, targeted, and caused by different semantic constituents. Therefore, it is necessary to incorporate these semantic constituents into the process of emotion detection. In this study, we propose a new task called emotion semantic parsing which aims to parse the emotion and semantic constituents into an abstract semantic tree structure. In particular, we design an end-to-end generation model to capture the relations between emotion and all the semantic constituents, and to generate them jointly. Furthermore, we employ a task decomposition strategy to capture the semantic relation among these constituents in a more cognitive and structural way. Experimental results demonstrate the importance of the proposed task, and indicate the proposed model gives superior performance compared to other models.
Large language models (LLMs) show great performance in various tasks, but face deployment challenges from limited memory capacity and bandwidth.Low-bit weight quantization can save memory and accelerate inference.Although floating-point (FP) formats show good performance in LLM quantization, they tend to perform poorly with small group sizes or sub-4 bits.We find the reason is that the absence of asymmetry in previous FP quantization makes it unsuitable for handling asymmetric value distribution of LLM weight tensors.In this work, we propose asymmetric FP quantization (AFPQ), which sets separate scales for positive and negative values.Our method leads to large accuracy improvements and can be easily plugged into other quantization methods, including GPTQ and AWQ, for better performance.Besides, no additional storage is needed compared with asymmetric integer (INT) quantization.The code is available at https://github.com/zhangsichengsjtu/AFPQ.
Keyphrase extraction aims to automatically extract salient phrases representing the critical information in the source document. Identifying salient phrases is challenging because there is a lot of noisy information in the document, leading to wrong extraction. To address this issue, in this paper, we propose a hybrid matching model for keyphrase extraction, which combines representation-focused and interaction-based matching modules into a unified framework for improving the performance of the keyphrase extraction task. Specifically, HybridMatch comprises (1) a PLM-based Siamese encoder component that represents both candidate phrases and documents, (2) an interaction-focused matching (IM) component that estimates word matches between candidate phrases and the corresponding document at the word level, and (3) a representation-focused matching (RM) component captures context-aware semantic relatedness of each candidate keyphrase at the phrase level. Extensive experimental results on the OpenKP dataset demonstrate that the performance of the proposed model HybridMatch outperforms the recent state-of-the-art keyphrase extraction baselines. Furthermore, we discuss the performance of large language models in keyphrase extraction based on recent studies and our experiments.
Controllable Data Augmentation for Few-Shot Text Mining with Chain-of-Thought Attribute Manipulation
PDF ↗Prompting large language models (LLMs) for data augmentation has recently become a common practice in few-shot NLP tasks. In this paper, we propose Chain-of-Thought Attribute Manipulation (CoTAM), a novel approach that generates new data from existing examples by only tweaking in the user-provided, task-specific attribute, e.g., sentiment polarity or topic in movie reviews. Instead of conventional latent representation controlling, we leverage the chain-of-thought prompting to directly edit the text in three steps, (1) attribute decomposition, (2) manipulation proposal, and (3) sentence reconstruction. Extensive results on various tasks, such as text (pair) classification and aspect-based sentiment analysis, verify the superiority of CoTAM over other LLM-based augmentation methods with the same number of training examples for both fine-tuning and in-context learning. Remarkably, the 2D visualization of the augmented dataset using principle component analysis revealed a human-recognizable decision boundary that is likely hinted by the attribute manipulation, demonstrating the potential of our proposed approach.
Image-based advertisements are complex multimodal stimuli that often contain unusual visual elements and figurative language. Previous research on automatic ad understanding has reported impressive zero-shot accuracy of contrastive vision-and-language models (VLMs) on an ad-explanation retrieval task. Here, we examine the original task setup and show that contrastive VLMs can solve it by exploiting grounding heuristics. To control for this confound, we introduce TRADE, a new evaluation test set with adversarial grounded explanations. While these explanations look implausible to humans, we show that they “fool” four different contrastive VLMs. Our findings highlight the need for an improved operationalisation of automatic ad understanding that truly evaluates VLMs’ multimodal reasoning abilities. We make our code and TRADE available at https://github.com/dmg-illc/trade.
Humor is a fundamental facet of human cognition and interaction. Yet, despite recent advances in natural language processing, humor detection remains a challenging task that is complicated by the scarcity of datasets that pair humorous texts with similar non-humorous counterparts. We investigate whether large language models (LLMs) can generate synthetic data for humor detection via editing texts. We benchmark LLMs on an existing human dataset and show that current LLMs display an impressive ability to “unfun” jokes, as judged by humans and as measured on the downstream task of humor detection. We extend our approach to a code-mixed English-Hindi humor dataset where we find that GPT-4’s synthetic data is highly rated by bilingual annotators and provides challenging adversarial examples for humor classifiers.
While generative multilingual models are rapidly being deployed, their safety and fairness evaluations are largely limited to resources collected in English. This is especially problematic for evaluations targeting inherently socio-cultural phenomena such as stereotyping, where it is important to build multilingual resources that reflect the stereotypes prevalent in respective language communities. However, gathering these resources, at scale, in varied languages and regions pose a significant challenge as it requires broad socio-cultural knowledge and can also be prohibitively expensive. To overcome this critical gap, we employ a recently introduced approach that couples LLM generations for scale with culturally situated validations for reliability, and build SeeGULL Multilingual, a global-scale multilingual dataset of social stereotypes, containing over 25K stereotypes, spanning 23 pairs of languages and regions they are common in, with human annotations, and demonstrate its utility in identifying gaps in model evaluations.
What Do Dialect Speakers Want? A Survey of Attitudes Towards Language Technology for German Dialects
PDF ↗Natural language processing (NLP) has largely focused on modelling standardized languages. More recently, attention has increasingly shifted to local, non-standardized languages and dialects. However, the relevant speaker populations’ needs and wishes with respect to NLP tools are largely unknown. In this paper, we focus on dialects and regional languages related to German – a group of varieties that is heterogeneous in terms of prestige and standardization. We survey speakers of these varieties (N=327) and present their opinions on hypothetical language technologies for their dialects. Although attitudes vary among subgroups of our respondents, we find that respondents are especially in favour of potential NLP tools that work with dialectal input (especially audio input) such as virtual assistants, and less so for applications that produce dialectal output such as machine translation or spellcheckers.
While subword tokenizers such as BPE and WordPiece are typically used to build vocabularies for NLP models, the method of decoding text into a sequence of tokens from these vocabularies is often left unspecified, or ill-suited to the method in which they were constructed. We provide a controlled analysis of seven tokenizer inference methods across four different algorithms and three vocabulary sizes, performed on a novel intrinsic evaluation suite we curated for English, combining measures rooted in morphology, cognition, and information theory. We show that for the most commonly used tokenizers, greedy inference performs surprisingly well; and that SaGe, a recently-introduced contextually-informed tokenizer, outperforms all others on morphological alignment.