EMNLP 2024 论文列表

The proliferation of inflammatory or misleading “fake” news content has become increasingly common in recent years. Simultaneously, it has become easier than ever to use AI tools to generate photorealistic images depicting any scene imaginable. Combining these two—AI-generated fake news content—is particularly potent and dangerous. To combat the spread of AI-generated fake news, we propose the MiRAGeNews Dataset, a dataset of 12,500 high-quality real and AI-generated image-caption pairs from state-of-the-art generators. We find that our dataset poses a significant challenge to humans (60% F-1) and state-of-the-art multi-modal LLMs (< 24% F-1). Using our dataset we train a multi-modal detector (MiRAGe) that improves by +5.1% F-1 over state-of-the-art baselines on image-caption pairs from out-of-domain image generators and news publishers. We release our code and data to aid future work on detecting AI-generated content.

Low-rank adaptation (LoRA) and its mixture-of-experts (MOE) variants are highly effective parameter-efficient fine-tuning (PEFT) methods. However, they introduce significant latency in multi-tenant settings due to the LoRA modules and MOE routers added to multiple linear modules in the Transformer layer. To address this issue, we propose Mixture of Low-Rank Adaptation (MiLoRA), a novel and efficient LoRA variant. MiLoRA differs from previous MOE-style LoRA methods by considering each LoRA module as an expert and employing a prompt-aware routing mechanism. This mechanism calculates expert routing results once before generating the first new token and reuses these results for subsequent tokens, reducing latency. Extensive experiments and analysis on commonsense reasoning tasks, math reasoning tasks, and widely used LLM evaluation benchmarks demonstrate that MiLoRA consistently outperforms strong PEFT baselines with comparable tunable parameter budgets. Additionally, MiLoRA significantly reduces latency in multi-tenant settings compared to previous LoRA-based methods.

This paper introduces the concept of actionability in the context of bias measures in natural language processing (NLP). We define actionability as the degree to which a measure’s results enable informed action and propose a set of desiderata for assessing it. Building on existing frameworks such as measurement modeling, we argue that actionability is a crucial aspect of bias measures that has been largely overlooked in the literature.We conduct a comprehensive review of 146 papers proposing bias measures in NLP, examining whether and how they provide the information required for actionable results. Our findings reveal that many key elements of actionability, including a measure’s intended use and reliability assessment, are often unclear or entirely absent.This study highlights a significant gap in the current approach to developing and reporting bias measures in NLP. We argue that this lack of clarity may impede the effective implementation and utilization of these measures. To address this issue, we offer recommendations for more comprehensive and actionable metric development and reporting practices in NLP bias research.

Code-switching (CS) is the process of speakers interchanging between two or more languages which in the modern world becomes increasingly common. In order to better describe CS speech the Matrix Language Frame (MLF) theory introduces the concept of a Matrix Language, which is the language that provides the grammatical structure for a CS utterance. In this work the MLF theory was used to develop systems for Matrix Language Identity (MLID) determination. The MLID of English/Mandarin and English/Spanish CS text and speech was compared to acoustic language identity (LID), which is a typical way to identify a language in monolingual utterances. MLID predictors from audio show higher correlation with the textual principles than LID in all cases while also outperforming LID in an MLID recognition task based on F1 macro (60%) and correlation score (0.38). This novel approach has identified that non-English languages (Mandarin and Spanish) are preferred over the English language as the ML contrary to the monolingual choice of LID.

The popularity of Large Language Models (LLMs) have unleashed a new age of Language Agents for solving a diverse range of tasks. While contemporary frontier LLMs are capable enough to power reasonably good Language agents, the closed-API model makes it hard to improve in cases they perform sub-optimally. To address this, recent works have explored techniques to improve their performance using self reflection and prompt optimization techniques. While techniques like self reflection work well in an online setup, contemporary prompt optimization techniques are designed to work on simpler tasks. To address this, we introduce METAREFLECTION, a novel offline reinforcement learning technique that enhances the performance of Language Agents by augmenting a semantic memory based on experiential learnings from past trials. We demonstrate the efficacy of METAREFLECTION by evaluating across multiple domains, including complex logical reasoning, biomedical semantic similarity, open world question answering, and vulnerability threat detection, in Infrastructure-as-Code, with different agent design. METAREFLECTION boosts Language agents’ performance by 4 % to 16.82 % over the raw GPT-4 baseline and performs on par with existing state-of-the-art prompt optimization techniques while requiring fewer LLM calls.

Traditional keyphrase prediction methods predict a single set of keyphrases per document, failing to cater to the diverse needs of users and downstream applications. To bridge the gap, we introduce on-demand keyphrase generation, a novel paradigm that requires keyphrases that conform to specific high-level goals or intents. For this task, we present MetaKP, a large-scale benchmark comprising four datasets, 7500 documents, and 3760 goals across news and biomedical domains with human-annotated keyphrases. Leveraging MetaKP, we design both supervised and unsupervised methods, including a multi-task fine-tuning approach and a self-consistency prompting method with large language models. The results highlight the challenges of supervised fine-tuning, whose performance is not robust to distribution shifts. By contrast, the proposed self-consistency prompting approach greatly improves the performance of large language models, enabling GPT-4o to achieve 0.548 SemF1, surpassing the performance of a fully fine-tuned BART-base model. Finally, we demonstrate the potential of our method to serve as a general NLP infrastructure, exemplified by its application in epidemic event detection from social media.

The advent of large language models (LLMs) like GPT-4 has catalyzed the exploration of multi-task learning (MTL), in which a single model demonstrates proficiency across diverse tasks. Task arithmetic has emerged as a cost-effective approach for MTL. It enables performance enhancement across multiple tasks by adding their corresponding task vectors to a pre-trained model. However, the current lack of a method that can simultaneously achieve optimal performance, computational efficiency, and data privacy limits their application to LLMs. In this paper, we propose Model Exclusive Task Arithmetic for merging GPT-scale models (MetaGPT) which formalizes the objective of model merging into a multi-task learning framework, aiming to minimize the average loss difference between the merged model and each individual task model. Since data privacy limits the use of multi-task training data, we leverage LLMs’ local linearity and task vectors’ orthogonality to separate the data term and scaling coefficients term and derive a model-exclusive task arithmetic method. Our proposed MetaGPT is data-agnostic and bypasses the heavy search process, making it cost-effective and easy to implement for LLMs. Extensive experiments demonstrate that MetaGPT leads to improvement of task arithmetic and achieves state-of-the-art performance on multiple tasks.

Emotion classification has wide applications in education, robotics, virtual reality, etc. However, identifying subtle differences between fine-grained emotion categories remains challenging. Current methods typically aggregate numerous token embeddings of a sentence into a single vector, which, while being an efficient compressor, may not fully capture complex semantic and temporal distributions. To solve this problem, we propose SEmantic ANchor Graph Neural Networks (SEAN-GNN) for fine-grained emotion classification. It learns a group of representative, multi-faceted semantic anchors in the token embedding space: using these anchors as a global reference, any sentence can be projected onto them to form a “semantic-anchor graph”, with node attributes and edge weights quantifying the semantic and temporal information respectively. The graph structure is well aligned across sentences and, importantly, allows for generating comprehensive emotion representations regarding K different anchors. Message passing on this graph can further integrate and refine the learned features. Empirically, SEAN-GNN can generate meaningful semantic anchors and discriminative graph patterns for different emotion, with promising classification results on 6 popular benchmark datasets against state-of-the-arts.

Adapting general-purpose language models to new skills is currently an expensive process that must be repeated as new instruction datasets targeting new skills are created, or can cause the models to forget older skills. In this work, we investigate the effectiveness of adding new skills to preexisting models by training on the new skills in isolation and later merging with the general model (e.g. using task vectors). In experiments focusing on scientific literature understanding, safety, and coding, we find that the parallel-train-then-merge procedure, which is significantly cheaper than retraining the models on updated data mixtures, is often comparably effective. Our experiments also show that parallel training is especially well-suited for enabling safety features in LMs relative to continued finetuning and retraining, as it dramatically improves model compliance with safe prompts while preserving its ability to refuse dangerous or harmful prompts.

Metaphor, as an advanced form of cognition, is challenging to understand their meaning. Current metaphor detection tasks only provide labels (i.e., metaphor or literal) without interpreting how to understand them. In this paper, we improve the metaphor detection task and explore the reason of metaphor. To the best of our knowledge, we are the first work to reason about metaphor using mainstream Large Language Models (LLMs). Specifically, we utilized ChatGPT3.5 to expand the mainstream datasets in current metaphor detection, including VUA ALL, TroFi, and MOH-X. We input the original sentence, target word, and usage (metaphor or literal) into ChatGPT, guiding it to generate corresponding metaphor reason. Then, we designed supervised baseline experiments (e.g., RoBERTa, GPT-2) and zero-shot experiments with LLMs (e.g., LLaMA3). For the results generated by the above experiments, we provided the case study. We devised four methods that include manual evaluation to evaluate the reason performance of the model, and discussed extensively the advantages and disadvantages of these evaluation methods. Our code is available at https://github.com/yc-cy/Metaphorical-Reasoning.

Large Language Models (LLMs) have displayed remarkable performances across various complex tasks by leveraging Chain-of-Thought (CoT) prompting. Recently, studies have proposed a Knowledge Distillation (KD) approach, reasoning distillation, which transfers such reasoning ability of LLMs through fine-tuning language models of multi-step rationales generated by LLM teachers. However, they have inadequately considered two challenges regarding insufficient distillation sets from the LLM teacher model, in terms of 1) data quality and 2) soft label provision. In this paper, we propose Mentor-KD, which effectively distills the multi-step reasoning capability of LLMs to smaller LMs while addressing the aforementioned challenges. Specifically, we exploit a mentor, intermediate-sized task-specific fine-tuned model, to augment additional CoT annotations and provide soft labels for the student model during reasoning distillation. We conduct extensive experiments and confirm Mentor-KD’s effectiveness across various models and complex reasoning tasks.

Mental disorders pose a global challenge, aggravated by the shortage of qualified mental health professionals. Mental disorder prediction from social media posts by current LLMs is challenging due to the complexities of sequential text data and the limited context length of language models. Current language model-based approaches split a single data instance into multiple chunks to compensate for limited context size. The predictive model is then applied to each chunk individually, and the most voted output is selected as the final prediction. This results in the loss of inter-post dependencies and important time variant information, leading to poor performance. We propose a novel framework which first compresses the large sequence of chronologically ordered social media posts into a series of numbers. We then use this time variant representation for mental disorder classification. We demonstrate the generalization capabilities of our framework by outperforming the current SOTA in three different mental conditions: depression, self-harm, and anorexia, by an absolute improvement of 5% in the F1 score. We also investigate the situation when current data instances fall within the context length of language models and present empirical results highlighting the importance of temporal properties of textual data. Furthermore, we utilize the proposed framework for a cross-domain study, exploring commonalities across disorders and the possibility of inter-domain data usage.

Fine-tuning provides an effective means to specialize pre-trained models for various downstream tasks. However, fine-tuning often incurs high memory overhead, especially for large transformer-based models, such as LLMs. While existing methods may reduce certain parts of the memory required for fine-tuning, they still require caching all intermediate activations computed in the forward pass to update weights during the backward pass. In this work, we develop TokenTune, a method to reduce memory usage, specifically the memory to store intermediate activations, in the fine-tuning of transformer-based models. During the backward pass, TokenTune approximates the gradient computation by backpropagating through just a subset of input tokens. Thus, with TokenTune, only a subset of intermediate activations are cached during the forward pass. Also, TokenTune can be easily combined with existing methods like LoRA, further reducing the memory cost. We evaluate our approach on pre-trained transformer models with up to billions of parameters, considering the performance on multiple downstream tasks such as text classification and question answering in a few-shot learning setup. Overall, TokenTune achieves performance on par with full fine-tuning or representative memory-efficient fine-tuning methods, while greatly reducing the memory footprint, especially when combined with other methods with complementary memory reduction mechanisms. We hope that our approach will facilitate the fine-tuning of large transformers, in specializing them for specific domains or co-training them with other neural components from a larger system. Our code is available at https://github.com/facebookresearch/tokentune.

The evolution of Large Language Models (LLMs) has led to significant advancements, with models like Claude and Gemini capable of processing contexts up to 1 million tokens. However, efficiently handling long sequences remains challenging, particularly during the prefilling stage when input lengths exceed GPU memory capacity. Traditional methods often segment sequence into chunks and compress them iteratively with fixed-size memory. However, our empirical analysis shows that the fixed-size memory results in wasted computational and GPU memory resources. Therefore, we introduces Incremental Memory (IM), a method that starts with a small memory size and gradually increases it, optimizing computational efficiency. Additionally, we propose Decremental Chunk based on Incremental Memory (IMDC), which reduces chunk size while increasing memory size, ensuring stable and lower GPU memory usage. Our experiments demonstrate that IMDC is consistently faster (1.45x) and reduces GPU memory consumption by 23.3% compared to fixed-size memory, achieving comparable performance on the LongBench Benchmark.

The complexity of text-embedded images presents a formidable challenge in machine learning given the need for multimodal understanding of multiple aspects of expression conveyed by them. While previous research in multimodal analysis has primarily focused on singular aspects such as hate speech and its subclasses, this study expands this focus to encompass multiple aspects of linguistics: hate, targets of hate, stance, and humor. We introduce a novel dataset PrideMM comprising 5,063 text-embedded images associated with the LGBTQ+ Pride movement, thereby addressing a serious gap in existing resources. We conduct extensive experimentation on PrideMM by using unimodal and multimodal baseline methods to establish benchmarks for each task. Additionally, we propose a novel framework MemeCLIP for efficient downstream learning while preserving the knowledge of the pre-trained CLIP model. The results of our experiments show that MemeCLIP achieves superior performance compared to previously proposed frameworks on two real-world datasets. We further compare the performance of MemeCLIP and zero-shot GPT-4 on the hate classification task. Finally, we discuss the shortcomings of our model by qualitatively analyzing misclassified samples. Our code and dataset are publicly available at: https://github.com/SiddhantBikram/MemeCLIP.

As we all know, hallucinations prevail in Large Language Models (LLMs), where the generated content is coherent but factually incorrect, which inflicts a heavy blow on the widespread application of LLMs. Previous studies have shown that LLMs could confidently state non-existent facts rather than answering “I don’t know”. Therefore, it is necessary to resort to external knowledge to detect and correct the hallucinated content. Since manual detection and correction of factual errors is labor-intensive, developing an automatic end-to-end hallucination-checking approach is indeed a needful thing. To this end, we present Medico, a Multi-source evidence fusion enhanced hallucination detection and correction framework. It fuses diverse evidence from multiple sources, detects whether the generated content contains factual errors, provides the rationale behind the judgment, and iteratively revises the hallucinated content. Experimental results on evidence retrieval (0.964 HR@5, 0.908 MRR@5), hallucination detection (0.927-0.951 F1), and hallucination correction (0.973-0.979 approval rate) manifest the great potential of Medico. A video demo of Medico can be found at https://youtu.be/RtsO6CSesBI.

Several recent works seek to develop foundation models specifically for medical applications, adapting general-purpose large language models (LLMs) and vision-language models (VLMs) via continued pretraining on publicly available biomedical corpora. These works typically claim that such domain-adaptive pretraining (DAPT) improves performance on downstream medical tasks, such as answering medical licensing exam questions. In this paper, we compare seven public “medical” LLMs and two VLMs against their corresponding base models, arriving at a different conclusion: all medical VLMs and nearly all medical LLMs fail to consistently improve over their base models in the zero-/few-shot prompting regime for medical question-answering (QA) tasks. For instance, across the tasks and model pairs we consider in the 3-shot setting, medical LLMs only outperform their base models in 12.1% of cases, reach a (statistical) tie in 49.8% of cases, and are significantly worse than their base models in the remaining 38.2% of cases. Our conclusions are based on (i) comparing each medical model head-to-head, directly against the corresponding base model; (ii) optimizing the prompts for each model separately; and (iii) accounting for statistical uncertainty in comparisons. While these basic practices are not consistently adopted in the literature, our ablations show that they substantially impact conclusions. Our findings suggest that state-of-the-art general-domain models may already exhibit strong medical knowledge and reasoning capabilities, and offer recommendations to strengthen the conclusions of future studies.

Framing is used to present some selective aspects of an issue and making them more salient, which aims to promote certain values, interpretations, or solutions (Entman, 1993). This study investigates the nuances of media framing on public perception and understanding by examining how events are presented within news articles. Unlike previous research that primarily focused on word choice as a framing device, this work explores the comprehensive narrative construction through events and their relations. Our method integrates event extraction, cross-document event coreference, and causal relationship mapping among events to extract framing devices employed by media to assess their role in framing the narrative. We evaluate our approach with a media attitude detection task and show that the use of event mentions, event cluster descriptors, and their causal relations effectively captures the subtle nuances of framing, thereby providing deeper insights into the attitudes conveyed by news articles. The experimental results show the framing device models surpass the baseline models and offers a more detailed and explainable analysis of media framing effects. We make the source code and dataset publicly available.

Medical task-oriented dialogue systems can assist doctors by collecting patient medical history, aiding in diagnosis, or guiding treatment selection, thereby reducing doctor burnout and expanding access to medical services. However, doctor-patient dialogue datasets are not readily available, primarily due to privacy regulations. Moreover, existing datasets lack comprehensive annotations involving medical slots and their different attributes, such as symptoms and their onset, progression, and severity. These comprehensive annotations are crucial for accurate diagnosis. Finally, most existing datasets are non-English, limiting their utility for the larger research community.In response, we introduce MediTOD, a new dataset of doctor-patient dialogues in English for the medical history-taking task. Collaborating with doctors, we devise a questionnaire-based labeling scheme tailored to the medical domain. Then, medical professionals create the dataset with high-quality comprehensive annotations, capturing medical slots and their attributes. We establish benchmarks in supervised and few-shot settings on MediTOD for natural language understanding, policy learning, and natural language generation subtasks, evaluating models from both TOD and biomedical domains. We make MediTOD publicly available for future research.