论文检索

Kai Sun, Hao Wu, Bin Shi, Samuel Mensah, Peng Liu, Bo Dong

Deep learning approaches for multimodal aspect-level sentiment classification (MALSC) often require extensive data, which is costly and time-consuming to obtain. To mitigate this, current methods typically fine-tune small-scale pretrained models like BERT and BART with few-shot examples. While these models have shown success, Large Vision-Language Models (LVLMs) offer significant advantages due to their greater capacity and ability to understand nuanced language in both zero-shot and few-shot settings. However, there is limited work on fine-tuning LVLMs for MALSC. A major challenge lies in selecting few-shot examples that effectively capture the underlying patterns in data for these LVLMs. To bridge this research gap, we propose an acquisition function designed to select challenging samples for the few-shot learning of LVLMs for MALSC. We compare our approach, Verification and ZERO-shot feedback acquisition (VERO), with diverse acquisition functions for few-shot learning in MALSC. Our experiments show that VERO outperforms prior methods, achieving an F1 score improvement of up to 6.07% on MALSC benchmark datasets.

Hugo Sousa, Satya Almasian, Ricardo Campos, Alipio Jorge

Language models have become foundational to many widely used systems. However, these seemingly advantageous models are double-edged swords. While they excel in tasks related to resource-rich languages like English, they often lose the fine nuances of language forms, dialects, and varieties that are inherent to languages spoken in multiple regions of the world. Languages like European Portuguese are neglected in favor of their more popular counterpart, Brazilian Portuguese, leading to suboptimal performance in various linguistic tasks. To address this gap, we introduce the first open-source translation model specifically tailored for European Portuguese, along with a novel dataset specifically designed for this task. Results from automatic evaluations on two benchmark datasets demonstrate that our best model surpasses existing open-source translation systems for Portuguese and approaches the performance of industry-leading closed-source systems for European Portuguese. By making our dataset, models, and code publicly available, we aim to support and encourage further research, fostering advancements in the representation of underrepresented language varieties.

Xin Song, Zhikai Xue, Guoxiu He, Jiawei Liu, Wei Lu

Parameter-efficient fine-tuning (PEFT) methods optimize large language models (LLMs) by modifying or introducing a small number of parameters to enhance alignment with downstream tasks. However, they can result in catastrophic forgetting, where LLMs prioritize new knowledge at the expense of comprehensive world knowledge. A promising approach to mitigate this issue is to recall prior memories based on the original knowledge. To this end, we propose a model-agnostic PEFT framework, IMSM, which Interweaves Memories of a Siamese Large Language Model. Specifically, our siamese LLM is equipped with an existing PEFT method. Given an incoming query, it generates two distinct memories based on the pre-trained and fine-tuned parameters. IMSM then incorporates an interweaving mechanism that regulates the contributions of both original and enhanced memories when generating the next token. This framework is theoretically applicable to all open-source LLMs and existing PEFT methods. We conduct extensive experiments across various benchmark datasets, evaluating the performance of popular open-source LLMs using the proposed IMSM, in comparison to both classical and leading PEFT methods. Our findings indicate that IMSM maintains comparable time and space efficiency to backbone PEFT methods while significantly improving performance and effectively mitigating catastrophic forgetting.

Shanbao Qiao, Xuebing Liu, Seung-Hoon Na

Model knowledge editing has become a widely researched topic because it enables efficient and rapid injection of new knowledge into language models or the correction of erroneous or outdated knowledge. Existing model knowledge editing methods typically categorized into single-instance sequential editing and massive one-time editing. However, in practical applications, the batched and iterative editing manner better aligns with model updating patterns. In this work, we explored the performance of parameter-update-based models in a new batched iterative editing benchmark. Our findings show that with an increase in the number of editing iterations, the accumulation of updated parameters leads to a greater change in the distribution of model parameters, making it more challenging to maintain editing performance and model stability. To address this degradation issue, we propose two methods: the Wasserstein distance constraint and update parameter sparsification, where the Wasserstein distance constraint optimizes the transition of parameter distribution before and after the editing, and update parameter sparsification significantly reduces the number of update parameters, thereby alleviating the issue of instability in the parameter distribution caused by the accumulation of update parameters through iterations. Our methods can be generally applied to different parameter-update-based knowledge editing models. Experiments on the zsRE and CounterFact datasets demonstrate that our methods can improve editing performance and enhance the later-stage stability of batched iterative editing across different models.

Maizhen Ning, Zihao Zhou, Qiufeng Wang, Xiaowei Huang, Kaizhu Huang

With the outstanding capabilities of Large Language Models (LLMs), solving math word problems (MWP) has greatly progressed, achieving higher performance on several benchmark datasets. However, it is more challenging to solve plane geometry problems (PGPs) due to the necessity of understanding, reasoning and computation on two modality data including both geometry diagrams and textual questions, where Multi-Modal Large Language Models (MLLMs) have not been extensively explored. Previous works simply regarded a plane geometry problem as multi-modal QA task, which ignored the importance of explicit parsing geometric elements from problems. To tackle this limitation, we propose to solve plane Geometry problems by Neural-Symbolic reasoning with MLLMs (GNS). We first leverage an MLLM to understand PGPs through knowledge prediction and symbolic parsing, next perform mathematical reasoning to obtain solutions, last adopt a symbolic solver to compute answers. Correspondingly, we introduce the largest PGPs dataset GNS-260K with multiple annotations including symbolic parsing, understanding, reasoning and computation. In experiments, our Phi3-Vision-based MLLM wins the first place on the PGPs solving task of MathVista benchmark, outperforming GPT-4o, Gemini Ultra and other much larger MLLMs. While LLaVA-13B-based MLLM markedly exceeded other close-source and open-source MLLMs on the MathVerse benchmark and also achieved the new SOTA on GeoQA dataset.

Shiwen Ni, Xiangtao Kong, Chengming Li, Xiping Hu, Ruifeng Xu, Jia Zhu, Min Yang

The success of Large Language Models (LLMs) relies heavily on the huge amount of pre-training data learned in the pre-training phase. The opacity of the pre-training process and the training data causes the results of many benchmark tests to become unreliable. If any model has been trained on a benchmark test set, it can seriously hinder the health of the field. In order to automate and efficiently test the capabilities of large language models, numerous mainstream benchmarks adopt a multiple-choice format. As the swapping of the contents of multiple-choice options does not affect the meaning of the question itself, we propose a simple and effective data leakage detection method based on this property. Specifically, we shuffle the contents of the options in the data to generate the corresponding derived data sets, and then detect data leakage based on the model's log probability distribution over the derived data sets. If there is a maximum and outlier in the set of log probabilities, it indicates that the data is leaked. Our method is able to work under gray-box conditions without access to model training data or weights, effectively identifying data leakage from benchmark test sets in model pre-training data, including both normal scenarios and complex scenarios where options may have been shuffled intentionally or unintentionally. Through experiments based on two LLMs and benchmark designs, we demonstrate the effectiveness of our method. In addition, we evaluate the degree of data leakage of 35 mainstream open-source LLMs on four benchmark datasets and give a ranking of the leaked LLMs for each benchmark, and we find that the Qwen family of LLMs has the highest degree of data leakage.

Mahdi Mostajabdaveh, Timothy Tin Long Yu, Samarendra Chandan Bindu Dash, Rindra Ramamonjison, Jabo Serge Byusa, Giuseppe Carenini, Zirui Zhou, Yong Zhang

In this paper, we introduce and apply Operations Research Question Answering (ORQA), a new benchmark, to assess the generalization capabilities of Large Language Models (LLMs) in the specialized technical domain of Operations Research (OR). This benchmark is designed to evaluate whether LLMs can emulate the knowledge and reasoning skills of OR experts when given diverse and complex optimization problems. The dataset, crafted by OR experts, presents real-world optimization problems that require multistep reasoning to build their mathematical models. Our evaluations of various open-source LLMs, such as LLaMA 3.1, DeepSeek, and Mixtral reveal their modest performance, indicating a gap in their aptitude to generalize to specialized technical domains. This work contributes to the ongoing discourse on LLMs’ generalization capabilities, providing insights for future research in this area. The dataset and evaluation code are publicly available.

Seyedreza Mohseni, Seyedali Mohammadi, Deepa Tilwani, Yash Saxena, Gerald Ketu Ndawula, Sriram Vema, Edward Raff, Manas Gaur

Malware authors often employ code obfuscations to make their malware harder to detect. Existing tools for generating obfuscated code often require access to the original source code (e.g., C++ or Java), and adding new obfuscations is a non-trivial, labor-intensive process. In this study, we ask the following question: Can Large Language Models (LLMs) potentially generate a new obfuscated assembly code? If so, this poses a risk to anti-virus engines and potentially increases the flexibility of attackers to create new obfuscation patterns. We answer this in the affirmative by developing the MetamorphASM benchmark comprising MetamorphASM Dataset (MAD) along with three code obfuscation techniques: dead code, register substitution, and control flow change. The MetamorphASM systematically evaluates the ability of LLMs to generate and analyze obfuscated code using MAD, which contains 328,200 obfuscated assembly code samples. We release this dataset and analyze the success rate of various LLMs (e.g., GPT-3.5/4, GPT-4o-mini, Starcoder, CodeGemma, CodeLlama, CodeT5, and LLaMA 3.1) in generating obfuscated assembly code. The evaluation was performed using established information-theoretic metrics and manual human review to ensure correctness and provide the foundation for researchers to study and develop remediations to this risk.

Ziyang Ma, Guanrou Yang, Yifan Yang, Zhifu Gao, Jiaming Wang, Zhihao Du, Fan Yu, Qian Chen, Siqi Zheng, Shiliang Zhang 等

In this paper, we focus on prompting one of the most important tasks in the field of speech processing, i.e., automatic speech recognition (ASR), with speech foundation encoders and large language models (LLM). Despite the growing body of research in this area, we find that many crucial design decisions in LLM-based ASR systems are often inadequately justified. This lack of clarity impedes the field's progress, making it challenging to pinpoint which design choices truly improve model performance. To address these challenges, we conduct a comprehensive series of experiments that explore various aspects, leading to the optimal LLM-based ASR system. We found that delicate designs are not necessary, while a clean setup with little task-specific design is competent. The models achieve strong performance on the Librispeech and Gigaspeech datasets, compared to both LLM-based models and non-LLM-based models. Finally, we explore the capability emergence of LLM-based ASR in the process of modal alignment. We hope that our study can facilitate the research on extending LLM with cross-modality capacity and shed light on the LLM-based ASR community.

Marcus Ma, Duong Minh Le, Junmo Kang, Yao Dou, John Cadigan, Dayne Freitag, Alan Ritter, Wei Xu

Authorship models have historically generalized poorly to new domains because of the wide distribution of author-identifying signals across domains. In particular, the effects of topic and genre are highly domain-dependent and impact authorship analysis performance greatly. This paper addresses the existing data gap in authorship for these resources by introducing CROSSNEWS, a novel cross-genre dataset that connects formal journalistic articles and casual social media posts. CROSSNEWS is the largest authorship dataset of its kind for supporting both verification and attribution tasks, with comprehensive topic and genre annotations. We use CROSSNEWS to demonstrate that current models exhibit poor performance in genre transfer scenarios, underscoring the need for authorship models robust to genre-specific effects. We also explore SELMA, a new LLM embedding approach for large-scale authorship setups that outperforms existing models in both same-genre and cross-genre settings.

Guangyuan Ma, Yongliang Ma, Xing Wu, Zhenpeng Su, Ming Zhou, Songlin Hu

Large Language Model-based Dense Retrieval (LLM-DR) optimizes over numerous heterogeneous fine-tuning collections from different domains. However, the discussion about its training data distribution is still minimal. Previous studies rely on empirically assigned dataset choices or sampling ratios, which inevitably lead to sub-optimal retrieval performances. In this paper, we propose a new task-level Distributionally Robust Optimization (tDRO) algorithm for LLM-DR fine-tuning, targeted at improving the universal domain generalization ability by end-to-end reweighting the data distribution of each task. The tDRO parameterizes the domain weights and updates them with scaled domain gradients. The optimized weights are then transferred to the LLM-DR fine-tuning to train more robust retrievers. Experiments show optimal improvements in large-scale retrieval benchmarks and reduce up to 30% dataset usage after applying our optimization algorithm with a series of different-sized LLM-DR models.

Penghao Lu, Xin Dong, Yuansheng Zhou, Lei Cheng, Chuan Yuan, Linjian Mo

Generative retrieval constitutes an innovative approach in information retrieval, leveraging generative language models(LM) to generate a ranked list of document identifiers (docid) for a given query. It simplifies the retrieval pipeline by replacing the large external index with model parameters. However, existing works merely learned the relationship between queries and document identifiers, which is unable to directly represent the relevance between queries and documents. To address the above problem, we propose a novel and general generative retrieval framework, namely Leveraging Document-Oriented Contrastive Learning in Generative Retrieval (DOGR), which leverages contrastive learning to improve generative retrieval tasks. It adopts a two-stage learning strategy that captures the relationship between queries and documents comprehensively through direct interactions. Furthermore, negative sampling methods and corresponding contrastive learning objectives are implemented to enhance the learning of semantic representations, thereby promoting a thorough comprehension of the relationship between queries and documents. Experimental results demonstrate that DOGR achieves state-of-the-art performance compared to existing generative retrieval methods on two public benchmark datasets. Further experiments have shown that our framework is generally effective for common identifier construction techniques.

Zesheng Liu, Qiannan Zhu, Cuiping Li, Hong Chen

Catastrophic forgetting is a key challenge in incremental named entity recognition (INER). Existing methods often address this issue through distillation-based approaches, which involve transferring previously learned knowledge from the old model to the new one. However, these methods may not fully equip the new model with an adequate understanding of the characteristics about old entity types, leading to confusion when classifying tokens associated with these entity types. To address this challenge, we propose a novel method called Prototypical Replay with Old-class Focusing Knowledge Distillation (POF) for INER. Our approach focuses on preserving the main characteristics of each previous entity type by storing compact prototypes and replaying them with appropriate frequency. This replay strategy makes the new model review the knowledge of old entity types while minimizing storage needs. Additionally, we introduce an old-class focusing knowledge distillation (OFKD) loss, which distills features only in old-class regions to maintain the quality of old-class prototypes and prevent ineffective prototypical replay while preserving sufficient plasticity for learning new entity types. We conducted experiments on three benchmark datasets (i.e., Few-NERD, I2B2 and OntoNotes5), and the results demonstrate that our method outperforms all previous state-of-the-art methods.

Yonghao Liu, Mengyu Li, Wei Pang, Fausto Giunchiglia, Lan Huang, Xiaoyue Feng, Renchu Guan

Short text classification, as a research subtopic in natural language processing, is more challenging due to its semantic sparsity and insufficient labeled samples in practical scenarios. We propose a novel model named MI-DELIGHT for short text classification in this work. Specifically, it first performs multi-source information (i.e., statistical information, linguistic information, and factual information) exploration to alleviate the sparsity issues. Then, the graph learning approach is adopted to learn the representation of short texts, which are presented in graph forms. Moreover, we introduce a dual-level (i.e., instance-level and cluster-level) contrastive learning auxiliary task to effectively capture different-grained contrastive information within massive unlabeled data. Meanwhile, previous models merely perform the main task and auxiliary tasks in parallel, without considering the relationship among tasks. Therefore, we introduce a hierarchical architecture to explicitly model the correlations between tasks. We conduct extensive experiments across various benchmark datasets, demonstrating that MI-DELIGHT significantly surpasses previous competitive models. It even outperforms popular large language models on several datasets.

Xukun Liu, Bowen Lei, Ruqi Zhang, Dongkuan (DK) Xu

Large language model (LLM) decoding involves generating a sequence of tokens based on a given context, where each token is predicted one at a time using the model's learned probabilities. The typical autoregressive decoding method requires a separate forward pass through the model for each token generated, which is computationally inefficient and poses challenges for deploying LLMs in latency-sensitive scenarios. The main limitations of current decoding methods stem from their inefficiencies and resource demands. Existing approaches either necessitate fine-tuning smaller models, which is resource-intensive, or relying on fixed retrieval schemes to construct drafts for the next tokens, which lack adaptability and fail to generalize across different models and contexts. To address these issues, we introduce a novel methodology called Adaptix, which accelerates LLM decoding without requiring fine-tuning. Our approach involves an adaptive draft-verification process that evolves over time to improve efficiency. We utilize a tri-gram matrix-based LLM representation to dynamically approximate the output distribution of the LLM, allowing the model to adjust to changing token probabilities during the decoding process. Additionally, we implement a draft construction mechanism that effectively balances exploration and exploitation, ensuring that the drafts generated are both diverse and close to the true output distribution of the LLM. The importance of this design lies in its ability to optimize the draft distribution adaptively, leading to faster and more accurate decoding. Through extensive experiments on various benchmark datasets and LLM architectures, we demonstrate that Adaptix significantly accelerates the decoding process while maintaining high accuracy, making it suitable for deployment in a wide range of practical applications.

Jiaqing Liu, Chong Deng, Qinglin Zhang, Shilin Zhou, Qian Chen, Hai Yu, Wen Wang

Automatic Speech Recognition (ASR) transcripts exhibit recognition errors and various spoken language phenomena such as disfluencies, ungrammatical sentences, and incomplete sentences, hence suffering from poor readability. To improve readability, we propose a Contextualized Spoken-to-Written conversion (CoS2W) task to address ASR and grammar errors and also transfer the informal text into the formal style with content preserved, utilizing contexts and auxiliary information. This task naturally matches the in-context learning capabilities of Large Language Models (LLMs). To facilitate comprehensive comparisons of various LLMs, we construct a document-level Spoken-to-Written conversion of ASR Transcripts Benchmark (SWAB) dataset. Using SWAB, we study the impact of different granularity levels on the CoS2W performance, and propose methods to exploit contexts and auxiliary information to enhance the outputs. Experimental results reveal that LLMs have the potential to excel in the CoS2W task, particularly in grammaticality and formality, our methods achieve effective understanding of contexts and auxiliary information by LLMs. We further investigate the effectiveness of using LLMs as evaluators and find that LLM evaluators show strong correlations with human evaluations on rankings of faithfulness and formality, which validates the reliability of LLM evaluators for the CoS2W task.

Jianjian Liu, Zhengtao Yu, Ying Li, Yuxin Huang, Shengxiang Gao

Pre-trained language models enhanced parsers have achieved outstanding performance in rich-resource languages. Cross-lingual dependency parsing aims to learn useful knowledge from high-resource languages to alleviate data scarcity in low-resource languages. However, effectively reducing the syntactic structure distributional bias and excavating the commonalities among languages is the key challenge for cross-lingual dependency parsing. To address this issue, we propose novel dynamic syntactic feature filtering and injecting networks based on the typical shared-private model that employs one shared and two private encoders to separate source and target language features. Concretely, a Language-Specific Filtering Network (LSFN) on private encoders emphasizes helpful information and ignores the irrelevant or harmful parts of it from the source language. Meanwhile, a Language-Invariant Injecting Network (LIIN) on the shared encoder integrates the advantages of BiLSTM and improved Transformer encoders to transcend language boundaries, thus amplifying syntactic commonalities across languages. Experiments on seven benchmark datasets show that our model achieves an average absolute gain of 1.84 UAS and 3.43 LAS compared with the shared-private model. Comparative experiments validate that both LSFN and LIIN components are complementary in transferring beneficial knowledge from source to target languages. Detailed analyses highlight that our model can effectively capture linguistic commonalities and mitigate the effect of distributional bias, showcasing its robustness and efficacy.

Haoxiong Liu, Yifan Zhang, Yifan Luo, Andrew C Yao

Despite the advancements in large language models (LLMs) for mathematical reasoning, solving competition-level math problems remains a significant challenge, especially for open-source LLMs without external tools. We introduce the MMIQC dataset, comprising a mixture of processed web data and synthetic question-response pairs, aimed at enhancing the mathematical reasoning capabilities of base language models. Models fine-tuned on MMIQC consistently surpass their counterparts in performance on the MATH benchmark across various model sizes. Notably, Qwen-72B-MMIQC achieves a 45.0% accuracy, exceeding the previous open-source state-of-the-art by 8.2% and outperforming the initial version GPT-4 released in 2023. Extensive evaluation results on Hungarian high school finals suggest that such improvement can generalize to unseen data. Our ablation study on MMIQC reveals that a large part of the improvement can be attributed to our novel augmentation method, Iterative Question Composing (IQC), which involves iteratively composing new questions from seed problems using an LLM and applying rejection sampling through another LLM.

Barys Liskavets, Maxim Ushakov, Shuvendu Roy, Mark Klibanov, Ali Etemad, Shane K. Luke

Large language models (LLMs) have triggered a new stream of research focusing on compressing the context length to reduce the computational cost while ensuring the retention of helpful information for LLMs to answer the given question. Token-based removal methods are one of the most prominent approaches in this direction, but risk losing the semantics of the context caused by intermediate token removal, especially under high compression ratios, while also facing challenges in computational efficiency. In this work, we propose context-aware prompt compression (CPC), a sentence-level prompt compression technique where its key innovation is a novel context-aware sentence encoder that provides a relevance score for each sentence for a given question. To train this encoder, we generate a new dataset consisting of questions, positives, and negative pairs where positives are sentences relevant to the question, while negatives are irrelevant context sentences. We train the encoder in a contrastive setup to learn context-aware sentence representations. Our method considerably outperforms prior works on prompt compression on benchmark datasets and is up to 10.93x faster at inference compared to the best token-level compression method. We also find better improvement for shorter length constraints in most benchmarks, showing the effectiveness of our proposed solution in the compression of relevant information in a shorter context. Finally, we release the code and the dataset for quick reproducibility and further development.

Xiaopeng Li, Shasha Li, Shezheng Song, Huijun Liu, Bin Ji, Xi Wang, Jun Ma, Jie Yu, Xiaodong Liu, Jing Wang 等

The general capabilities of large language models (LLMs) make them the infrastructure for various AI applications, but updating their inner knowledge requires significant resources. Recent model editing is a promising technique for efficiently updating a small amount of knowledge of LLMs and has attracted much attention. In particular, local editing methods, which directly update model parameters, are proven suitable for updating small amounts of knowledge. Local editing methods update weights by computing least squares closed-form solutions and identify edited knowledge by vector-level matching in inference, which achieve promising results. However, these methods still require a lot of time and resources to complete the computation. Moreover, vector-level matching lacks reliability, and such updates disrupt the original organization of the model's parameters. To address these issues, we propose a detachable and expandable Subject Word Embedding Altering (SWEA) framework, which finds the editing embeddings through token-level matching and adds them to the subject word embeddings in Transformer input. To get these editing embeddings, we propose optimizing then suppressing fusion method, which first optimizes learnable embedding vectors for the editing target and then suppresses the Knowledge Embedding Dimensions (KEDs) to obtain final editing embeddings. We thus propose SWEAOS method for editing factual knowledge in LLMs. We demonstrate the overall state-of-the-art (SOTA) performance of SWEAOS on the CounterFact and zsRE datasets. To further validate the reasoning ability of SWEAOS in editing knowledge, we evaluate it on the more complex RippleEdits benchmark. The results demonstrate that SWEAOS possesses SOTA reasoning ability.