Arrow Research search

Author name cluster

Le Sun

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

31 papers
2 author rows

Possible papers

31

JBHI Journal 2026 Journal Article

A 6G-Enabled Hierarchical Contrastive Learning Framework for Multi-Scale Medical Time Series Analysis

  • Le Sun
  • Jie Lin
  • Zhiguo Qu
  • Yimin Yu
  • Jinliang Liu
  • Deepak Gupta
  • Yanchun Zhang

Medical time series analysis, particularly for electrocardiogram (ECG) and electroencephalogram (EEG) signals, is essential in modern diagnostics, supporting early detection of conditions such as arrhythmias and epileptic seizures. However, existing approaches often struggle to capture multi-scale periodic patterns and longrange dependencies while meeting real-time processing demands. The envisioned 6G networks, with their terahertz communication and integrated sensing and communication (ISAC) capabilities, will generate vast volumes of high-fidelity physiological data at the network edge. This paradigm shift intensifies the conflict between the computational complexity of advanced AI models and the limited resources of edge devices, creating a critical bottleneck for deploying sophisticated analytics in real-world healthcare scenarios. To overcome these limitations, this paper introduces a 6G-enabled hierarchical contrastive learning framework, referred to as Hierarchical Contrastive Learning for Multi-Scale Medical time series analysis (HCL-MSM), which integrates three core components: a signal-adaptive encoder based on multi-period decomposition and 2D convolution, a patient-level contrastive module enhanced with decomposable multi-scale mixing, and a 6G-edge deployment module optimized via quantization and pruning. The framework effectively models nested physiological rhythms and cross-time dependencies in medical data, while maintaining low-latency operation under resource-constrained edge environments. We evaluated HCL-MSM on multiple clinical datasets under simulated 6G settings. Our framework achieves significant gains in arrhythmia detection, seizure prediction, and neurological monitoring. We evaluated HCL-MSM on multiple clinical datasets under simulated 6G settings. Our framework achieves significant gains in arrhythmia detection (F1-score: 86. 39 percent), seizure prediction (Recall: 87. 72 percent), and neurological monitoring (Recall: 87. 8 percent), outperforming existing state-of- the-art methods.

Details DOI

AAAI Conference 2026 Conference Paper

AI-Salesman: Towards Reliable Large Language Model Driven Telemarketing

  • Qingyu Zhang
  • Chunlei Xin
  • Xuanang Chen
  • Yaojie Lu
  • Hongyu Lin
  • Xianpei Han
  • Le Sun
  • Qing Ye

Goal-driven persuasive dialogue, exemplified by applications like telemarketing, requires sophisticated multi-turn planning and strict factual faithfulness, which remains a significant challenge for even state-of-the-art Large Language Models (LLMs). A lack of task-specific data often limits previous works, and direct LLM application suffers from strategic brittleness and factual hallucination. In this paper, we first construct and release TeleSalesCorpus, the first real-world-grounded dialogue dataset for this domain. We then propose AI-Salesman, a novel framework featuring a dual-stage architecture. For the training stage, we design a Bayesian-supervised reinforcement learning algorithm that learns robust sales strategies from noisy dialogues. For the inference stage, we introduce the Dynamic Outline-Guided Agent (DOGA), which leverages a pre-built script library to provide dynamic, turn-by-turn strategic guidance. Moreover, we design a comprehensive evaluation framework that combines fine-grained metrics for key sales skills with the LLM-as-a-Judge paradigm. Experimental results demonstrate that our proposed AI-Salesman significantly outperforms baseline models in both automatic metrics and comprehensive human evaluations, showcasing its effectiveness in complex persuasive scenarios.

PDF Details DOI

AAAI Conference 2026 Conference Paper

Does Question Really Matter? The Attribution of Answer Bias in LLM Evaluation

  • Boxi Cao
  • Ruotong Pan
  • Hongyu Lin
  • Xianpei Han
  • Le Sun

Multiple-choices question answering (MCQA) has emerged as one of the most popular task formats for large language models (LLMs) evaluation. Unfortunately, there exist substantial evidence that the evaluation of current MCQA benchmarks suffers from significant answer bias, which severely undermines the reliability of the evaluation conclusions. Specifically, many LLMs achieve performance significantly higher than random selection even when the questions are omitted from input information. To this end, we conduct a systematic investigation of the attribution of answer bias, and demonstrate a strong correlation between the degree of data contamination and the severity of answer bias, while the position of options and the popularity of answers have relatively minor effects. Building on these insights, we further propose OPD, a straightforward yet effective tool for contamination detection and dataset debiasing without requiring access to the model’s internal training data. Our findings and algorithms provide valuable insights for the design of future trustworthy LLM evaluation protocols.

PDF Details DOI

JBHI Journal 2026 Journal Article

Energy-Efficient Online Continual Learning for Time Series Classification in Nanorobot-Based Smart Health

  • Le Sun
  • Qingyuan Chen
  • Min Zheng
  • Xin Ning
  • Deepak Gupta
  • Prayag Tiwari

Nanorobots have been used in smart health to collect time series data such as electrocardiograms and electroencephalograms. Real-time classification of dynamic time series signals in nanorobots is a challenging task. Nanorobots in the nanoscale range require a classification algorithm with low computational complexity. First, the classification algorithm should be able to dynamically analyze time series signals and update itself to process the concept drift (CD). Second, the classification algorithm should have the ability to handle catastrophic forgetting (CF) and classify historical data. Most importantly, the classification algorithm should be energy-efficient to use less computing power and memory to classify signals in real-time on a smart nanorobot. To solve these challenges, we design an algorithm that can Prevent Concept Drift in Online continual Learning for time series classification (PCDOL). The prototype suppression item in PCDOL can reduce the impact caused by CD. It also solves the CF problem through the replay feature. The computation per second and the memory consumed by PCDOL are only 3. 572 M and 1 KB, respectively. The experimental results show that PCDOL is better than several state-of-the-art methods for dealing with CD and CF in energy-efficient nanorobots.

Details DOI

AAAI Conference 2025 Conference Paper

DOMAINEVAL: An Auto-Constructed Benchmark for Multi-Domain Code Generation

  • Qiming Zhu
  • Jialun Cao
  • Yaojie Lu
  • Hongyu Lin
  • Xianpei Han
  • Le Sun
  • Shing-Chi Cheung

Code benchmarks such as HumanEval are widely adopted to evaluate the capabilities of Large Language Models (LLMs), providing insights into their strengths and weaknesses. However, current benchmarks primarily exercise LLMs' capability on common coding tasks (e.g., bubble sort, greatest common divisor), leaving domain-specific coding tasks (e.g., computation, system, cryptography) unexplored. To fill this gap, we propose a multi-domain code benchmark, DOMAINEVAL, designed to evaluate LLMs' coding capabilities thoroughly. Our pipeline works in a fully automated manner, enabling a push-button construction from code repositories into formatted subjects under study. Interesting findings are observed by evaluating 12 representative LLMs against DOMAINEVAL. We notice that LLMs are generally good at computation tasks while falling short on cryptography and system coding tasks. The performance gap can be as much as 68.94% (80.94% - 12.0%) in some LLMs. We also observe that generating more samples can increase the overall performance of LLMs, while the domain bias may even increase. The contributions of this study include a code generation benchmark dataset DOMAINEVAL, encompassing six popular domains, a fully automated pipeline for constructing code benchmarks, and an identification of the limitations of LLMs in code generation tasks based on their performance on DOMAINEVAL, providing directions for future research improvements.

PDF Details DOI

ICLR Conference 2025 Conference Paper

Enhancing Learning with Label Differential Privacy by Vector Approximation

  • Puning Zhao
  • Jiafei Wu
  • Zhe Liu 0001
  • Li Shen 0008
  • Zhikun Zhang 0001
  • Rongfei Fan
  • Le Sun
  • Qingming Li

Label differential privacy (DP) is a framework that protects the privacy of labels in training datasets, while the feature vectors are public. Existing approaches protect the privacy of labels by flipping them randomly, and then train a model to make the output approximate the privatized label. However, as the number of classes K increases, stronger randomization is needed, thus the performances of these methods become significantly worse. In this paper, we propose a vector approximation approach for learning with label local differential privacy, which is easy to implement and introduces little additional computational overhead. Instead of flipping each label into a single scalar, our method converts each label into a random vector with K components, whose expectations reflect class conditional probabilities. Intuitively, vector approximation retains more information than scalar labels. A brief theoretical analysis shows that the performance of our method only decays slightly with K. Finally, we conduct experiments on both synthesized and real datasets, which validate our theoretical analysis as well as the practical performance of our method.

Details

AAAI Conference 2024 Conference Paper

Benchmarking Large Language Models in Retrieval-Augmented Generation

  • Jiawei Chen
  • Hongyu Lin
  • Xianpei Han
  • Le Sun

Retrieval-Augmented Generation (RAG) is a promising approach for mitigating the hallucination of large language models (LLMs). However, existing research lacks rigorous evaluation of the impact of retrieval-augmented generation on different large language models, which make it challenging to identify the potential bottlenecks in the capabilities of RAG for different LLMs. In this paper, we systematically investigate the impact of Retrieval-Augmented Generation on large language models. We analyze the performance of different large language models in 4 fundamental abilities required for RAG, including noise robustness, negative rejection, information integration, and counterfactual robustness. To this end, we establish Retrieval-Augmented Generation Benchmark (RGB), a new corpus for RAG evaluation in both English and Chinese. RGB divides the instances within the benchmark into 4 separate testbeds based on the aforementioned fundamental abilities required to resolve the case. Then we evaluate 6 representative LLMs on RGB to diagnose the challenges of current LLMs when applying RAG. Evaluation reveals that while LLMs exhibit a certain degree of noise robustness, they still struggle significantly in terms of negative rejection, information integration, and dealing with false information. The aforementioned assessment outcomes indicate that there is still a considerable journey ahead to effectively apply RAG to LLMs.

PDF Details DOI

JBHI Journal 2024 Journal Article

DCNet: A Self-Supervised EEG Classification Framework for Improving Cognitive Computing-Enabled Smart Healthcare

  • Yiyang Zhang
  • Le Sun
  • Deepak Gupta
  • Xin Ning
  • Prayag Tiwari

Cognitive computing endeavors to construct models that emulate brain functions, which can be explored through electroencephalography (EEG). Developing precise and robust EEG classification models is crucial for advancing cognitive computing. Despite the high accuracy of supervised EEG classification models, they are constrained by labor-intensive annotations and poor generalization. Self-supervised models address these issues but encounter difficulties in matching the accuracy of supervised learning. Three challenges persist: 1) capturing temporal dependencies in EEG; 2) adapting loss functions to describe feature similarities in self-supervised models; and 3) addressing the prevalent issue of data imbalance in EEG. This study introduces the DreamCatcher Network (DCNet), a self-supervised EEG classification framework with a two-stage training strategy. The first stage extracts robust representations through contrastive learning, and the second stage transfers the representation encoder to a supervised EEG classification task. DCNet utilizes time-series contrastive learning to autonomously construct representations that comprehensively capture temporal correlations. A novel loss function, SelfDreamCatcherLoss, is proposed to evaluate the similarities between these representations and enhance the performance of DCNet. Additionally, two data augmentation methods are integrated to alleviate class imbalances. Extensive experiments show the superiority of DCNet over the current state-of-the-art models, achieving high accuracy on both the Sleep-EDF and HAR datasets. It holds substantial promise for revolutionizing sleep disorder detection and expediting the development of advanced healthcare systems driven by cognitive computing.

Details DOI

JBHI Journal 2024 Journal Article

Few-Shot Class-Incremental Learning for Medical Time Series Classification

  • Le Sun
  • Mingyang Zhang
  • Benyou Wang
  • Prayag Tiwari

Continuously analyzing medical time series as new classes emerge is meaningful for health monitoring and medical decision-making. Few-shot class-incremental learning (FSCIL) explores the classification of few-shot new classes without forgetting old classes. However, little of the existing research on FSCIL focuses on medical time series classification, which is more challenging to learn due to its large intra-class variability. In this paper, we propose a framework, the Meta self-Attention Prototype Incrementer (MAPIC) to address these problems. MAPIC contains three main modules: an embedding encoder for feature extraction, a prototype enhancement module for increasing inter-class variation, and a distance-based classifier for reducing intra-class variation. To mitigate catastrophic forgetting, MAPIC adopts a parameter protection strategy in which the parameters of the embedding encoder module are frozen at incremental stages after being trained in the base stage. The prototype enhancement module is proposed to enhance the expressiveness of prototypes by perceiving inter-class relations using a self-attention mechanism. We design a composite loss function containing the sample classification loss, the prototype non-overlapping loss, and the knowledge distillation loss, which work together to reduce intra-class variations and resist catastrophic forgetting. Experimental results on three different time series datasets show that MAPIC significantly outperforms state-of-the-art approaches by 27. 99%, 18. 4%, and 3. 95%, respectively.

Details DOI

AAAI Conference 2024 Conference Paper

Mitigating Large Language Model Hallucinations via Autonomous Knowledge Graph-Based Retrofitting

  • Xinyan Guan
  • Yanjiang Liu
  • Hongyu Lin
  • Yaojie Lu
  • Ben He
  • Xianpei Han
  • Le Sun

Incorporating factual knowledge in knowledge graph is regarded as a promising approach for mitigating the hallucination of large language models (LLMs). Existing methods usually only use the user's input to query the knowledge graph, thus failing to address the factual hallucination generated by LLMs during its reasoning process. To address this problem, this paper proposes Knowledge Graph-based Retrofitting (KGR), a new framework that incorporates LLMs with KGs to mitigate factual hallucination during the reasoning process by retrofitting the initial draft responses of LLMs based on the factual knowledge stored in KGs. Specifically, KGR leverages LLMs to extract, select, validate, and retrofit factual statements within the model-generated responses, which enables an autonomous knowledge verifying and refining procedure without any additional manual efforts. Experiments show that KGR can significantly improve the performance of LLMs on factual QA benchmarks especially when involving complex reasoning processes, which demonstrates the necessity and effectiveness of KGR in mitigating hallucination and enhancing the reliability of LLMs.

PDF Details DOI

EAAI Journal 2024 Journal Article

Randomized attention and dual-path system for electrocardiogram identity recognition

  • Le Sun
  • Huiyun Li
  • Ghulam Muhammad

With the advancement in digital communication and artificial intelligence-based applications, the emphasis on information security has intensified. Traditional authentication techniques like passwords, fingerprints, and facial recognition present risks of unauthorized access and duplication. The distinctive nature of electrocardiogram (ECG) signals is drawing considerable attention as an identity verification approach. However, current ECG-based identification methods tend to possess intricate preprocessing phases and excessively rely on the accuracy of fiducial point detection. This prevents a comprehensive capture of the information within the ECG signals, thereby affecting the accuracy of identification. To address these issues, we introduce RandSaD (Random-window Split attention Dual-path), an innovative ECG identity recognition model. RandSaD combines random window extraction with a dual-path residual neural network and integrates the split attention (SpAt) mechanism. During the preprocessing stage, RandSaD operates directly on wavelet-denoised ECG signals. It extracts random windows with the objective of preserving the intrinsic characteristics of the ECG signal while minimizing information loss. Subsequently, we integrate a dual-path residual network integrated with SpAt. Each path can independently extract features of a specific scale or type, enhancing the capability of the model to extract features. The SpAt amplifies the capacity of the model to distinguish and merge information from different channels. Experimental results indicate that the RandSaD model achieves a 99. 6% accuracy rate in ECG identity recognition tasks.

Details DOI

NeurIPS Conference 2024 Conference Paper

Self-Retrieval: End-to-End Information Retrieval with One Large Language Model

  • Qiaoyu Tang
  • Jiawei Chen
  • Zhuoqun Li
  • Bowen Yu
  • Yaojie Lu
  • Cheng Fu
  • Haiyang Yu
  • Hongyu Lin

The rise of large language models (LLMs) has significantly transformed both the construction and application of information retrieval (IR) systems. However, current interactions between IR systems and LLMs remain limited, with LLMs merely serving as part of components within IR systems, and IR systems being constructed independently of LLMs. This separated architecture restricts knowledge sharing and deep collaboration between them. In this paper, we introduce Self-Retrieval, a novel end-to-end LLM-driven information retrieval architecture. Self-Retrieval unifies all essential IR functions within a single LLM, leveraging the inherent capabilities of LLMs throughout the IR process. Specifically, Self-Retrieval internalizes the retrieval corpus through self-supervised learning, transforms the retrieval process into sequential passage generation, and performs relevance assessment for reranking. Experimental results demonstrate that Self-Retrieval not only outperforms existing retrieval approaches by a significant margin, but also substantially enhances the performance of LLM-driven downstream applications like retrieval-augmented generation.

PDF Details DOI

JBHI Journal 2023 Journal Article

A Scalable and Transferable Federated Learning System for Classifying Healthcare Sensor Data

  • Le Sun
  • Jin Wu

With the development of Internet of Medical Things, massive healthcare sensor data (HSD) are transmitted in the Internet, which faces various security problems. Healthcare data are sensitive and important for patients. Automatic classification of HSD has significant value for protecting the privacy of patients. Recently, the edge computing-based federated learning has brought new opportunities and challenges. It is difficult to develop a lightweight HSD classification system for edge computing. In particular, the classification system should consider the dynamic characteristics of HSD, e. g. , the change of data distributions and the appearance of initially unknown classes. To solve these problems, the paper proposes a scalable and transferable classification system, called SCALT. It is a one-classifier-per-class system based on federated learning. It comprises a one-dimensional convolution-based network for feature extraction, and an individual mini-classifier for each class. It is easy to be scaled when new class appears since only a mini-classifier will be trained. The feature extractor is updated only when it is transferred to a new task. SCALT has a parameter protection mechanism, which can avoid catastrophic forgetting in sequential HSD classification tasks. We conduct comprehensive experiments to evaluate SCALT on three different physiological signal datasets: Electrocardiogram, Electroencephalogram and Photoplethysmograph. The accuracies on the three datasets are 98. 65%, 91. 10% and 89. 93% respectively, which are higher than the compared state-of-the-art works. At last, an application of applying SCALT to protect the privacy of patients is presented.

Details DOI

AAAI Conference 2023 Conference Paper

Dialogue Rewriting via Skeleton-Guided Generation

  • Chunlei Xin
  • Hongyu Lin
  • Shan Wu
  • Xianpei Han
  • Bo Chen
  • Wen Dai
  • Shuai Chen
  • Bin Wang

Dialogue rewriting aims to transform multi-turn, context-dependent dialogues into well-formed, context-independent text for most NLP systems. Previous dialogue rewriting benchmarks and systems assume a fluent and informative utterance to rewrite. Unfortunately, dialogue utterances from real-world systems are frequently noisy and with various kinds of errors that can make them almost uninformative. In this paper, we first present Real-world Dialogue Rewriting Corpus (RealDia), a new benchmark to evaluate how well current dialogue rewriting systems can deal with real-world noisy and uninformative dialogue utterances. RealDia contains annotated multi-turn dialogues from real scenes with ASR errors, spelling errors, redundancies and other noises that are ignored by previous dialogue rewriting benchmarks. We show that previous dialogue rewriting approaches are neither effective nor data-efficient to resolve RealDia. Then this paper presents Skeleton-Guided Rewriter (SGR), which can resolve the task of dialogue rewriting via a skeleton-guided generation paradigm. Experiments show that RealDia is a much more challenging benchmark for real-world dialogue rewriting, and SGR can effectively resolve the task and outperform previous approaches by a large margin.

PDF Details DOI

AAAI Conference 2023 Conference Paper

Universal Information Extraction as Unified Semantic Matching

  • Jie Lou
  • Yaojie Lu
  • Dai Dai
  • Wei Jia
  • Hongyu Lin
  • Xianpei Han
  • Le Sun
  • Hua Wu

The challenge of information extraction (IE) lies in the diversity of label schemas and the heterogeneity of structures. Traditional methods require task-specific model design and rely heavily on expensive supervision, making them difficult to generalize to new schemas. In this paper, we decouple IE into two basic abilities, structuring and conceptualizing, which are shared by different tasks and schemas. Based on this paradigm, we propose to universally model various IE tasks with Unified Semantic Matching (USM) framework, which introduces three unified token linking operations to model the abilities of structuring and conceptualizing. In this way, USM can jointly encode schema and input text, uniformly extract substructures in parallel, and controllably decode target structures on demand. Empirical evaluation on 4 IE tasks shows that the proposed method achieves state-of-the-art performance under the supervised experiments and shows strong generalization ability in zero/few-shot transfer settings.

PDF Details DOI

IJCAI Conference 2022 Conference Paper

Bridging the Gap between Reality and Ideality of Entity Matching: A Revisiting and Benchmark Re-Construction

  • Tianshu Wang
  • Hongyu Lin
  • Cheng Fu
  • Xianpei Han
  • Le Sun
  • Feiyu Xiong
  • Hui Chen
  • Minlong Lu

Entity matching (EM) is the most critical step for entity resolution (ER). While current deep learning-based methods achieve very impressive performance on standard EM benchmarks, their real-world application performance is much frustrating. In this paper, we highlight that such the gap between reality and ideality stems from the unreasonable benchmark construction process, which is inconsistent with the nature of entity matching and therefore leads to biased evaluations of current EM approaches. To this end, we build a new EM corpus and re-construct EM benchmarks to challenge critical assumptions implicit in the previous benchmark construction process by step-wisely changing the restricted entities, balanced labels, and single-modal records in previous benchmarks into open entities, imbalanced labels, and multi-modal records in an open environment. Experimental results demonstrate that the assumptions made in the previous benchmark construction process are not coincidental with the open environment, which conceal the main challenges of the task and therefore significantly overestimate the current progress of entity matching. The constructed benchmarks and code are publicly released at https: //github. com/tshu-w/ember.

PDF Details DOI

AIJ Journal 2022 Journal Article

End-to-end neural event coreference resolution

  • Yaojie Lu
  • Hongyu Lin
  • Jialong Tang
  • Xianpei Han
  • Le Sun

Conventional event coreference systems commonly use a pipeline architecture and rely heavily on handcrafted features, which often causes error propagation problems and leads to poor generalization ability. In this paper, we propose a neural network-based end-to-end event coreference architecture ( E 3 C ) that can jointly model event detection and event coreference resolution tasks and learn to extract features from raw text automatically. Furthermore, because event mentions are highly diversified and event coreference is intricately governed by long-distance and semantically-dependent decisions, a type-enhanced event coreference mechanism is further proposed in our E 3 C neural network. Experiments show that our method achieves a new state-of-the-art performance on both standard datasets.

Details DOI

JBHI Journal 2022 Journal Article

PerAE: An Effective Personalized AutoEncoder for ECG-Based Biometric in Augmented Reality System

  • Le Sun
  • Zhaoyi Zhong
  • Zhiguo Qu
  • N. Xiong

With the development of the Augmented and Virtual Reality (AR/VR) technologies, massive biometric data are collected by different organizations. These data have great significance but also worsen the privacy risks. Electro-CardioGram (ECG)-based Identity Recognition (EIR) is a popular Biometric technology. An ECG record is an internal Biology feature of a person and has time continuity. Thus, compared with traditional Biometric methods like face recognition, EIR may be less vulnerable to attack. We propose an Autoencoder-based EIR system, called P ersonalized A uto E ncoder (PerAE). PerAE maintains a small autoencoder model (called Attention-MemAE) for each registered user of a system. The Attention-MemAE enhances the autoencoder by using a memory module and two attention mechanisms. A user’s Attention-MemAE classifies the hearbeats of other users as anomalies. An Attention-MemAE can be updated when the distribution of the user’s ECG data is changed. By using personalized autoencoder, PerAE can improve the time efficiency and reduce the memory overhead. It improves the adaptability, scalability, and maintainability of EIR systems. Experiment results show that to train an Attention-MemAE with 90 $\%$ identification accuracy for a user, we can just take five minutes to collect the user’s ECG data (around 500 heartbeat samples).

Details DOI

AAAI Conference 2022 Conference Paper

Procedural Text Understanding via Scene-Wise Evolution

  • Jialong Tang
  • Hongyu Lin
  • Meng Liao
  • Yaojie Lu
  • Xianpei Han
  • Le Sun
  • Weijian Xie
  • Jin Xu

Procedural text understanding requires machines to reason about entity states within the dynamical narratives. Current procedural text understanding approaches are commonly entity-wise, which separately track each entity and independently predict different states of each entity. Such an entity-wise paradigm does not consider the interaction between entities and their states. In this paper, we propose a new scene-wise paradigm for procedural text understanding, which jointly tracks states of all entities in a scene-by-scene manner. Based on this paradigm, we propose Scene Graph Reasoner (SGR), which introduces a series of dynamically evolving scene graphs to jointly formulate the evolution of entities, states and their associations throughout the narrative. In this way, the deep interactions between all entities and states can be jointly captured and simultaneously derived from scene graphs. Experiments show that SGR not only achieves the new state-of-the-art performance but also significantly accelerates the speed of reasoning.

PDF Details

IJCAI Conference 2022 Conference Paper

Towards Robust Dense Retrieval via Local Ranking Alignment

  • Xuanang Chen
  • Jian Luo
  • Ben He
  • Le Sun
  • Yingfei Sun

Dense retrieval (DR) has extended the employment of pre-trained language models, like BERT, for text ranking. However, recent studies have raised the robustness issue of DR model against query variations, like query with typos, along with non-trivial performance losses. Herein, we argue that it would be beneficial to allow the DR model to learn to align the relative positions of query-passage pairs in the representation space, as query variations cause the query vector to drift away from its original position, affecting the subsequent DR effectiveness. To this end, we propose RoDR, a novel robust DR model that learns to calibrate the in-batch local ranking of query variation to that of original query for the DR space alignment. Extensive experiments on MS MARCO and ANTIQUE datasets show that RoDR significantly improves the retrieval results on both the original queries and different types of query variations. Meanwhile, RoDR provides a general query noise-tolerate learning framework that boosts the robustness and effectiveness of various existing DR models. Our code and models are openly available at https: //github. com/cxa-unique/RoDR.

PDF Details DOI

AAAI Conference 2021 Conference Paper

Benchmarking Knowledge-Enhanced Commonsense Question Answering via Knowledge-to-Text Transformation

  • Ning Bian
  • Xianpei Han
  • Bo Chen
  • Le Sun

A fundamental ability of humans is to utilize commonsense knowledge in language understanding and question answering. In recent years, many knowledge-enhanced Commonsense Question Answering (CQA) approaches have been proposed. However, it remains unclear: (1) How far can we get by exploiting external knowledge for CQA? (2) How much potential of knowledge has been exploited in current CQA models? (3) Which are the most promising directions for future CQA? To answer these questions, we benchmark knowledge-enhanced CQA by conducting extensive experiments on multiple standard CQA datasets using a simple and effective knowledgeto-text transformation framework. Experiments show that: (1) Our knowledge-to-text framework is effective and achieves state-of-the-art performance on CommonsenseQA dataset, providing a simple and strong knowledge-enhanced baseline for CQA; (2) The potential of knowledge is still far from being fully exploited in CQA — there is a significant performance gap from current models to our models with golden knowledge; and (3) Context-sensitive knowledge selection, heterogeneous knowledge exploitation, and commonsense-rich language models are promising CQA directions.

PDF Details

AAAI Conference 2021 Conference Paper

Denoising Distantly Supervised Named Entity Recognition via a Hypergeometric Probabilistic Model

  • Wenkai Zhang
  • Hongyu Lin
  • Xianpei Han
  • Le Sun
  • Huidan Liu
  • Zhicheng Wei
  • Nicholas Yuan

Denoising is the essential step for distant supervision based named entity recognition. Previous denoising methods are mostly based on instance-level confidence statistics, which ignore the variety of the underlying noise distribution on different datasets and entity types. This makes them difficult to be adapted to high noise rate settings. In this paper, we propose Hypergeometric Learning (HGL), a denoising algorithm for distantly supervised NER that takes both noise distribution and instance-level confidence into consideration. Specifically, during neural network training, we naturally model the noise samples in each batch following a hypergeometric distribution parameterized by the noise-rate. Then each instance in the batch is regarded as either correct or noisy one according to its label confidence derived from previous training step, as well as the noise distribution in this sampled batch. Experiments show that HGL can effectively denoise the weaklylabeled data retrieved from distant supervision, and therefore results in significant improvements on the trained models.

PDF Details

AIJ Journal 2021 Journal Article

Enhanced aspect-based sentiment analysis models with progressive self-supervised attention learning

  • Jinsong Su
  • Jialong Tang
  • Hui Jiang
  • Ziyao Lu
  • Yubin Ge
  • Linfeng Song
  • Deyi Xiong
  • Le Sun

In aspect-based sentiment analysis (ABSA), many neural models are equipped with an attention mechanism to quantify the contribution of each context word to sentiment prediction. However, such a mechanism suffers from one drawback: only a few frequent words with sentiment polarities are tended to be taken into consideration for final sentiment decision while abundant infrequent sentiment words are ignored by models. To deal with this issue, we propose a progressive self-supervised attention learning approach for attentional ABSA models. In this approach, we iteratively perform sentiment prediction on all training instances, and continually learn useful attention supervision information in the meantime. During training, at each iteration, context words with the highest impact on sentiment prediction, identified based on their attention weights or gradients, are extracted as words with active/misleading influence on the correct/incorrect prediction for each instance. Words extracted in this way are masked for subsequent iterations. To exploit these extracted words for refining ABSA models, we augment the conventional training objective with a regularization term that encourages ABSA models to not only take full advantage of the extracted active context words but also decrease the weights of those misleading words. We integrate the proposed approach into three state-of-the-art neural ABSA models. Experiment results and in-depth analyses show that our approach yields better attention results and significantly enhances the performance of all three models. We release the source code and trained models at https: //github. com/DeepLearnXMU/PSSAttention.

Details DOI

AAAI Conference 2020 Conference Paper

End-to-End Bootstrapping Neural Network for Entity Set Expansion

  • Lingyong Yan
  • Xianpei Han
  • Ben He
  • Le Sun

Bootstrapping for entity set expansion (ESE) has long been modeled as a multi-step pipelined process. Such a paradigm, unfortunately, often suffers from two main challenges: 1) the entities are expanded in multiple separate steps, which tends to introduce noisy entities and results in the semantic drift problem; 2) it is hard to exploit the high-order entity-pattern relations for entity set expansion. In this paper, we propose an end-to-end bootstrapping neural network for entity set expansion, named BootstrapNet, which models the bootstrapping in an encoder-decoder architecture. In the encoding stage, a graph attention network is used to capture both the first- and the high-order relations between entities and patterns, and encode useful information into their representations. In the decoding stage, the entities are sequentially expanded through a recurrent neural network, which outputs entities at each stage, and its hidden state vectors, representing the target category, are updated at each expansion step. Experimental results demonstrate substantial improvement of our model over previous ESE approaches.

PDF Details

IJCAI Conference 2020 Conference Paper

Global Structure and Local Semantics-Preserved Embeddings for Entity Alignment

  • Hao Nie
  • Xianpei Han
  • Le Sun
  • Chi Man Wong
  • Qiang Chen
  • Suhui Wu
  • Wei Zhang

Entity alignment (EA) aims to identify entities located in different knowledge graphs (KGs) that refer to the same real-world object. To learn the entity representations, most EA approaches rely on either translation-based methods which capture the local relation semantics of entities or graph convolutional networks (GCNs), which exploit the global KG structure. Afterward, the aligned entities are identified based on their distances. In this paper, we propose to jointly leverage the global KG structure and entity-specific relational triples for better entity alignment. Specifically, a global structure and local semantics preserving network is proposed to learn entity representations in a coarse-to-fine manner. Experiments on several real-world datasets show that our method significantly outperforms other entity alignment approaches and achieves the new state-of-the-art performance.

PDF Details DOI

IJCAI Conference 2020 Conference Paper

Hierarchical Matching Network for Heterogeneous Entity Resolution

  • Cheng Fu
  • Xianpei Han
  • Jiaming He
  • Le Sun

Entity resolution (ER) aims to identify data records referring to the same real-world entity. Most existing ER approaches rely on the assumption that the entity records to be resolved are homogeneous, i. e. , their attributes are aligned. Unfortunately, entities in real-world datasets are often heterogeneous, usually coming from different sources and being represented using different attributes. Furthermore, the entities’ attribute values may be redundant, noisy, missing, misplaced, or misspelled—we refer to it as the dirty data problem. To resolve the above problems, this paper proposes an end-to-end hierarchical matching network (HierMatcher) for entity resolution, which can jointly match entities in three levels—token, attribute, and entity. At the token level, a cross-attribute token alignment and comparison layer is designed to adaptively compare heterogeneous entities. At the attribute level, an attribute-aware attention mechanism is proposed to denoise dirty attribute values. Finally, the entity level matching layer effectively aggregates all matching evidence for the final ER decisions. Experimental results show that our method significantly outperforms previous ER methods on homogeneous, heterogeneous and dirty datasets.

PDF Details DOI

AAAI Conference 2020 Conference Paper

Learning to Map Frequent Phrases to Sub-Structures of Meaning Representation for Neural Semantic Parsing

  • Bo Chen
  • Xianpei Han
  • Ben He
  • Le Sun

Neural semantic parsers usually generate meaning representation tokens from natural language tokens via an encoderdecoder model. However, there is often a vocabularymismatch problem between natural language utterances and logical forms. That is, one word maps to several atomic logical tokens, which need to be handled as a whole, rather than individual logical tokens at multiple steps. In this paper, we propose that the vocabulary-mismatch problem can be effectively resolved by leveraging appropriate logical tokens. Specifically, we exploit macro actions, which are of the same granularity of words/phrases, and allow the model to learn mappings from frequent phrases to corresponding substructures of meaning representation. Furthermore, macro actions are compact, and therefore utilizing them can significantly reduce the search space, which brings a great benefit to weakly supervised semantic parsing. Experiments show that our method leads to substantial performance improvement on three benchmarks, in both supervised and weakly supervised settings.

PDF Details

IJCAI Conference 2019 Conference Paper

End-to-End Multi-Perspective Matching for Entity Resolution

  • Cheng Fu
  • Xianpei Han
  • Le Sun
  • Bo Chen
  • Wei Zhang
  • Suhui Wu
  • Hao Kong

Entity resolution (ER) aims to identify data records referring to the same real-world entity. Due to the heterogeneity of entity attributes and the diversity of similarity measures, one main challenge of ER is how to select appropriate similarity measures for different attributes. Previous ER methods usually employ heuristic similarity selection algorithms, which are highly specialized to specific ER problems and are hard to be generalized to other situations. Furthermore, previous studies usually perform similarity learning and similarity selection independently, which often result in error propagation and are hard to be optimized globally. To resolve the above problems, this paper proposes an end-to-end multi-perspective entity matching model, which can adaptively select optimal similarity measures for heterogenous attributes by jointly learning and selecting similarity measures in an end-to-end way. Experiments on two real-world datasets show that our method significantly outperforms previous ER methods.

PDF Details

AAAI Conference 2017 Conference Paper

Distant Supervision via Prototype-Based Global Representation Learning

  • Xianpei Han
  • Le Sun

Distant supervision (DS) is a promising technique for relation extraction. Currently, most DS approaches build relation extraction models in local instance feature space, often suffer from the multi-instance problem and the missing label problem. In this paper, we propose a new DS method Ñ prototype-based global representation learning, which can effectively resolve the multi-instance problem and the missing label problem by learning informative entity pair representations, and building discriminative extraction models at the entity pair level, rather than at the instance level. Specifically, we propose a prototype-based embedding algorithm, which can embed entity pairs into a prototype-based global feature space; we then propose a neural network model, which can classify entity pairs into target relation types by summarizing relevant information from multiple instances. Experimental results show that our method can achieve significant performance improvement over traditional DS methods.

PDF Details

AAAI Conference 2016 Conference Paper

A Joint Model for Entity Set Expansion and Attribute Extraction from Web Search Queries

  • Zhenzhong Zhang
  • Le Sun
  • Xianpei Han

Entity Set Expansion (ESE) and Attribute Extraction (AE) are usually treated as two separate tasks in Information Extraction (IE). However, the two tasks are tightly coupled, and each task can benefit significantly from the other by leveraging the inherent relationship between entities and attributes. That is, 1) an attribute is important if it is shared by many typical entities of a class; 2) an entity is typical if it owns many important attributes of a class. Based on this observation, we propose a joint model for ESE and AE, which models the inherent relationship between entities and attributes as a graph. Then a graph reinforcement algorithm is proposed to jointly mine entities and attributes of a specific class. Experimental results demonstrate the superiority of our method for discovering both new entities and new attributes.

PDF Details

AAAI Conference 2016 Conference Paper

Global Distant Supervision for Relation Extraction

  • Xianpei Han
  • Le Sun

Machine learning approaches to relation extraction are typically supervised and require expensive labeled data. To break the bottleneck of labeled data, a promising approach is to exploit easily obtained indirect supervision knowledge – which we usually refer to as distant supervision (DS). However, traditional DS methods mostly only exploit one specific kind of indirect supervision knowledge – the relations/facts in a given knowledge base, thus often suffer from the problem of lack of supervision. In this paper, we propose a global distant supervision model for relation extraction, which can: 1) compensate the lack of supervision with a wide variety of indirect supervision knowledge; and 2) reduce the uncertainty in DS by performing joint inference across relation instances. Experimental results show that, by exploiting the consistency between relation labels, the consistency between relations and arguments, and the consistency between neighbor instances using Markov logic, our method significantly outperforms traditional DS approaches.

PDF Details
v2026.09.13