Author name cluster

Chao Deng

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.

12 papers

2 author rows

EAAI Journal 2026 Journal Article

A method for extracting emotion–cause pairs based on bidirectional machine reading comprehension

Guorui Li
Yaxin Wen
Cong Wang
Lihong Cao
Sancheng Peng
Chao Deng

Emotion–cause pair extraction (ECPE) is an emerging task aimed at extracting emotional clauses and their corresponding causal clauses from unlabeled documents, which possesses wide application potentials in areas such as e-commerce, medical health, public opinion mining, and empathetic companion. The challenges of inadequate emotion-leading extraction mechanism and deficient inter-clause relationship modeling ability still limit its performance. In this study, we present a method to extract potential emotion–cause pairs based on bidirectional machine reading comprehension. It transforms the ECPE task into a three-round machine reading comprehension task in both emotion–cause and cause–emotion directions, which effectively improves the coverage ratio of valid emotion–cause pairs. Furthermore, the graph attention network version 2 and the emotion filter are also integrated into the proposed method to model and identify the complicated relationship between emotional clauses and causal clauses. Extensive experiments were conducted on both Chinese and English ECPE benchmark datasets to verify the effectiveness of the proposed method. The experimental results demonstrate that our proposed method achieves optimal performance, outperforming the state-of-the-art method by 2. 28%, 4. 67%, 3. 25% and 3. 64%, 9. 16%, 5. 86% for precision, recall, and F1 score on two benchmark datasets, respectively. The performance analysis further confirms the effectiveness of its core components, including the bidirectional ECPE mechanism, the dynamic graph attention network, and the emotion filter module.

Details DOI

AAAI Conference 2026 Conference Paper

Self-Correction Distillation for Structured Data Question Answering

Yushan Zhu
Wen Zhang
Long Jin
Mengshu Sun
Ling Zhong
Zhiqiang Liu
Juan Li
Lei Liang

Structured data question answering (QA), including table QA, Knowledge Graph (KG) QA, and temporal KG QA, is a pivotal research area. Advances in large language models (LLMs) have driven significant progress in unified structural QA frameworks like TrustUQA. However, these frameworks face challenges when applied to small-scale LLMs since small-scale LLMs are prone to errors in generating structured queries. To improve the structured data QA ability of small-scale LLMs, we propose a self-correction distillation (SCD) method. In SCD, an error prompt mechanism (EPM) is designed to detect errors and provide customized error messages during inference, and a two-stage distillation strategy is designed to transfer large-scale LLMs' query-generation and error-correction capabilities to small-scale LLM. Experiments across 5 benchmarks with 3 structured data types demonstrate that our SCD achieves the best performance and superior generalization on small-scale LLM (8B) compared to other distillation methods, and closely approaches the performance of GPT4 on some datasets. Furthermore, large-scale LLMs equipped with EPM surpass the state-of-the-art results on most datasets.

PDF Details DOI

AAAI Conference 2025 Conference Paper

FatesGS: Fast and Accurate Sparse-View Surface Reconstruction Using Gaussian Splatting with Depth-Feature Consistency

Han Huang
Yulun Wu
Chao Deng
Ge Gao
Ming Gu
Yu-Shen Liu

Recently, Gaussian Splatting has sparked a new trend in the field of computer vision. Apart from novel view synthesis, it has also been extended to the area of multi-view reconstruction. The latest methods facilitate complete, detailed surface reconstruction while ensuring fast training speed. However, these methods still require dense input views, and their output quality significantly degrades with sparse views. We observed that the Gaussian primitives tend to overfit the few training views, leading to noisy floaters and incomplete reconstruction surfaces. In this paper, we present an innovative sparse-view reconstruction framework that leverages intra-view depth and multi-view feature consistency to achieve remarkably accurate surface reconstruction. Specifically, we utilize monocular depth ranking information to supervise the consistency of depth distribution within patches and employ a smoothness loss to enhance the continuity of the distribution. To achieve finer surface reconstruction, we optimize the absolute position of depth through multi-view projection features. Extensive experiments on DTU and BlendedMVS demonstrate that our method outperforms state-of-the-art methods with a speedup of 60x to 200x, achieving swift and fine-grained mesh reconstruction without the need for costly pre-training.

PDF Details DOI

AAAI Conference 2025 Conference Paper

Improving Cancer Gene Prediction by Enhancing Common Information Between the PPI Network and Gene Functional Association

Chao Deng
Hongdong Li
Jianxin Wang

Identifying cancer genes is crucial for treatment and understanding pathogenesis. Recent methods typically leverage protein-protein interaction (PPI) networks or gene functional association data from annotated gene sets. There may be some shared neighborhood structure information between these two types of gene association data. While this common information may contain more accurate gene association information, existing methods often overlook this potential. To address this gap, we introduce DISFusion, which integrates multi-omics cancer data, PPI networks, and gene functional associations to identify cancer genes. A key innovation of DISFusion is the cross-view decorrelation loss, which enhances the common information between PPI networks and gene functional associations, thereby improving prediction accuracy. Extensive experiments indicate that DISFusion outperforms state-of-the-art methods and exhibits greater generalization ability. Moreover, analysis of CPTAC pan-cancer proteomic data highlights significant associations between the 30 novel cancer genes predicted by DISFusion and multiple cancer types, underscoring its practical utility. These findings validate the effectiveness of enhancing common information and provide new insights into cancer gene identification.

PDF Details DOI

ICLR Conference 2025 Conference Paper

LOIRE: LifelOng learning on Incremental data via pre-trained language model gRowth Efficiently

Xue Han 0018
Yitong Wang
Junlan Feng
Wenchun Gao
Qian Hu
Chao Deng

Large-scale pre-trained language models (PLMs) require significant computational resources to train from scratch on large volumes of data. But in the real world, emerging data from diverse sources may not be initially available for pre-training. Recent studies on lifelong learning have tried to solve this problem by exploring the use of model growth techniques to effectively incorporate new knowledge without the need for complete re-training. However, model growth approaches utilized have issues with growth operators that do not ensure strict function preservation or growth schedules that only include a few growth dimensions, reducing lifelong learning's effect. Furthermore, existing approaches often assume that emerging data has the same distribution as pre-training data, causing catastrophic forgetting of previously acquired knowledge. To address the aforementioned issues, we introduce LOIRE, a framework for lifelong learning that enables PLMs to effectively grow their capacity using incremental data. LOIRE employs growth operators for all feasible dimensions and a growth schedule to generate the optimal expansion sequence in the field of lifelong learning. Specifically, we present a novel plug-in layer growth operator with residual connections that skip the newly added layer during initial training while ensuring function preservation. We additionally propose an iterative distillation strategy for LOIRE that allows an intermediate model in the growth stages to switch between being a student and a teacher, reducing catastrophic forgetting during growth. Experiments show that LOIRE can reduce computational expenses by an average of 29.22\% while retaining equivalent or better downstream performance.

Details

AAMAS Conference 2025 Conference Paper

MacLight: Multi-scene Aggregation Convolutional Learning for Traffic Signal Control

Sunbowen Lee
Hongqin Lyu
Yicheng Gong
Yingying Sun
Chao Deng

Reinforcement learning methods have proposed promising traffic signal control policy that can be trained on large road networks. Current SOTA methods model road networks as topological graph structures, incorporate graph attention into deep Q-learning, and merge local and global embeddings to improve policy. However, graph-based methods are difficult to parallelize, resulting in huge time overhead. Moreover, none of the current peer studies have deployed dynamic traffic systems for experiments, which is far from the actual situation. In this context, we propose Multi-Scene Aggregation Convolutional Learning for traffic signal control (MacLight), which offers faster training speeds and more stable performance. Our approach consists of two main components. The first is the global representation, where we utilize variational autoencoders to compactly compress and extract the global representation. The second component employs the proximal policy optimization algorithm as the backbone, allowing value evaluation to consider both local features and global embedding representations. This backbone model significantly reduces time overhead and ensures stability in policy updates. We validated our method across multiple traffic scenarios under both static and dynamic traffic systems. Experimental results demonstrate that, compared to general and domian SOTA ∗Contribution equal to the first author †Corresponding author This work is licensed under a Creative Commons Attribution International 4. 0 License. Proc. of the 24th International Conference on Autonomous Agents and Multiagent Systems (AAMAS 2025), Y. Vorobeychik, S. Das, A. Nowé (eds.), May 19 – 23, 2025, Detroit, Michigan, USA. © 2025 International Foundation for Autonomous Agents and Multiagent Systems (www. ifaamas. org). methods, our approach achieves superior stability, optimized convergence levels and the highest time efficiency. The code is under https: //github. com/Aegis1863/MacLight.

PDF

AAAI Conference 2025 Conference Paper

MoE-LPR: Multilingual Extension of Large Language Models Through Mixture-of-Experts with Language Priors Routing

Hao Zhou
Zhijun Wang
Shujian Huang
Xin Huang
Xue Han
Junlan Feng
Chao Deng
Weihua Luo

Large Language Models (LLMs) are often English-centric due to the disproportionate distribution of languages in their pre-training data. Enhancing non-English language capabilities through post-pretraining often results in catastrophic forgetting of high-resource languages. Previous methods either achieve good expansion with severe forgetting or slight forgetting with poor expansion, indicating the challenge of balancing language expansion while preventing forgetting. In this paper, we propose a method called MoE-LPR (Mixture-of-Experts with Language Priors Routing) to alleviate this problem. MoE-LPR employs a two-stage training approach to enhance the multilingual capability. First, the model is post-pretrained into a Mixture-of-Experts(MoE) architecture by upcycling, where all the original parameters are frozen and new experts are added. In this stage, we focus improving the ability on expanded languages, without using any original language data. Then, the model reviews the knowledge of the original languages with replay data amounting to less than 1% of post-pretraining, where we incorporate language priors routing to better recover the abilities of the original languages. Evaluations on multiple benchmarks show that MoE-LPR outperforms other post-pretraining methods. Freezing original parameters preserves original language knowledge while adding new experts preserves the learning ability. Reviewing with LPR enables effective utilization of multilingual knowledge within the parameters. Additionally, the MoE architecture maintains the same inference overhead while increasing total model parameters. Extensive experiments demonstrate MoE-LPR’s effectiveness in improving expanded languages and preserving original language proficiency with superior scalability.

PDF Details DOI

AAAI Conference 2025 Conference Paper

Sparis: Neural Implicit Surface Reconstruction of Indoor Scenes from Sparse Views

Yulun Wu
Han Huang
Wenyuan Zhang
Chao Deng
Ge Gao
Ming Gu
Yu-Shen Liu

In recent years, reconstructing indoor scene geometry from multi-view images has achieved encouraging accomplishments. Current methods incorporate monocular priors into neural implicit surface models to achieve high-quality reconstructions. However, these methods require hundreds of images for scene reconstruction. When only a limited number of views are available as input, the performance of monocular priors deteriorates due to scale ambiguity, leading to the collapse of the reconstructed scene geometry. In this paper, we propose a new method, named Sparis, for indoor surface reconstruction from sparse views. Specifically, we investigate the impact of monocular priors on sparse scene reconstruction, introducing a novel prior based on inter-image matching information. Our prior offers more accurate depth information while ensuring cross-view matching consistency. Additionally, we employ an angular filter strategy and an epipolar matching weight function, aiming to reduce errors due to view matching inaccuracies, thereby refining the inter-image prior for improved reconstruction accuracy. The experiments conducted on widely used benchmarks demonstrate superior performance in sparse-view scene reconstruction.

PDF Details DOI

EAAI Journal 2025 Journal Article

Temporal-constrained parallel graph neural networks for recognizing motion patterns and gait phases in class-imbalanced scenarios

Changhe Zhang
Zidong Yu
Xiaoyun Wang
Ze-Jian Chen
Chao Deng

Exoskeletons have emerged as a promising technology in the field of motor rehabilitation, particularly for individuals with lower limb motor dysfunction. Related research on human motion intention recognition based on wearable sensors and deep learning is garnering increasing attention from scholars. This study proposes a novel two-stage framework that first performs Motion Pattern Recognition (MPR) and then achieves Gait Phase Recognition (GPR) using surface electromyography and kinematic data from subjects' lower limbs. In view of the short-range dependency of convolutional neural networks, this study exploits Graph Neural Networks (GNNs) to adaptively fuse multimodal signals, and designs several graph generation methods to tackle the issue of lacking original topology in multimodal signals. On this basis, a Parallel GNNs (PGNNs) architecture is constructed to fuse graph-level and domain knowledge features, and a dynamically decaying weighted cross-entropy loss is presented to enhance the recognition performance of PGNNs in class-imbalanced scenarios. Additionally, a categorical hidden Markov model is established using the prior temporal constraints among subject's gait phases to reduce the “phase mutation” errors in PGNNs' predictions without additional training costs. The effectiveness of the proposed framework is verified on a benchmark dataset containing five common lower limb motions of ten healthy subjects. It achieves an average accuracy of 99. 97% in the MPR task, and average GPR accuracies of 94. 06%, 95. 27%, and 95. 43% in level walking, upstairs and downstairs motions, respectively, outperforming existing state-of-the-art methods. Experimental results indicate potential support for the controller design of human-robot highly-coupled rehabilitation exoskeletons.

Details DOI

EAAI Journal 2024 Journal Article

Interpretable Dual-branch EMGNet: A transfer learning-based network for inter-subject lower limb motion intention recognition

Changhe Zhang
Xiaoyun Wang
Zidong Yu
Bingjin Wang
Chao Deng

Currently, the fusion of surface Electromyography (EMG) and deep learning is gradually showing immense potential in the research of Lower Limb Motion Intention Recognition (LLMIR). Nevertheless, most deep learning algorithms have poor interpretability without special design or the help of other post-hoc analysis tools, as well as unsatisfactory performance in cross-subject prediction. Hence, this paper presents a novel Interpretable Dual-Branch EMG Network (IDB-EMGNet), in which one branch is dedicated to lower limb motion recognition, and the other is able to predict knee joint angles in advance. The shallow feature extraction module of IDB-EMGNet is constructed using an ante-hoc interpretable SincNet technique, which enables the detection of the spectral range of EMG used for the LLMIR task. An improved bottleneck block with shuffle attention is designed for deep feature extraction, which enhances model performance with only a little increase in complexity. The performance of IDB-EMGNet in both intra-subject and inter-subject scenarios is investigated, where the latter integrates the transfer learning technique. Specifically, by conducting model pre-training on source-domain subjects and transferring the learned knowledge to target-domain subjects, satisfactory performance can be achieved even with less computing resource. Experimental results on two publicly available datasets indicate that the proposed approach exhibits superior applicability to both normal subjects and knee-pathology patients, showing a promising prospect in the controller design of human-robot collaborative exoskeletons.

Details DOI

TIST Journal 2024 Journal Article

KGDA: A Knowledge Graph Driven Decomposition Approach for Cellular Traffic Prediction

Jiahui Gong
Tong Li
Huandong Wang
Yu Liu
Xing Wang
Zhendong Wang
Chao Deng
Junlan Feng

Understanding and accurately predicting cellular traffic data is vital for communication operators and device users, as it facilitates efficient resource allocation and ensures superior service quality. However, large-scale cellular traffic data forecasting remains challenging due to intricate temporal variations and complex spatial relationships. This article proposes a Knowledge Graph Driven Decomposition Approach (KGDA) for precise cellular traffic prediction. The KGDA breaks down the impact of static environmental factors and dynamic autocorrelations of cellular traffic time series, enabling the capture of overall traffic changes and understanding of traffic dependence on past values. Specifically, we propose an urban knowledge graph to capture the static environmental context of base stations, mapping these entities into the same latent space while retaining static environmental knowledge. The cellular traffic is divided into a regular pattern and fluctuating residual components, with the KGDA comprising four modules: a Knowledge Graph Representation Learning model, a traffic regular pattern prediction module, a traffic residual dynamic prediction module, and an attentional fusion module. The first leverages graph neural networks to extract spatial contexts and predict regular patterns, the second utilizes the Bi-directional Long Short-Term Memory (Bi-LSTM) model to capture autocorrelations of traffic time series, and the final module integrates the patterns and residuals to produce the final prediction result. Comprehensive experiments demonstrate that our proposed model outperforms state-of-the-art models by more than 10% in forecasting cellular traffic.

Details DOI

AAAI Conference 2017 Conference Paper

Variational Autoencoder for Semi-Supervised Text Classification

Weidi Xu
Haoze Sun
Chao Deng
Ying Tan

Although semi-supervised variational autoencoder (SemiVAE) works in image classiﬁcation task, it fails in text classiﬁcation task if using vanilla LSTM as its decoder. From a perspective of reinforcement learning, it is veriﬁed that the decoder’s capability to distinguish between different categorical labels is essential. Therefore, Semi-supervised Sequential Variational Autoencoder (SSVAE) is proposed, which increases the capability by feeding label into its decoder RNN at each time-step. Two speciﬁc decoder structures are investigated and both of them are veriﬁed to be effective. Besides, in order to reduce the computational complexity in training, a novel optimization method is proposed, which estimates the gradient of the unlabeled objective function by sampling, along with two variance reduction techniques. Experimental results on Large Movie Review Dataset (IMDB) and AG’s News corpus show that the proposed approach signiﬁcantly improves the classiﬁcation accuracy compared with pure-supervised classiﬁers, and achieves competitive performance against previous advanced methods. State-of-the-art results can be obtained by integrating other pretraining-based methods.

PDF Details