Yulong Li
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Yulong Li has a total of 20 co-authors. Here are the top co-authors by number of shared publications:
- Feilong Tang
Monash University, Clayton, Australia
3 shared publications - Haochen Xue
Xi'an Jiaotong-Liverpool University, Suzhou, China
3 shared publications - Imran Razzak
Mohamed bin Zayed University of Artificial Intelligence, Abu Dhabi, United Arab Emirates
3 shared publications - Jionglong Su
Xi'an Jiaotong-Liverpool University
3 shared publications - Mian Zhou
Xi'an Jiaotong-Liverpool University, Suzhou, Jiangsu, CHINA
3 shared publications - Yuxuan Zhang
111 Ren’ai Road, Xi'an Jiaotong-Liverpool University, Suzhou, China
3 shared publications
Latest Publications
- Open Access
- 20 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 2723-2734
CNText2Sign and CNSign: Unified Chinese Sign Language Datasets for Bidirectional Accessibility
Sign language is the primary communication mode for 72 million hearing-impaired individuals worldwide, necessitating effective bidirectional Sign Language Production and Sign Language Translation systems. However, functional bidirectional systems require a unified linguistic environment, hindered by the lack of suitable unified datasets, particularly those providing the necessary pose information for accurate Sign Language Production (SLP) evaluation. Concurrently, current SLP evaluation methods like back-translation ignore pose accuracy, and high-quality coordinated generation remains challenging. To create this crucial environment and overcome these challenges, we introduce CNText2Sign and CNSign, which together constitute the first unified dataset aimed at supporting bidirectional accessibility systems for Chinese sign language; CNText2Sign provides 15,000 natural language-to-sign mappings and standardized skeletal keypoints for 8,643 vocabulary items supporting pose assessment. Building upon this foundation, we propose the AuraLLM model, which leverages a decoupled architecture with CNText2Sign's pose data for novel direct gesture accuracy assessment. The model employs retrieval augmentation and Cascading Vocabulary Resolution to handle semantic mapping and out-of-vocabulary words, and achieves all-scenario production with controllable coordination of gestures and facial expressions via pose-conditioned video synthesis. Concurrently, our Sign Language Translation model SignMST-C employs targeted self-supervised pretraining for dynamic feature capture, achieving new SOTA results on PHOENIX2014-T with BLEU-4 scores up to 32.08. AuraLLM establishes a strong performance baseline on CNText2Sign with a BLEU-4 score of 50.41 under direct evaluation.
- 20 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 2723-2734
Open AccessCNText2Sign and CNSign: Unified Chinese Sign Language Datasets for Bidirectional Accessibility
Sign language is the primary communication mode for 72 million hearing-impaired individuals worldwide, necessitating effective bidirectional Sign Language Production and Sign Language Translation systems. However, functional bidirectional systems require a unified linguistic environment, hindered by the lack of suitable unified datasets, particularly those providing the necessary pose information for accurate Sign Language Production (SLP) evaluation. Concurrently, current SLP evaluation methods like back-translation ignore pose accuracy, and high-quality coordinated generation remains challenging. To create this crucial environment and overcome these challenges, we introduce CNText2Sign and CNSign, which together constitute the first unified dataset aimed at supporting bidirectional accessibility systems for Chinese sign language; CNText2Sign provides 15,000 natural language-to-sign mappings and standardized skeletal keypoints for 8,643 vocabulary items supporting pose assessment. Building upon this foundation, we propose the AuraLLM model, which leverages a decoupled architecture with CNText2Sign's pose data for novel direct gesture accuracy assessment. The model employs retrieval augmentation and Cascading Vocabulary Resolution to handle semantic mapping and out-of-vocabulary words, and achieves all-scenario production with controllable coordination of gestures and facial expressions via pose-conditioned video synthesis. Concurrently, our Sign Language Translation model SignMST-C employs targeted self-supervised pretraining for dynamic feature capture, achieving new SOTA results on PHOENIX2014-T with BLEU-4 scores up to 32.08. AuraLLM establishes a strong performance baseline on CNText2Sign with a BLEU-4 score of 50.41 under direct evaluation.
- Open Access
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 4611-4622
Rhythm of Opinion: Interpretable Hawkes-Graph Networks for Hierarchical Opinion Propagation
Opinion propagation research primarily focuses on phenomenon prediction rather than mechanism understanding, lacking interpretable frameworks to reveal underlying propagation dynamics. This limitation stems from two sources: existing methods employ end-to-end paradigms where parameters lack physical meanings, while available datasets suffer from incomplete hierarchical structures, coarse sentiment annotations, and limited domain coverage. To address these limitations, we introduce VISTA, a multi-dimensional opinion propagation dataset providing complete hierarchical structures, fine-grained emotional annotations, and cross-domain coverage. Based on this dataset, we propose an interpretable modeling framework integrating high-dimensional Hawkes processes with graph neural networks, enabling parametric expression of propagation mechanisms through event space constructed from emotional and reply level combinations. Through interpretable parameter analysis, we reveal three mechanistic patterns: differential emotional propagation strength, asymmetric hierarchical excitation, and temporal memory effects. Our framework establishes quantitative foundations for understanding opinion propagation dynamics, achieving best performance in sentiment prediction and structural consistency tasks while providing the first benchmark for multi-dimensional propagation mechanism analysis.
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 4611-4622
Open AccessRhythm of Opinion: Interpretable Hawkes-Graph Networks for Hierarchical Opinion Propagation
Opinion propagation research primarily focuses on phenomenon prediction rather than mechanism understanding, lacking interpretable frameworks to reveal underlying propagation dynamics. This limitation stems from two sources: existing methods employ end-to-end paradigms where parameters lack physical meanings, while available datasets suffer from incomplete hierarchical structures, coarse sentiment annotations, and limited domain coverage. To address these limitations, we introduce VISTA, a multi-dimensional opinion propagation dataset providing complete hierarchical structures, fine-grained emotional annotations, and cross-domain coverage. Based on this dataset, we propose an interpretable modeling framework integrating high-dimensional Hawkes processes with graph neural networks, enabling parametric expression of propagation mechanisms through event space constructed from emotional and reply level combinations. Through interpretable parameter analysis, we reveal three mechanistic patterns: differential emotional propagation strength, asymmetric hierarchical excitation, and temporal memory effects. Our framework establishes quantitative foundations for understanding opinion propagation dynamics, achieving best performance in sentiment prediction and structural consistency tasks while providing the first benchmark for multi-dimensional propagation mechanism analysis.
- Open Access
- 27 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 12651-12658
Genesis: A Large-Scale Benchmark for Multimodal Large Language Model in Emotional Causality Analysis
- 27 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 12651-12658
Open AccessGenesis: A Large-Scale Benchmark for Multimodal Large Language Model in Emotional Causality Analysis