Associate Professor at Shenzhen University, China
Dr. Junyang Chen is a tenured Associate Professor at the College of Computer Science and Software Engineering, Shenzhen University, China. His research bridges multimodal data mining, spatiotemporal modeling, and robust recommendation systems. Dr. Chen has developed influential frameworks for multimedia recommendation and user intent modeling, with notable deployments in industry such as Tencent QQ Browser. He is a Senior Member of IEEE/CCF and actively contributes to top-tier conferences and journals in artificial intelligence and multimedia systems.
🔹Professional Profile:
Scopus Profile
Google Scholar Profile
🎓Education Background
💼 Professional Development
-
Associate Professor (Tenured), Shenzhen University, Jan 2025 – Present
-
Assistant Professor, Shenzhen University, Mar 2021 – Dec 2024
At Shenzhen University, Dr. Chen leads research in multimodal data mining, user behavior analysis, and spatiotemporal modeling. He has collaborated with leading industry players, including Tencent, where his algorithms significantly improved engagement metrics such as CTR and recommendation accuracy.
🔬Research Focus
-
Multimodal Data Mining
-
Robust Recommendation Systems in Open Environments
-
Spatiotemporal User Modeling
-
Few-shot and Zero-shot Learning
-
Graph Representation Learning
-
Social Media Behavior Analysis
📈Author Metrics:
🏆Awards and Honors:
-
2024 Shenzhen AI Natural Science Award (First Recipient)
-
2024 Tencent “Rhinoceros Bird” Young Faculty Research Fund
-
2020 National Postdoctoral Forum Best Paper Award (1st Class)
-
Principal Investigator of 1 NSFC project and 3 provincial/municipal projects
-
Editorial and Leadership Roles:
-
Guest Editor, IEEE Transactions on Computational Social Systems
-
Executive Committee Member, CCF Multimedia & Collaborative Computing
-
PC Member: AAAI, ACL, ACM MM, NeurIPS, and more
📝Publication Top Notes
1. LUSTER: Link Prediction Utilizing Shared-Latent Space Representation in Multi-Layer Networks
-
Authors: Ruohan Yang, Muhammad Asif Ali, Huan Wang, Junyang Chen, Di Wang
-
Conference: The Web Conference (WWW), 2025
-
Summary: This paper introduces LUSTER, an end-to-end framework designed for link prediction in multi-layer networks. It addresses the limitations of existing methods by leveraging a shared-latent space representation to capture inter-layer dependencies. The framework comprises four modules: representation extractor, latent space learner, complementary enhancer, and link predictor. Experimental results demonstrate that LUSTER outperforms state-of-the-art methods, achieving up to a 15.87% improvement in AUC metrics. OpenReview
2. Deconfound Semantic Shift and Incompleteness in Incremental Few-shot Semantic Segmentation
-
Authors: Yirui Wu, Yuhang Xia, Hao Li, Lixin Yuan, Junyang Chen, Jun Liu, Tong Lu, Shaohua Wan
-
Conference: Proceedings of the AAAI Conference on Artificial Intelligence, 2025
-
Summary: This study tackles the challenges of semantic shift and data incompleteness in incremental few-shot semantic segmentation. The authors propose a dual-stream incremental learning framework that disentangles semantic shifts and addresses data incompleteness, enhancing the model’s adaptability to new classes with limited data.
3. Reliable Service Recommendation: A Multi-modal Adversarial Method for Personalized Recommendation under Uncertain Missing Modalities
-
Authors: Junyang Chen, Ruohan Yang, Jingcai Guo, Huan Wang, Kaishun Wu, Liangjie Zhang
-
Journal: IEEE Transactions on Services Computing, 2025
-
Summary: The paper presents a robust adversarial learning framework for personalized service recommendation, particularly effective when dealing with uncertain or missing data modalities. The proposed method enhances recommendation reliability by adapting to incomplete multi-modal data scenarios.
4. Adaptive Density Estimation for Personalized Recommendations Across Varied User Activity Levels
-
Authors: Weiwen Liu, Hao Zhu, Jun Yu, Liang Zheng, Jun Yin, Ruijia Liang, Xiaoyu Liu, Junyang Chen, Victor C.M. Leung
-
Journal: IEEE Transactions on Computational Social Systems, 2025
-
Summary: This research introduces a user-activity-aware density estimation technique aimed at improving personalized recommendation systems. By accounting for varying levels of user activity, the proposed method enhances recommendation performance for both highly active and less active users.
5. Unveiling User Interests: A Deep User Interest Exploration Network for Sequential Location Recommendation
-
Authors: Junyang Chen, Jingcai Guo, Qin Zhang, Kaishun Wu, Liangjie Zhang, Victor C.M. Recommendation systemLeung, Huan Wang, Zhiguo Gong
-
Journal: Information Sciences, Volume 689, Article 121416, 2025
-
Summary: The authors propose the Deep User Interest Exploration Network (DUIEN) to enhance sequential location recommendation systems. DUIEN employs a bidirectional neural network with a Cloze objective to predict masked points of interest (POIs) within user trajectories and integrates a user interest exploration layer that utilizes POI tag information. This approach effectively captures the evolving interests of users, leading to improved recommendation accuracy
Conclusion:
Assoc. Prof. Dr. Junyang Chen exemplifies the ideal candidate for the Research for Young Scientist Award through a combination of:
-
Cutting-edge and practical research in Recommendation Systems
-
Demonstrated real-world impact via industry deployments
-
Recognized leadership in high-impact publishing and community service
-
Strategic research directions that align with future challenges in AI personalization, multimodal learning, and robust modeling
Given his academic maturity, research depth, early-career accomplishments, and societal relevance of his work, Dr. Chen is a highly deserving nominee for this award. His work is not only forward-looking but also sets foundational pillars for next-generation intelligent systems.