Short Bio

I am a Ph.D. candidate (2021-) in the School of Computer Science at Beihang University , advised by Associate Prof. Rui Liu and Associate Prof. Yong Chen.

My research interests include machine learning and its applications in computer vision, currently focusing on continual learning and multimodal large language models .

News

[2025-12] One paper about class-incremental learning is submitted to ICME 2026 .
[2025-09] One paper about few-shot class-incremental learning is submitted to IJCV .
[2025-09] One paper about class-incremental learning is submitted to Pattern Recognition .
[2020-01] One paper about 3D track planning is accepted to Mathematical Problems in Engineering .

Publications

Journal Papers

IJCV Paper

Decoupled and Distilled: Task-Adaptive LoRA-Teachers with Ensemble Knowledge Transfer for Few-Shot Class-Incremental Learning

Hongwei Zhao , Rui Liu, Yansong Liu, Zhiyuan Zou, Yong Chen

International Journal of Computer Vision. IJCV (Under Review)

PR Paper

Dynamic LoRA-Experts and Prototype-Ensemble Matching for Class-Incremental Learning

Hongwei Zhao , Rui Liu, Ruoheng Li, Yong Chen

Pattern Recognition. PR (Under Review)

Conference Papers

ICME Paper

Hyperbolic Prototype Routing for Rehearsal-Free Class-Incremental Learning

Hongwei Zhao , Rui Liu, Yong Chen

IEEE International Conference on Multimedia and Expo. ICME 2026 (Under Review)

MPE Paper

On the 3D Track Planning for Electric Power Inspection Based on the Improved Ant Colony Optimization and A* Algorithm

Zheng Huang, Xuefeng Zhai, Hongxing Wang, Hang Zhou, Hongwei Zhao , Mingduan Feng

Mathematical Problems in Engineering, 2020. SCI

Projects

🏛️ National Network Management Platform for Major Scientific Research Infrastructures Website

Ministry of Science and Technology Major Project | Student Lead
Overview: In recent years, China's scientific research facilities and instruments have continued to grow, but issues such as low utilization, redundant construction, and idle waste remain prominent. To improve resource utilization efficiency, the State Council issued guidelines requiring the establishment of an open sharing system and unified network management platform. Currently, the platform integrates 600+ categories and 100,000+ sets of instruments (valued at ~160 billion RMB), serving nearly 5,000 institutions across 31 provinces and 25 departments with 8 million annual visits.
Responsibilities: (1) System integration and data verification; (2) Platform front-end and back-end upgrade; (3) Coordinating technical solutions and personnel arrangements, communicating with multiple institutions, completing business/technical documentation and contract signing.

📋 Large-scale Scientific Instrument Procurement System Website

Ministry of Science and Technology Major Project | Student Lead
Overview: The Large-scale Scientific Instrument Procurement System is the unified channel for institutions to apply for instrument procurement on the national platform. It serves as the data source for the duplicate review system and a powerful tool for management departments such as the Ministry of Science and Technology and the Ministry of Finance. Following unified data standards, it standardizes the application process for all institutions, supports procurement applications and appeals, and provides task management and result publication functions.
Responsibilities: (1) Designing system architecture using Vue + SpringBoot front-end/back-end separation; (2) Coordinating personnel and managing project progress; (3) Completing business/technical documentation and contract signing.

🔬 Shenzhen Major Scientific Infrastructure and Instrument Sharing Platform Website

Shenzhen Municipal Project | Student Lead
Overview: A regional instrument sharing platform following the unified data standards of the national network management platform. Standardizes application processes for all institutions, supports instrument procurement applications and appeals, and provides task management and result publication functions for management departments.
Responsibilities: (1) Designing system architecture using Vue + SpringBoot front-end/back-end separation; (2) Coordinating personnel and managing project progress; (3) Completing business/technical documentation and contract signing.

🚁 UAV Intelligent Inspection Key Technologies and 3D Platform

Power Grid Inspection Project | Student Lead
Overview: This project uses airborne LiDAR to complete high-precision holographic digital modeling of power lines and towers in Jiangsu Province. Based on 3D point cloud data, it achieves real-time centimeter-level trajectory planning for multi-rotor UAV power tower inspection. Finally, a complete power tower inspection system was developed to achieve digital, automated, and refined inspection.
Responsibilities: (1) Flight space grid modeling based on 3D point cloud data and UAV hovering accuracy via ArcGIS interface; (2) Constructing objective functions for minimum energy consumption and minimum lighting cost, considering natural wind and lighting effects; (3) Improving Ant Colony + A* hybrid algorithm for real-time path planning under different times/wind directions; (4) Developing power inspection system prototype tools including path planning, safety analysis, knowledge base, parameter library, and visualization engine.

🖥️ Code Debugging and Runtime Monitoring Technology

13th Five-Year Pre-research Project (Equipment Development Department) | Student Lead
Overview: This project focuses on the Tianyi integrated development environment, researching code debugging and runtime monitoring technology for multi-task parallel programs based on the Tianyi embedded operating system. Based on the Eclipse plugin platform, using program instrumentation, host-target network communication, and SWT technologies, key modules including parallel program process replay, dynamic tracking and precise positioning, real-time monitoring and analysis were developed. The project ultimately achieved program debugging with process replay, precise error scene tracking and positioning, and a low-overhead runtime state real-time monitoring and analysis debugging prototype tool.
Responsibilities: (1) Designing and implementing the interface and operations of parallel program debugging tools; (2) Implementing parallel program error detection and analysis, Loongson 3A target machine runtime state real-time monitoring, and host-target network communication modules; (3) Preparing software design documentation.

Education

Beihang University - Ph.D. Candidate in Computer Software and Theory
2021.09 - Present | Beijing, China
Nanjing University of Aeronautics and Astronautics - M.S. in Transportation Planning and Management
2018.09 - 2021.03 | Nanjing, China
Lyuliang University - B.E. in Safety Engineering
2013.09 - 2017.06 | Shanxi, China