About Me
I'm a Research Scientist at TCL AI Lab, Hong Kong, where I focus on applying machine learning and AI for real-world industrial applications at scale.
I received my PhD in Electrical and Electronic Engineering from The Hong Kong Polytechnic University in 2024, where my thesis focused on multimedia signal processing and social network analysis.
14+
Publications
6+
Yrs of Research
PhD · PolyU
TCL AI Lab
Research Interests
Multimedia Signal Processing
Leveraging deep learning to fuse heterogeneous data sources including vision and text for holistic understanding.
Social Network Analysis
Building AI systems that detect misinformation, verify claims, and understand how content propagates through online communities.
Industrial Artificial Intelligence
Applying LLMs and vision models to industrial challenges including knowledge management, predictive maintenance, and materials discovery.
Development Skills
PyTorch
HuggingFace
Scrapy
LangChain
React.js
FastAPI
MySQL
Tableau
List of Projects
-
Post-Training of Large Language Models for Patent Analytics in Scientific Discovery (2025 - Now)
-
RAG-Powered Large Language Models for Industrial AI Assistant (2024 - 2025)
-
Multimodal Deep Learning for Fake News and Hate Speech Detection on Social Media (2020 - 2023)
-
Natural Language Processing for Social Media Business Intelligence (2019 - 2020)
Publications
- Ka-Chun Fung, Tsun-Hin Cheung, Songjiang Lai, et al., "Zero-Shot Cross-Domain Transfer for Remaining Useful Life Prediction of Aero-Engines via Iterative Diffusion Refinement," Reliability Engineering & System Safety, 112786, 2026.
- Tsun-Hin Cheung, Wei Xu, Xinshi Lin, et al., "Enhancing Chemical Capabilities of Large Language Models for OLED Materials Design," International Conference on Display Technology, Chongqing, China, 2026.
- Kuong-Hon Pong, Kin-Chung Chan, Tsun-Hin Cheung, et al., "Towards Robust Ergonomic Assessment for Lifting Postures with Vision-Language Models," International Conference on Control Engineering and Artificial Intelligence, Kowloon, Hong Kong, 2026.
- Songjiang Lai, Tsun-Hin Cheung, Ka-Chun Fung, et al., "Enhancing Technical Documents Retrieval for RAG," Asia Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC), Singapore, 2025.
- Songjiang Lai, Tsun-Hin Cheung, Ka-Chun Fung, et al., "Fault Diagnosis of Rolling Bearings Using Denoising Multi-Channel Mixture of CNN and Mamba-Enhanced Adaptive Self-Attention LSTM," Sensors, vol. 25, no. 21, pp. 6652, 2025.
- Songjiang Lai, Tsun-Hin Cheung, Ka-Chun Fung, et al., "HAAT: Hybrid Attention Aggregation Transformer for Image Super-Resolution," SPIE International Workshop on Advanced Imaging Technology (IWAIT), Yunlin, Taiwan, 2025.
- Songjiang Lai, Tsun-Hin Cheung, Jiayi Zhao, et al., "Residual Attention Single-Head Vision Transformer Network for Rolling Bearing Fault Diagnosis in Noisy Environments," ACM International Conference on Video, Signal and Image Processing, Ningbo, China, 2024.
- Songjiang Lai, Tsun-Hin Cheung, Ka-Chun Fung, et al., "An End-to-End Two-Stream Network Based on RGB Flow and Representation Flow for Human Action Recognition," Asia Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC), Macau, China, 2024.
- Tsun-Hin Cheung, Ka-Chun Fung, Songjiang Lai, et al., "Automatic Prompt Generation and Grounding Object Detection for Zero-Shot Image Anomaly Detection," Asia Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC), Macau, China, 2024.
- Tsun-Hin Cheung and Kin-Man Lam, "FactLLaMA: Optimizing Instruction-Following Language Models with External Knowledge for Automated Fact-Checking," Asia Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC), Taipei, Taiwan, 2023.
- Tsun-Hin Cheung and Kin-Man Lam, "Causal Diffused Graph-Transformer Network with Stacked Early Classification Loss for Efficient Stream Classification of Rumours," Knowledge-Based Systems, vol. 277, pp. 110807, 2023.
- Tsun-Hin Cheung and Kin-Man Lam, "Crossmodal Bipolar Attention for Multimodal Classification on Social Media," Neurocomputing, vol. 514, pp. 1-12, 2022.
- Tsun-Hin Cheung and Kin-Man Lam, "Simultaneous Fake News and Topic Classification via Auxiliary Task Learning," Asia Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC), Auckland, New Zealand, 2020.