Mingyang Wu
Undergraduate@SUSTech Research Assistant@CUHK MMLab
I like aeroplanes and travelling. Canton, China
Profile
I am currently an RA at CUHK MMLab, advised by Prof. Xiangyu Yue, where I study data-centric AI and RL to improve audio-video understanding in multimodal LLMs.
Previously, I worked on BCI and neural decoding at SUSTech NCC Lab with Prof. Quanying Liu. I am an UG at SUSTech and Co-Founder of WormForce Corp.
Core Skills
- Data-Centric AI
- Reinforcement Learning
- Multimodal LLMs
- Audio-Video Understanding
- BCI & Neural Decoding
Links
Publications
AVCap: Reinforcing Audio-Video Joint Caption with Detail-Aware Reward
Under Review
- RealMind: Advancing visual decoding and language interaction via EEG signals
2025 IEEE International Conference on Multimedia and Expo (ICME), 1-6 (2025)
PDF - An EEG Dataset for Multimodal Semantic Alignment and Neural Decoding during Reading and Listening
Scientific Data (2025)
PDF - MindPilot: Closed-loop Visual Stimulation Optimization for Brain Modulation with EEG-guided Diffusion
ICLR 2026 (2026)
PDF - BrainFLORA: Uncovering Brain Concept Representation via Multimodal Neural Embeddings
Proceedings of the 33rd ACM International Conference on Multimedia, 5577-5586 (2025)
PDF