Zhaoyi Song
Zhaoyi Song
Home
Publications
Projects
Contact
Light
Dark
Automatic
paper-conference
HELSA: Hierarchical Reinforcement Learning with Spatiotemporal Abstraction for Large-Scale Multi-Agent Path Finding
Proposed a hierarchical reinforcement learning framework for large-scale multi-agent pathfinding, addressing the challenges of sample inefficiency resulting from sparse rewards and partial observability.
Zhaoyi Song
,
Rongqing Zhang
,
Xiang Cheng
PDF
Cite
Poster
Slides
Video
Image change detection with only a few samples
A power plant aimed to detect anomalies in surveillance footage. Given the scarcity of real-world data, realized a strategy to develop a synthetic dataset and train models to detect discrepancies in the footage.
Ke Liu
,
Zhaoyi Song
,
Haoyue Bai
PDF
Cite
Cite
×