Papers
1
Total Citations
4
H-Index
1
About
Jiakai Song is a researcher advancing the frontiers of reinforcement learning, with a primary focus on multi-agent systems, self-play mechanisms, and parameterized action spaces. His most-cited work, "Distributed Reinforcement Learning with Self-Play in Parameterized Action Space" (2021), addresses a critical challenge in competitive environments: stabilizing training when rewards are sparse, such as in scoring tasks with a goalkeeper. By integrating self-play with distributed learning, Song demonstrates how agents can autonomously generate a robust curriculum without direct supervision, improving performance in complex, adversarial settings. This contribution has earned his paper 4 citations, marking early recognition of its potential to enhance AI training in domains like robotics and game AI. Song’s research is particularly notable for tackling the instability of self-play in sparse-reward scenarios, a persistent hurdle in reinforcement learning. His work offers practical insights for developing more resilient autonomous agents, making it a valuable reference for students and researchers exploring multi-agent coordination and competitive AI.
Research Focus
Key Achievements
Top Papers
- 1