Papers
3
Total Citations
17
H-Index
3
About
Shasha Liu is a rising force in embodied AI and multi-agent reinforcement learning (MARL), whose work bridges the critical gap between simulated training and real-world robot deployment. Her research centers on developing robust policy learning frameworks that enable autonomous agents—from single robots to cooperative teams—to transfer skills from virtual environments to physical systems with minimal performance loss. Liu’s most impactful contribution, the NeuronsMAE environment (9 citations), provides a standardized benchmark for evaluating MARL algorithms in both cooperative and competitive multi-robot tasks, addressing a pressing need for realistic, reproducible testing grounds. She further advanced sim-to-real transfer through her comparative study of domain randomization methods (5 citations), systematically analyzing techniques that make policies resilient to environmental mismatches. Her NeuronsGym hybrid framework (3 citations) integrates high-fidelity simulation with real-robot validation, offering a complete pipeline for embodied AI research. By creating tools that accelerate the development of generalizable robot policies, Liu is helping to democratize access to cutting-edge reinforcement learning research, making it easier for labs worldwide to move from simulation to real-world impact.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3