Papers
6
Total Citations
72
H-Index
5
About
Shangqi Guo is a leading researcher in reinforcement learning (RL) and robotics, whose work focuses on making autonomous agents more efficient, scalable, and biologically plausible. His primary contributions lie in hierarchical reinforcement learning (HRL), where he tackles the critical challenge of training efficiency. In his highly cited work, "Adjacency Constraint for Efficient Hierarchical Reinforcement Learning" (2022, 31 citations), Guo introduced a novel method to constrain the high-level goal space, dramatically accelerating learning in complex tasks. He further advanced HRL with "Hierarchical reinforcement learning from imperfect demonstrations through reachable coverage-based subgoal filtering" (2024, 10 citations), enabling agents to learn from suboptimal human data. Beyond HRL, Guo has pioneered continual imitation learning for robots (CRIL, 2021, 17 citations), allowing skill acquisition without catastrophic forgetting, and developed data-efficient self-supervised learning from demonstration videos (2021, 6 citations). His recent exploration into brain-inspired RL, including spiking neural networks (2024, 5 citations), and fast counterfactual inference for history-based RL (2023, 3 citations) demonstrates his commitment to bridging artificial and biological intelligence. Guo’s work is shaping the future of sample-efficient, lifelong learning robots.
Research Focus
Key Achievements
Top Papers
- 1Adjacency Constraint for Efficient Hierarchical Reinforcement Learning31 citations · 2022
- 2CRIL: Continual Robot Imitation Learning via Generative and Prediction Model17 citations · 2021
- 3
- 4
- 5
- 6Fast Counterfactual Inference for History-Based Reinforcement Learning3 citations · 2023