Balancing Value Underestimation and Overestimation with Realistic Actor-Critic
Li, Sicen, Tang, Qinyun, Pang, Yiming, Ma, Xinmeng, Wang, Gang
–arXiv.org Artificial Intelligence
Model-free deep reinforcement learning (RL) has been successfully applied to challenging continuous control domains. However, poor sample efficiency prevents these methods from being widely used in real-world domains. This paper introduces a novel model-free algorithm, Realistic Actor-Critic(RAC), which can be incorporated with any off-policy RL algorithms to improve sample efficiency. RAC employs Universal Value Function Approximators (UVFA) to simultaneously learn a policy family with the same neural network, each with different trade-offs between underestimation and overestimation. To learn such policies, we introduce uncertainty punished Q-learning, which uses uncertainty from the ensembling of multiple critics to build various confidence-bounds of Q-function. We evaluate RAC on the MuJoCo benchmark, achieving 10x sample efficiency and 25\% performance improvement on the most challenging Humanoid environment compared to SAC.
arXiv.org Artificial Intelligence
Oct-26-2022
- Country:
- Oceania > Australia
- New South Wales (0.04)
- Asia > China
- Heilongjiang Province > Harbin (0.04)
- Oceania > Australia
- Genre:
- Research Report > Promising Solution (0.34)
- Technology: