Efficient Hyperparameter Optimization for Physics-based Character Animation
This addresses the computational bottleneck in hyperparameter optimization for physics-based character animation, which is an incremental improvement within a specific domain.
The paper tackles the problem of computationally expensive hyperparameter tuning for Deep Reinforcement Learning-based physics-based character animation by proposing a Curriculum-based Multi-Fidelity Bayesian Optimization framework (CMFBO). The result is at least 5x efficiency gain compared to author-released settings in DeepMimic.
Physics-based character animation has seen significant advances in recent years with the adoption of Deep Reinforcement Learning (DRL). However, DRL-based learning methods are usually computationally expensive and their performance crucially depends on the choice of hyperparameters. Tuning hyperparameters for these methods often requires repetitive training of control policies, which is even more computationally prohibitive. In this work, we propose a novel Curriculum-based Multi-Fidelity Bayesian Optimization framework (CMFBO) for efficient hyperparameter optimization of DRL-based character control systems. Using curriculum-based task difficulty as fidelity criterion, our method improves searching efficiency by gradually pruning search space through evaluation on easier motor skill tasks. We evaluate our method on two physics-based character control tasks: character morphology optimization and hyperparameter tuning of DeepMimic. Our algorithm significantly outperforms state-of-the-art hyperparameter optimization methods applicable for physics-based character animation. In particular, we show that hyperparameters optimized through our algorithm result in at least 5x efficiency gain comparing to author-released settings in DeepMimic.