Evolution Strategies as a Scalable Alternative to Reinforcement Learning
원문 초록
We explore the use of Evolution Strategies (ES), a class of black box optimization algorithms, as an alternative to popular MDP-based RL techniques such as Q-learning and Policy Gradients. Experiments on MuJoCo and Atari show that ES is a viable solution strategy that scales extremely well with the number of CPUs available: By using a novel communication strategy based on common random numbers, our ES implementation only needs to communicate scalars, making it possible to scale to over a thousand parallel workers. This allows us to solve 3D humanoid walking in 10 minutes and obtain competitive results on most Atari games after one hour of training. In addition, we highlight several advantages of ES as a black box optimization technique: it is invariant to action frequency and delayed rewards, tolerant of extremely long horizons, and does not need temporal discounting or value function approximation.
함께 살펴볼 논문
공통 태그를 근거로 연결했습니다. 실제 인용 관계나 논문의 우수성을 뜻하지 않습니다.
Generalized Adaptive Transfer Network: Enhancing Transfer Learning in Reinforcement Learning Across Domains
Deep Reinforcement Learning with Gradient Eligibility Traces
Aligning Humans and Robots via Reinforcement Learning from Implicit Human Feedback
Reinforcement Learning-Based Policy Optimisation For Heterogeneous Radio Access
PGDQN: Preference-Guided Deep Q-Network
Task Independent Capsule-Based Agents for Deep Q-Learning
토론
이 논문과 연결된 포럼 글이 아직 없습니다.