ACTKR

2000년 도입 · 논문 2편에서 사용

ACKTR, or Actor Critic with Kronecker-factored Trust Region, is an actor-critic method for reinforcement learning that applies trust region optimization using a recently proposed Kronecker-factored approximation to the curvature. The method extends the framework of natural policy gradient and optimizes both the actor and the critic using Kronecker-factored approximate curvature (K-FAC) with trust region.

출처: Scalable trust-region method for deep reinforcement learning using Kronecker-factored approximation

소개 논문: Scalable trust-region method for deep reinforcement learning using Kronecker-factored approximation