NoisyNet-A3C
2000년 도입 · 논문 1편에서 사용
NoisyNet-A3C is a modification of A3C that utilises noisy linear layers for exploration instead of $\epsilon$-greedy exploration as in the original DQN formulation.
출처: Noisy Networks for Exploration
소개 논문: Noisy Networks for Exploration
Policy Gradient Methods · Reinforcement Learning