Charel_CHEN的博客

OpenAI_gym的官网案例Tags: openAI_gym创建，渲染，随机选择动作当然这只是gym的一个游戏，还有一些如： MountainCar-v0, MsPacman-v0 (requires the Atari dependency), or Hopper-v1 (requires the MuJoCo dependencies). Environments all descend fr

FeUdal Networks for Hierarchical Reinforcement Learning标签（空格分隔）： 论文笔记 增强学习算法FeUdal Networks for Hierarchical Reinforcement LearningAbstractIntroductionmodelLearningTransition Policy GradientsArch

DARLA: Improving Zero-Shot Transfer in Reinforcement Learning标签（空格分隔）： 论文笔记 增强学习算法DARLA Improving Zero-Shot Transfer in Reinforcement Learning目的和意义训练领域和应用领域source domain and target domain算法细则该论文主要讲

DARLA 源码解析标签（空格分隔）： 增强学习算法 源码'''Implementation of DARLA preprocessing, as found in DARLA: Improving Zero-Shot Transfer in Reinforcement Learningby Higgins and Pal et al (https://arxiv.org/pdf/1707.08

Dynamic Programming标签（空格分隔）： Reinforcement Learning: An Introduction Dynamic ProgrammingDynamic ProgrammingPolicy EvaluationPolicy ImprovementPolicy IterationValue IterationAsynchronous Dynamic P

A2C Advantage Actor-Critic (离散空间)标签（空格分隔）： 增强学习算法 源码import numpy as npimport tensorflow as tfimport gymnp.random.seed(2)tf.set_random_seed(2) # reproducible# SuperparametersOUTPUT_GRAPH = False #

Asynchronous Methods for Deep Reinforcement Learning 阅读笔记标签（空格分隔）： 增强学习算法 论文笔记本文的贡献在于提出了异步学习的算法，并应用在A2C Q-learning等算法中该论文作者提出了异步训练（Asynchronous Methods）的方法应用到强化学习的各个算法中（Sarsa,one-step Q-learning n-ste