A model based approach to exploration of continuous-state MDPs using Divergence-to-Go

著录项

• 来源
  《IEEE International Workshop on Machine Learning for Signal Processing》|2015年|1-6|共6页
• 会议地点
• 作者

  Emigh Matthew; Kriminger Evan; Principe Jose C.;

  展开▼
• 作者单位

  展开▼
• 会议组织
• 原文格式 PDF
• 正文语种
• 中图分类
• 关键词

  Markov processes; decision theory; dynamic programming; information theory; learning (artificial intelligence); Markov decision process; continuous-state MDPs; divergence-to-go; dynamic programming; exploration method; information-theoretic estimators; model based approach; reinforcement learning; sequential decision problem; Computational modeling; Kernel; Learning (artificial intelligence); Markov processes; Measurement uncertainty; Monte Carlo methods; Uncertainty;

  机译：马尔可夫过程;决策理论;动态规划;信息论;学习（人工智能）;马尔可夫决策过程;连续状态MDPs;待发散;动态规划;探索方法;信息理论估计量;基于模型的方法;强化学习;顺序决策问题;计算建模;内核;学习（人工智能）;马尔可夫过程;测量不确定度;蒙特卡洛方法;不确定度;