Kaixhin/Atari

Implement Pop-Art

Open

#6 建立於 2016年4月13日

在 GitHub 查看
 (0 留言) (0 反應) (0 負責人)Lua (72 fork)github user discovery
enhancementhelp wanted

倉庫指標

Star
 (263 star)
PR 合併指標
 (30 天內沒有已合併 PR)

描述

Learning functions across many orders of magnitudes introduces Preserving Outputs Precisely, while Adaptively Rescaling Targets (Pop-Art). In summary it normalises outputs across orders of magnitudes and gets rid of the clipping (i.e. counting) rewards heuristic for Atari games. The normalisation is also better for non-stationary problems, i.e., any decent real world problem.

The below is a picture of extra notes from the authors, next to their poster at NIPS 2016: img_20161207_190428

貢獻者指南