ML Anthology

The Paths Perspective on Value Learning

Sam Greydanus, Chris Olah
Distill 2019
doi:10.23915/distill.00020 /distill/2019/greydanus2019distill-paths/

Abstract

Distill articles are interactive publications and do not include traditional abstracts. This summary was written for the ML Anthology. Introduces the paths perspective as a framework for understanding value learning in reinforcement learning. Demonstrates how temporal difference learning merges intersecting trajectories to achieve greater statistical efficiency than Monte Carlo methods.

Distill Code Semantic Scholar

Cite

Text

Greydanus and Olah. "The Paths Perspective on Value Learning." Distill, 2019. doi:10.23915/distill.00020

Markdown

[Greydanus and Olah. "The Paths Perspective on Value Learning." Distill, 2019.](https://mlanthology.org/distill/2019/greydanus2019distill-paths/) doi:10.23915/distill.00020

BibTeX

@article{greydanus2019distill-paths,
  title     = {{The Paths Perspective on Value Learning}},
  author    = {Greydanus, Sam and Olah, Chris},
  journal   = {Distill},
  year      = {2019},
  doi       = {10.23915/distill.00020},
  url       = {https://mlanthology.org/distill/2019/greydanus2019distill-paths/}
}