search

reproducibility-challenge / iclr_2019

ICLR Reproducibility Challenge 2019
https://reproducibility-challenge.github.io/iclr_2019/
219 stars 40 forks source link

submission for #67 #160

Open mhtsbt opened 5 years ago

mhtsbt commented 5 years ago

submission for #67

reproducibility-org commented 5 years ago

Hi, please find below a review submitted by one of the reviewers:

Score: 5 Reviewer 1 comment : Successor Options is a novel strategy to discover options for RL algorithms. The authors have reproduced a subset of the original experiments proposed in the main paper. Since they considered the simplest environment, the scope of reproducibility study is quite limited. Further, the report does not direct compare with the results reported in the paper. The overall writing is easy to follow though there are some grammatical mistake (eg good canidate subgoals)

Confidence : 3