A Survey on Reproducibility by Evaluating Deep Reinforcement Learning Algorithms on Real-World Robots

Nicolai Anton Lynnerup*, Laura Nolling, Rasmus Hasle, John Hallam

*Kontaktforfatter for dette arbejde

Publikation: Bidrag til tidsskriftKonferenceartikelForskningpeer review

Abstrakt

As reinforcement learning (RL) achieves more success in solving complex tasks, more care is needed to ensure that RL research is reproducible and that algorithms herein can be compared easily and fairly with minimal bias. RL results are, however, notoriously hard to reproduce due to the algorithms’ intrinsic variance, the environments’ stochasticity, and numerous (potentially unreported) hyper-parameters. In this work we investigate the many issues leading to irreproducible research and how to manage those. We further show how to utilise a rigorous and standardised evaluation approach for easing the process of documentation, evaluation and fair comparison of different algorithms, where we emphasise the importance of choosing the right measurement metrics and conducting proper statistics on the results, for unbiased reporting of the results.
OriginalsprogEngelsk
TidsskriftJournal of Machine Learning Research
Vol/bind100
Sider (fra-til)466-489
ISSN1532-4435
StatusUdgivet - 2020
Begivenhed3rd Conference on Robot Learning: CoRL 2019 - Osaka, Japan
Varighed: 30. okt. 20191. nov. 2019

Konference

Konference3rd Conference on Robot Learning
LandJapan
ByOsaka
Periode30/10/201901/11/2019

Fingeraftryk Dyk ned i forskningsemnerne om 'A Survey on Reproducibility by Evaluating Deep Reinforcement Learning Algorithms on Real-World Robots'. Sammen danner de et unikt fingeraftryk.

Citationsformater