UAI 2023poster4 citations

Loosely consistent emphatic temporal-difference learning

Jiamin He, Fengdi Che, Yi Wan, A. Rupam Mahmood

Abstract

There has been significant interest in searching for off-policy Temporal-Difference (TD) algorithms that find the same solution that would have been obtained in the on-policy regime. An important property of such algorithms is that their expected update has the same fixed point as that of On-policy TD($\lambda$), which we call

BibTeX
@InProceedings{pmlr-v216-he23a,
  title = 	 {Loosely consistent emphatic temporal-difference learning},
  author =       {He, Jiamin and Che, Fengdi and Wan, Yi and Mahmood, A. Rupam},
  booktitle = 	 {Proceedings of the Thirty-Ninth Conference on Uncertainty in Artificial Intelligence},
  pages = 	 {849--859},
  year = 	 {2023},
  editor = 	 {Evans, Robin J. and Shpitser, Ilya},
  volume = 	 {216},
  series = 	 {Proceedings of Machine Learning Research},
  month = 	 {31 Jul--04 Aug},
  publisher =    {PMLR},
  pdf = 	 {https://proceedings.mlr.press/v216/he23a/he23a.pdf},
  url = 	 {https://proceedings.mlr.press/v216/he23a.html},
  abstract = 	 {There has been significant interest in searching for off-policy Temporal-Difference (TD) algorithms that find the same solution that would have been obtained in the on-policy regime. An important property of such algorithms is that their expected update has the same fixed point as that of On-policy TD($\lambda$), which we call