Automatic Reward Shaping from Confounded Offline Data

Mingxuan Li, Junzhe Zhang, Elias Bareinboim
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:36765-36793, 2025.

Abstract

Reward shaping has been demonstrated to be an effective technique for accelerating the learning process of reinforcement learning (RL) agents. While successful in empirical applications, the design of a good shaping function is less well understood in principle and thus often relies on domain expertise and manual design. To overcome this limitation, we propose a novel automated approach for designing reward functions from offline data, possibly contaminated with the unobserved confounding bias.We propose to use causal state value upper bounds calculated from offline datasets as a conservative optimistic estimation of the optimal state value, which is then used as state potentials in Potential-Based Reward Shaping (PBRS). When applying our shaping function to a model-free learner based on UCB principles, we show that it enjoys a better gap-dependent regret bound than the learner without shaping. To the best of our knowledge, this is the first gap-dependent regret bound for PBRS in model-free learning with online exploration. Simulations support the theoretical findings.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-li25dr, title = {Automatic Reward Shaping from Confounded Offline Data}, author = {Li, Mingxuan and Zhang, Junzhe and Bareinboim, Elias}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {36765--36793}, year = {2025}, editor = {Singh, Aarti and Fazel, Maryam and Hsu, Daniel and Lacoste-Julien, Simon and Berkenkamp, Felix and Maharaj, Tegan and Wagstaff, Kiri and Zhu, Jerry}, volume = {267}, series = {Proceedings of Machine Learning Research}, month = {13--19 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v267/main/assets/li25dr/li25dr.pdf}, url = {https://proceedings.mlr.press/v267/li25dr.html}, abstract = {Reward shaping has been demonstrated to be an effective technique for accelerating the learning process of reinforcement learning (RL) agents. While successful in empirical applications, the design of a good shaping function is less well understood in principle and thus often relies on domain expertise and manual design. To overcome this limitation, we propose a novel automated approach for designing reward functions from offline data, possibly contaminated with the unobserved confounding bias.We propose to use causal state value upper bounds calculated from offline datasets as a conservative optimistic estimation of the optimal state value, which is then used as state potentials in Potential-Based Reward Shaping (PBRS). When applying our shaping function to a model-free learner based on UCB principles, we show that it enjoys a better gap-dependent regret bound than the learner without shaping. To the best of our knowledge, this is the first gap-dependent regret bound for PBRS in model-free learning with online exploration. Simulations support the theoretical findings.} }
Endnote
%0 Conference Paper %T Automatic Reward Shaping from Confounded Offline Data %A Mingxuan Li %A Junzhe Zhang %A Elias Bareinboim %B Proceedings of the 42nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Aarti Singh %E Maryam Fazel %E Daniel Hsu %E Simon Lacoste-Julien %E Felix Berkenkamp %E Tegan Maharaj %E Kiri Wagstaff %E Jerry Zhu %F pmlr-v267-li25dr %I PMLR %P 36765--36793 %U https://proceedings.mlr.press/v267/li25dr.html %V 267 %X Reward shaping has been demonstrated to be an effective technique for accelerating the learning process of reinforcement learning (RL) agents. While successful in empirical applications, the design of a good shaping function is less well understood in principle and thus often relies on domain expertise and manual design. To overcome this limitation, we propose a novel automated approach for designing reward functions from offline data, possibly contaminated with the unobserved confounding bias.We propose to use causal state value upper bounds calculated from offline datasets as a conservative optimistic estimation of the optimal state value, which is then used as state potentials in Potential-Based Reward Shaping (PBRS). When applying our shaping function to a model-free learner based on UCB principles, we show that it enjoys a better gap-dependent regret bound than the learner without shaping. To the best of our knowledge, this is the first gap-dependent regret bound for PBRS in model-free learning with online exploration. Simulations support the theoretical findings.
APA
Li, M., Zhang, J. & Bareinboim, E.. (2025). Automatic Reward Shaping from Confounded Offline Data. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:36765-36793 Available from https://proceedings.mlr.press/v267/li25dr.html.

Related Material