More for Less: Safe Policy Improvement with Stronger Performance Guarantees

From International Center for Computational Logic

"8" is not in the list (Januar, Februar, März, April, Mai, Juni, Juli, August, September, Oktober, ...) of allowed values for the "Month" property.

Toggle side column

More for Less: Safe Policy Improvement with Stronger Performance Guarantees

Patrick WienhöftPatrick Wienhöft,  Marnix SuilenMarnix Suilen,  Thiago D. SimãoThiago D. Simão,  Clemens DubslaffClemens Dubslaff,  Christel BaierChristel Baier,  Nils JansenNils Jansen
Patrick Wienhöft, Marnix Suilen, Thiago D. Simão, Clemens Dubslaff, Christel Baier, Nils Jansen
More for Less: Safe Policy Improvement with Stronger Performance Guarantees
In Edith Elkind, eds., Proceedings of the Thirty-Second International Joint Conference on Artificial Intelligence, IJCAI-23, 4406--4415,  2023. International Joint Conferences on Artificial Intelligence Organization
  • KurzfassungAbstract
    In an offline reinforcement learning setting, the safe policy improvement (SPI) problem aims to improve the performance of a behavior policy according to which sample data has been generated. State-of-the-art approaches to SPI require a high number of samples to provide practical probabilistic guarantees on the improved policy's performance. We present a novel approach to the SPI problem that provides the means to require less data for such guarantees. Specifically, to prove the correctness of these guarantees, we devise implicit transformations on the data set and the underlying environment model that serve as theoretical foundations to derive tighter improvement bounds for SPI. Our empirical evaluation, using the well-established SPI with baseline bootstrapping (SPIBB) algorithm, on standard benchmarks shows that our method indeed significantly reduces the sample complexity of the SPIBB algorithm.
  • Forschungsgruppe:Research Group: Algebraische und logische Grundlagen der InformatikAlgebraic and Logical Foundations of Computer Science
@inproceedings{WSSDBJ2023,
  author    = {Patrick Wienh{\"{o}}ft and Marnix Suilen and Thiago D.
               Sim{\~{a}}o and Clemens Dubslaff and Christel Baier and Nils
               Jansen},
  title     = {More for Less: Safe Policy Improvement with Stronger Performance
               Guarantees},
  editor    = {Edith Elkind},
  booktitle = {Proceedings of the Thirty-Second International Joint Conference
               on Artificial Intelligence, {IJCAI-23}},
  publisher = {International Joint Conferences on Artificial Intelligence
               Organization},
  year      = {2023},
  pages     = {4406--4415},
  doi       = {10.24963/ijcai.2023/490}
}