Abstraction for Deep Reinforcement Learning

Murray Shanahan; Melanie Mitchell

IJCAI 2022poster36 citations

Abstraction for Deep Reinforcement Learning

Abstract

We characterise the problem of abstraction in the context of deep reinforcement learning. Various well established approaches to analogical reasoning and associative memory might be brought to bear on this issue, but they present difficulties because of the need for end-to-end differentiability. We review developments in AI and machine learning that could facilitate their adoption.

Survey Track: Machine Learning

BibTeX

@inproceedings{ijcai2022p780,
  title     = {Abstraction for Deep Reinforcement Learning},
  author    = {Shanahan, Murray and Mitchell, Melanie},
  booktitle = {Proceedings of the Thirty-First International Joint Conference on
               Artificial Intelligence, {IJCAI-22}},
  publisher = {International Joint Conferences on Artificial Intelligence Organization},
  editor    = {Lud De Raedt},
  pages     = {5588--5596},
  year      = {2022},
  month     = {7},
  note      = {Survey Track},
  doi       = {10.24963/ijcai.2022/780},
  url       = {https://doi.org/10.24963/ijcai.2022/780},
}