The Wasserstein Believer: Learning Belief Updates for Partially Observable Environments through Reliable Latent Space Models.
Raphaël AvalosFlorent DelgrangeAnn NowéGuillermo A. PérezDiederik M. RoijersPublished in: CoRR (2023)
Keyphrases
- partially observable environments
- learning algorithm
- reinforcement learning
- inverse reinforcement learning
- active learning
- learning tasks
- prior knowledge
- dynamic programming
- probabilistic model
- search algorithm
- model selection
- closed form
- latent variables
- hidden variables
- partially observable
- latent space
- cooperative games