Recurrent networks, hidden states and beliefs in partially observable environments.
Gaspard LambrechtsAdrien BollandDamien ErnstPublished in: Trans. Mach. Learn. Res. (2022)
Keyphrases
- recurrent networks
- partially observable environments
- hidden states
- hidden markov models
- hidden state
- partially observable
- recurrent neural networks
- feed forward
- inverse reinforcement learning
- reinforcement learning
- conditional random fields
- hidden variables
- biologically inspired
- reinforcement learning algorithms
- neural network
- partially observable markov decision processes
- belief revision
- dynamic bayesian networks
- decision making
- exponential family
- state space
- decision problems
- generative model
- markov decision processes
- bayesian networks
- knowledge base
- finite state
- probabilistic model
- belief state
- fuzzy logic