Conjugated Discrete Distributions for Distributional Reinforcement Learning.
Björn LindenbergJonas NordqvistKarl-Olof LindahlPublished in: AAAI (2022)
Keyphrases
- reinforcement learning
- function approximation
- efficient algorithms to compute
- probability distribution
- continuous state and action spaces
- sequence prediction
- co occurrence
- random variables
- robotic control
- continuous domains
- model free
- machine learning
- temporal difference
- control problems
- kullback leibler divergence
- continuous state
- discrete geometry
- learning process
- discrete version
- multi agent
- learning algorithm
- markov decision processes
- transfer learning
- partially observable
- optimal policy
- markov decision process
- continuous attributes
- data sets