Neural Map: Structured Memory for Deep Reinforcement Learning.
Emilio ParisottoRuslan SalakhutdinovPublished in: CoRR (2017)
Keyphrases
- reinforcement learning
- associative memory
- fitted q iteration
- function approximation
- neural network
- network architecture
- memory usage
- maximum a posteriori
- memory requirements
- memory size
- optimal control
- temporal difference
- structured data
- machine learning
- model free
- reinforcement learning algorithms
- state space
- real world
- markov decision processes
- information processing
- bio inspired
- data sets
- computing power
- memory space
- limited memory
- robotic control
- data streams