A dual-memory architecture for reinforcement learning on neuromorphic platforms.
Wilkie Olin-AmmentorpYury SokolovMaxim BazhenovPublished in: Neuromorph. Comput. Eng. (2021)
Keyphrases
- reinforcement learning
- real time
- memory usage
- software platform
- memory management
- machine learning
- computing power
- function approximation
- memory requirements
- dynamic programming
- associative memory
- software architecture
- temporal difference
- management system
- reinforcement learning algorithms
- learning capabilities
- memory space
- multi agent systems
- memory hierarchy
- memory access
- memory size
- memory subsystem
- level parallelism
- learning algorithm
- multithreading
- computing platform
- data flow