DARL: Distributed Reconfigurable Accelerator for Hyperdimensional Reinforcement Learning.
Hanning ChenMariam IssaYang NiMohsen ImaniPublished in: ICCAD (2022)
Keyphrases
- reinforcement learning
- multi agent
- distributed systems
- function approximation
- compute intensive
- field programmable gate array
- general purpose
- cooperative
- distributed data
- lightweight
- state space
- fault tolerant
- transfer learning
- peer to peer
- learning process
- hardware implementation
- reconfigurable architecture
- data sets
- communication cost
- load balancing
- dynamic programming
- web services
- neural network