LRP-based Policy Pruning and Distillation of Reinforcement Learning Agents for Embedded Systems.
Rui XuSiyu LuanZonghua GuQingling ZhaoGang ChenPublished in: ISORC (2022)
Keyphrases
- embedded systems
- reinforcement learning agents
- dynamic environments
- reinforcement learning
- low cost
- embedded software
- embedded devices
- real time systems
- computing power
- optimal policy
- resource limited
- software systems
- hw sw
- state abstraction
- search space
- multi agent environments
- multi agent
- transfer learning
- action selection
- search algorithm
- hardware software
- embedded real time systems
- real time
- field programmable gate array
- learning algorithm
- flash memory
- initial state
- autonomous agents
- petri net
- protocol stack