GCN-RL Circuit Designer: Transferable Transistor Sizing with Graph Neural Networks and Reinforcement Learning.
Hanrui WangKuan WangJiacheng YangLinxiao ShenNan SunHae-Seung LeeSong HanPublished in: DAC (2020)
Keyphrases
- reinforcement learning
- neural network
- high speed
- function approximation
- function approximators
- circuit design
- reinforcement learning algorithms
- optimal policy
- learning algorithm
- state space
- learning classifier systems
- model free
- markov decision processes
- rl algorithms
- temporal difference
- dynamic programming
- back propagation
- machine learning
- graph structure
- low power
- power dissipation
- transfer learning
- artificial neural networks
- learning process
- learning problems
- genetic algorithm
- reinforcement learning methods
- pattern recognition
- learning agent
- learning capabilities
- action selection
- random walk
- integrated circuit
- weighted graph
- action space
- directed graph
- actor critic
- reinforcement learning problems
- direct policy search
- fuzzy logic
- low cost
- learning agents
- multi agent
- neural network model
- state and action spaces