RL-Scope: Cross-Stack Profiling for Deep Reinforcement Learning Workloads.
James GleesonSrivatsan KrishnanMoshe GabelVijay Janapa ReddiEyal de LaraGennady PekhimenkoPublished in: CoRR (2021)
Keyphrases
- reinforcement learning
- function approximation
- model free
- state space
- reinforcement learning algorithms
- rl algorithms
- database systems
- markov decision processes
- multi agent
- action selection
- learning algorithm
- optimal policy
- learning problems
- control problems
- transfer learning
- continuous state
- actor critic
- partially observable domains
- markov decision process
- multi agent reinforcement learning
- temporal difference
- temporal difference learning
- state action
- policy iteration
- partially observable
- machine learning
- supervised learning
- action space
- control strategies
- autonomous learning
- computer systems
- state abstraction
- policy search
- mobile robot
- continuous state and action spaces