PEARL: Enabling Portable, Productive, and High-Performance Deep Reinforcement Learning using Heterogeneous Platforms.
Yuan MengMichael KinsnerDeshanand P. SinghMahesh A. IyerViktor K. PrasannaPublished in: CF (2024)
Keyphrases
- heterogeneous platforms
- reinforcement learning
- web services
- function approximation
- state space
- deep learning
- markov decision processes
- optimal policy
- learning algorithm
- shared memory
- machine learning
- dynamic programming
- multi agent
- robotic control
- collaborative learning
- high reliability
- causal models
- temporal difference
- reinforcement learning algorithms
- open ended
- lightweight
- conditional independence
- multi agent reinforcement learning
- distributed memory
- model free
- high efficiency
- optimal control
- belief revision
- multi view