Towered Actor Critic For Handling Multiple Action Types In Reinforcement Learning For Drug Discovery.
Sai Krishna GottipatiYashaswi PathakBoris Sattarov SahirRohan NuttallMohammad AminiMatthew E. TaylorSarath ChandarPublished in: AAAI (2021)
Keyphrases
- reinforcement learning
- actor critic
- drug discovery
- policy gradient
- temporal difference
- reinforcement learning algorithms
- function approximation
- optimal control
- action selection
- neuro fuzzy
- gradient method
- state space
- dynamic programming
- learning algorithm
- approximate dynamic programming
- neural network
- policy iteration
- action space
- state action
- markov decision processes
- data analysis
- machine learning
- data mining