Adversarial Policies: Attacking Deep Reinforcement Learning.
Adam GleaveMichael DennisCody WildNeel KantSergey LevineStuart RussellPublished in: ICLR (2020)
Keyphrases
- reinforcement learning
- optimal policy
- policy search
- multi agent
- markov decision process
- control policies
- reward function
- partially observable markov decision processes
- reinforcement learning agents
- cooperative multi agent systems
- state space
- control policy
- markov decision processes
- function approximation
- markov decision problems
- total reward
- fitted q iteration
- decision problems
- continuous state
- macro actions
- hierarchical reinforcement learning
- model free
- policy gradient methods
- reinforcement learning algorithms
- dynamic programming
- learning algorithm
- learning process
- multi agent reinforcement learning
- optimal control
- information security
- search space
- robotic control
- supervised learning
- state and action spaces
- monte carlo
- function approximators
- revenue management
- infinite horizon
- temporal difference
- data mining
- decision processes