Adversarial Policies: Attacking Deep Reinforcement Learning.
Adam GleaveMichael DennisNeel KantCody WildSergey LevineStuart RussellPublished in: CoRR (2019)
Keyphrases
- reinforcement learning
- optimal policy
- policy search
- markov decision process
- control policies
- multi agent
- reward function
- state space
- markov decision processes
- total reward
- reinforcement learning agents
- fitted q iteration
- function approximation
- partially observable markov decision processes
- hierarchical reinforcement learning
- control policy
- cooperative multi agent systems
- reinforcement learning algorithms
- dynamic programming
- markov decision problems
- model free
- decision problems
- learning process
- long run
- macro actions
- multiagent reinforcement learning
- infinite horizon
- average reward
- management policies
- learning problems
- temporal difference
- approximate policy iteration
- robotic control
- sufficient conditions
- continuous state
- temporal difference learning
- learning algorithm
- function approximators
- supervised learning
- deep learning