PatchAttack: A Black-Box Texture-Based Attack with Reinforcement Learning.
Chenglin YangAdam KortylewskiCihang XieYinzhi CaoAlan L. YuillePublished in: ECCV (26) (2020)
Keyphrases
- black box
- reinforcement learning
- black boxes
- function approximation
- model free
- white box
- markov decision processes
- state space
- temporal difference
- test cases
- learning algorithm
- optimal policy
- machine learning
- reinforcement learning algorithms
- dynamic programming
- state transition
- texture features
- countermeasures
- artificial intelligence
- multi agent
- rule extraction
- action space
- function approximators
- integration testing
- database