Binary Black-Box Attacks Against Static Malware Detectors with Reinforcement Learning in Discrete Action Spaces.
Mohammadreza EbrahimiJason PachecoWeifeng LiJames Lee HuHsinchun ChenPublished in: SP Workshops (2021)
Keyphrases
- black box
- action space
- reinforcement learning
- continuous state
- continuous state spaces
- continuous action
- state space
- detect malicious
- markov decision processes
- state and action spaces
- policy search
- dynamic analysis
- real valued
- black boxes
- white box
- control policies
- state action
- detecting malicious
- action selection
- test cases
- integration testing
- finite state
- function approximators
- static analysis
- state transition
- partially observable markov decision processes
- markov decision process
- temporal difference
- model free
- reverse engineering
- planning problems
- stochastic processes
- reinforcement learning algorithms
- optimal policy
- function approximation
- white box testing
- machine learning
- malware detection
- infinite horizon
- optimal control
- model checking
- markov chain
- dynamic programming
- multi agent
- training data