RL, I Choose You: Using RL to learn mutation strategies in fuzzing
Lightning Talk
In this talk I will motivate some of the key challenges in fuzzing, particularly related to increasing code coverage and finding bugs and vulnerabilities in programs. I will then discuss how this problem can be consider an exploration/exploitation problem particularly with relation Reinforcement Learning (RL). I will then propose an approach to use a special case of RL to select the strategy by selecting a location for mutating a test case. I will then demonstrate how this can be applied in practice and share some preliminary results.