Predictive Response Optimization: Using Reinforcement Learning to Fight Online Social Network Abuse
Garrett Wilson
34th USENIX Security Symposium (USENIX Security '25) · Day 2 · Web and Mobile Security
In the realm of online social networks, the battle against abuse is a perpetual arms race. Traditional approaches have largely focused on the *detection* of malicious activities, often grappling with the inherent trade-off between precision and recall. However, as Garrett Wilson articulates in his USENIX Security talk, "Predictive Response Optimization: Using Reinforcement Learning to Fight Online Social Network Abuse," the true challenge and opportunity lie not just in detection, but in the intelligent *selection of enforcement actions*. This presentation introduces a paradigm shift, reframing the problem from a binary classification task to an **action selection** challenge, which can be optimally addressed using **reinforcement learning (RL)**.
AI review
Solid applied ML security research with real production deployments and honest metrics — this is exactly the kind of work USENIX Security exists for. The RL+MPC framing for enforcement action selection is a genuine conceptual contribution, and the case studies (adversarial adaptation detection, bug response in under 48 hours) are the kind of operational honesty that separates real systems papers from academic theater.