Zero-Regret Performative Prediction Under Inequality Constraints

To date, work on performative prediction has only focused on unconstrained scenarios, neglecting the fact that many real-world learning problems are subject to constraints.