Dynamic Regret Bounds for Online Omniprediction with Long Term Constraints
This work addresses a foundational challenge in online learning for multi-agent systems with constraints, offering a novel solution with broad implications for decision-making under uncertainty.
The paper tackles the problem of online omniprediction with long-term constraints by developing an algorithm that provides dynamic regret bounds for all agents, ensuring worst-case utility guarantees and vanishing constraint violations without requiring agents to maintain state.
We present an algorithm guaranteeing dynamic regret bounds for online omniprediction with long term constraints. The goal in this recently introduced problem is for a learner to generate a sequence of predictions which are broadcast to a collection of downstream decision makers. Each decision maker has their own utility function, as well as a vector of constraint functions, each mapping their actions and an adversarially selected state to reward or constraint violation terms. The downstream decision makers select actions "as if" the state predictions are correct, and the goal of the learner is to produce predictions such that all downstream decision makers choose actions that give them worst-case utility guarantees while minimizing worst-case constraint violation. Within this framework, we give the first algorithm that obtains simultaneous \emph{dynamic regret} guarantees for all of the agents -- where regret for each agent is measured against a potentially changing sequence of actions across rounds of interaction, while also ensuring vanishing constraint violation for each agent. Our results do not require the agents themselves to maintain any state -- they only solve one-round constrained optimization problems defined by the prediction made at that round.