<ul><li>A new approach to regret analysis of online convex optimization algorithms is proposed.</li><li>The approach leverages Integral Quadratic Constraints (IQCs) to derive a semi-definite program for regret guarantee.</li><li>The concept of variational IQCs is introduced, which generalizes IQCs to time-varying monotone operators.</li><li>The results do not require the assumption of gradient boundedness or a bounded feasible set.</li></ul>

Online Convex Optimization and Integral Quadratic Constraints: A new approach to regret analysis

Discover more