Timezone: »

Provable Acceleration of Heavy Ball beyond Quadratics for a Class of Polyak-Lojasiewicz Functions when the Non-Convexity is Averaged-Out
Jun-Kun Wang · Chi-Heng Lin · Andre Wibisono · Bin Hu

Wed Jul 20 03:30 PM -- 05:30 PM (PDT) @ Hall E #1317

Heavy Ball (HB) nowadays is one of the most popular momentum methods in non-convex optimization. It has been widely observed that incorporating the Heavy Ball dynamic in gradient-based methods accelerates the training process of modern machine learning models. However, the progress on establishing its theoretical foundation of acceleration is apparently far behind its empirical success. Existing provable acceleration results are of the quadratic or close-to-quadratic functions, as the current techniques of showing HB's acceleration are limited to the case when the Hessian is fixed. In this work, we develop some new techniques that help show acceleration beyond quadratics, which is achieved by analyzing how the change of the Hessian at two consecutive time points affects the convergence speed. Based on our technical results, a class of Polyak-Lojasiewicz (PL) optimization problems for which provable acceleration can be achieved via HB is identified. Moreover, our analysis demonstrates a benefit of adaptively setting the momentum parameter.

Author Information

Jun-Kun Wang (Yale University)
Chi-Heng Lin (Georgia Institute of Technology)
Andre Wibisono (Yale University)
Bin Hu (University of Illinois at Urbana-Champaign)

Related Events (a corresponding poster, oral, or spotlight)

More from the Same Authors