NEON2: Finding Local Minima via First-Order Oracles
Zeyuan Allen-Zhu · Yuanzhi Li
Keywords:
Non-Convex Optimization
2018 Poster
Abstract
We propose a reduction for non-convex optimization that can (1) turn an stationary-point finding algorithm into an local-minimum finding one, and (2) replace the Hessian-vector product computations with only gradient computations. It works both in the stochastic and the deterministic settings, without hurting the algorithm's performance.
As applications, our reduction turns Natasha2 into a first-order method without hurting its theoretical performance. It also converts SGD, GD, SCSG, and SVRG into algorithms finding approximate local minima, outperforming some best known results.
Chat is not available.
Successful Page Load