Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Single Loop Gaussian Homotopy Method for Non-convex Optimization

About

The Gaussian homotopy (GH) method is a popular approach to finding better stationary points for non-convex optimization problems by gradually reducing a parameter value $t$, which changes the problem to be solved from an almost convex one to the original target one. Existing GH-based methods repeatedly call an iterative optimization solver to find a stationary point every time $t$ is updated, which incurs high computational costs. We propose a novel single loop framework for GH methods (SLGH) that updates the parameter $t$ and the optimization decision variables at the same. Computational complexity analysis is performed on the SLGH algorithm under various situations: either a gradient or gradient-free oracle of a GH function can be obtained for both deterministic and stochastic settings. The convergence rate of SLGH with a tuned hyperparameter becomes consistent with the convergence rate of gradient descent, even though the problem to be solved is gradually changed due to $t$. In numerical experiments, our SLGH algorithms show faster convergence than an existing double loop GH method while outperforming gradient descent-based methods in terms of finding a better solution.

Hidenori Iwakiri, Yuhang Wang, Shinji Ito, Akiko Takeda• 2022

Related benchmarks

TaskDatasetResultRank
Black-box Adversarial AttackCIFAR-10 (test)
Success Rate93
11
Global OptimizationGriewank
MSE2.77
9
Global OptimizationAckley d = 500
MSE0.57
9
Black-Box Targeted Adversarial AttacksCIFAR-10
Success Rate (SR)100
9
Global OptimizationRosenbrock $d=500$
MSE0.4
9
Black-box Targeted Adversarial AttackVitalDB
Success Rate (SR)100
9
Zeroth-order optimizationTheoretical Objective Functions
Iteration Complexity2
8
Adversarial AttackMNIST 100 images
Success Rate96
5
Showing 8 of 8 rows

Other info

Follow for update