4 papers
Self-concordant smoothing in proximal quasi-Newton algorithms for large-scale convex composite optimization
Adeyemi D. Adeoye, Alberto Bemporad
We introduce a notion of self-concordant smoothing for minimizing the sum of two convex functions, one of which is smooth and the other nonsmooth. The key highlight is a natural pr…
Exact Gauss-Newton Optimization for Training Deep Neural Networks
Mikalai Korbit, Adeyemi D. Adeoye, Alberto Bemporad +1
We present Exact Gauss-Newton (EGN), a stochastic second-order optimization algorithm that combines the generalized Gauss-Newton (GN) Hessian approximation with low-rank linear alg…
A proximal augmented Lagrangian method for nonconvex optimization with equality and inequality constraints
Adeyemi D. Adeoye, Puya Latafat, Alberto Bemporad
We propose an inexact proximal augmented Lagrangian method (P-ALM) for nonconvex structured optimization problems. The proposed method features an easily implementable rule not onl…
Regularized Gauss-Newton for Optimizing Overparameterized Neural Networks
Adeyemi D. Adeoye, Philipp Christian Petersen, Alberto Bemporad
The generalized Gauss-Newton (GGN) optimization method incorporates curvature estimates into its solution steps, and provides a good approximation to the Newton method for large-sc…