Fast convex optimization for two-layer relu networks: Equivalent model classes and cone decompositions

A Mishkin, A Sahiner, M Pilanci - … Conference on Machine …, 2022‏ - proceedings.mlr.press
We develop fast algorithms and robust software for convex optimization of two-layer neural
networks with ReLU activation functions. Our work leverages a convex re-formulation of the …

Evaluation methodology for deep learning imputation models

O Boursalie, R Samavi… - Experimental Biology and …, 2022‏ - journals.sagepub.com
There is growing interest in imputing missing data in tabular datasets using deep learning.
Existing deep learning–based imputation models have been commonly evaluated using root …

[PDF][PDF] Globally convergent derivative-free methods in nonconvex optimization with and without noise

PD Khanh, BS Mordukhovich, DB Tran - 2024‏ - optimization-online.org
This paper addresses the study of nonconvex derivative-free optimization problems, where
only information of either smooth objective functions or their noisy approximations is …

Interpolation, growth conditions, and stochastic gradient descent

A Mishkin - 2020‏ - open.library.ubc.ca
Current machine learning practice requires solving huge-scale empirical risk minimization
problems quickly and robustly. These problems are often highly under-determined and …

Glocal Smoothness: Line Search can really help!

C Fox, M Schmidt - OPT 2024: Optimization for Machine Learning‏ - openreview.net
Iteration complexities are bounds on the number of iterations of an algorithm. Iteration
complexities for first-order numerical optimization algorithms are typically stated in terms of a …

Evaluation methodology for deep learning imputation models

R Samavi, O Boursalie, TE Doyle‏ - rshare.library.torontomu.ca
There is growing interest in imputing missing data in tabular datasets using deep learning.
Existing deep learning–based imputation models have been commonly evaluated using root …