Understanding generalization and optimization performance of deep CNNs

P Zhou, J Feng - International Conference on Machine …, 2018 - proceedings.mlr.press
This work aims to provide understandings on the remarkable success of deep convolutional
neural networks (CNNs) by theoretically analyzing their generalization performance and …

Uncertainty quantification for online learning and stochastic approximation via hierarchical incremental gradient descent

WJ Su, Y Zhu - arxiv preprint arxiv:1802.04876, 2018 - arxiv.org
Stochastic gradient descent (SGD) is an immensely popular approach for online learning in
settings where data arrives in a stream or data sizes are very large. However, despite an …

Local convergence properties of SAGA/Prox-SVRG and acceleration

C Poon, J Liang, C Schoenlieb - … Conference on Machine …, 2018 - proceedings.mlr.press
In this paper, we present a local convergence anal-ysis for a class of stochastic optimisation
meth-ods: the proximal variance reduced stochastic gradient methods, and mainly focus on …

Fast rates for bandit optimization with upper-confidence frank-wolfe

Q Berthet, V Perchet - Advances in Neural Information …, 2017 - proceedings.neurips.cc
We consider the problem of bandit optimization, inspired by stochastic optimization and
online learning problems with bandit feedback. In this problem, the objective is to minimize a …

Stochastic Composite Least-Squares Regression with Convergence Rate

N Flammarion, F Bach - Conference on Learning Theory, 2017 - proceedings.mlr.press
We consider the minimization of composite objective functions composed of the expectation
of quadratic functions and an arbitrary convex function. We study the stochastic dual …

Optimization of smooth functions with noisy observations: Local minimax rates

Y Wang, S Balakrishnan… - Advances in Neural …, 2018 - proceedings.neurips.cc
We consider the problem of global optimization of an unknown non-convex smooth function
with noisy zeroth-order feedback. We propose a local minimax framework to study the …

Higrad: Uncertainty quantification for online learning and stochastic approximation

WJ Su, Y Zhu - Journal of Machine Learning Research, 2023 - jmlr.org
Stochastic gradient descent (SGD) is an immensely popular approach for online learning in
settings where data arrives in a stream or data sizes are very large. However, despite an …

Optimization of smooth functions with noisy observations: Local minimax rates

Y Wang, S Balakrishnan, A Singh - IEEE Transactions on …, 2019 - ieeexplore.ieee.org
We consider the problem of global optimization of an unknown non-convex smooth function
with noisy zeroth-order feedback. We propose a local minimax framework to study the …

Asymptotic Behaviors and Phase Transitions in Projected Stochastic Approximation: A Jump Diffusion Approach

J Liang, Y Han, X Li, Z Zhang - arxiv preprint arxiv:2304.12953, 2023 - arxiv.org
In this paper we consider linearly constrained optimization problems and propose a loopless
projection stochastic approximation (LPSA) algorithm. It performs the projection with …

Stochastic approximation and least-squares regression, with applications to machine learning

N Flammarion - 2017 - theses.hal.science
Many problems in machine learning are naturally cast as the minimization of a smooth
function defined on a Euclidean space. For supervised learning, this includes least-squares …