OCTANE -- Optimal Control for Tensor-based Autoencoder Network Emergence: Explicit Case

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Khatri, Ratna, Kolshorn, Anthony, Olson, Colin, Antil, Harbir
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912579786375168
author Khatri, Ratna
Kolshorn, Anthony
Olson, Colin
Antil, Harbir
author_facet Khatri, Ratna
Kolshorn, Anthony
Olson, Colin
Antil, Harbir
contents This paper presents a novel, mathematically rigorous framework for autoencoder-type deep neural networks that combines optimal control theory and low-rank tensor methods to yield memory-efficient training and automated architecture discovery. The learning task is formulated as an optimization problem constrained by differential equations representing the encoder and decoder components of the network and the corresponding optimality conditions are derived via a Lagrangian approach. Efficient memory compression is enabled by approximating differential equation solutions on low-rank tensor manifolds using an adaptive explicit integration scheme. These concepts are combined to form OCTANE (Optimal Control for Tensor-based Autoencoder Network Emergence) -- a unified training framework that yields compact autoencoder architectures, reduces memory usage, and enables effective learning, even with limited training data. The framework's utility is illustrated with application to image denoising and deblurring tasks and recommendations regarding governing hyperparameters are provided.
format Preprint
id arxiv_https___arxiv_org_abs_2509_08169
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle OCTANE -- Optimal Control for Tensor-based Autoencoder Network Emergence: Explicit Case
Khatri, Ratna
Kolshorn, Anthony
Olson, Colin
Antil, Harbir
Optimization and Control
Machine Learning
34H05, 37N40, 49K15, 49M41, 65K10, 68T05, 65Z05
This paper presents a novel, mathematically rigorous framework for autoencoder-type deep neural networks that combines optimal control theory and low-rank tensor methods to yield memory-efficient training and automated architecture discovery. The learning task is formulated as an optimization problem constrained by differential equations representing the encoder and decoder components of the network and the corresponding optimality conditions are derived via a Lagrangian approach. Efficient memory compression is enabled by approximating differential equation solutions on low-rank tensor manifolds using an adaptive explicit integration scheme. These concepts are combined to form OCTANE (Optimal Control for Tensor-based Autoencoder Network Emergence) -- a unified training framework that yields compact autoencoder architectures, reduces memory usage, and enables effective learning, even with limited training data. The framework's utility is illustrated with application to image denoising and deblurring tasks and recommendations regarding governing hyperparameters are provided.
title OCTANE -- Optimal Control for Tensor-based Autoencoder Network Emergence: Explicit Case
topic Optimization and Control
Machine Learning
34H05, 37N40, 49K15, 49M41, 65K10, 68T05, 65Z05
url https://arxiv.org/abs/2509.08169