Unsupervised representation learning aims at describing raw data efficiently
to solve various downstream tasks. It has been approached with many techniques,
such as manifold learning, diffusion maps, or more recently self-supervised
learning. Those techniques are arguably all based on the underlying assumption
that target functions, associated with future downstream tasks, have low
variations in densely populated regions of the input space. Unveiling minimal
variations as a guiding principle behind unsupervised representation learning
paves the way to better practical guidelines for self-supervised learning
algorithms.
Description
On minimal variations for unsupervised representation learning
%0 Generic
%1 cabannes2022minimal
%A Cabannes, Vivien
%A Bietti, Alberto
%A Balestriero, Randall
%D 2022
%K ai self-supervised
%T On minimal variations for unsupervised representation learning
%U http://arxiv.org/abs/2211.03782
%X Unsupervised representation learning aims at describing raw data efficiently
to solve various downstream tasks. It has been approached with many techniques,
such as manifold learning, diffusion maps, or more recently self-supervised
learning. Those techniques are arguably all based on the underlying assumption
that target functions, associated with future downstream tasks, have low
variations in densely populated regions of the input space. Unveiling minimal
variations as a guiding principle behind unsupervised representation learning
paves the way to better practical guidelines for self-supervised learning
algorithms.
@misc{cabannes2022minimal,
abstract = {Unsupervised representation learning aims at describing raw data efficiently
to solve various downstream tasks. It has been approached with many techniques,
such as manifold learning, diffusion maps, or more recently self-supervised
learning. Those techniques are arguably all based on the underlying assumption
that target functions, associated with future downstream tasks, have low
variations in densely populated regions of the input space. Unveiling minimal
variations as a guiding principle behind unsupervised representation learning
paves the way to better practical guidelines for self-supervised learning
algorithms.},
added-at = {2022-12-12T15:57:56.000+0100},
author = {Cabannes, Vivien and Bietti, Alberto and Balestriero, Randall},
biburl = {https://www.bibsonomy.org/bibtex/24ccedc36556a8c758684703d11fac63a/tcorbettclark},
description = {On minimal variations for unsupervised representation learning},
interhash = {34b14046c21a54407f5295a37c6640a1},
intrahash = {4ccedc36556a8c758684703d11fac63a},
keywords = {ai self-supervised},
note = {cite arxiv:2211.03782Comment: 5 pages, 1 figure; 1 table},
timestamp = {2022-12-12T15:57:56.000+0100},
title = {On minimal variations for unsupervised representation learning},
url = {http://arxiv.org/abs/2211.03782},
year = 2022
}