Atlas Based Representation and Metric Learning on Manifolds

06/13/2021
by   Eric O. Korman, et al.
0

We explore the use of a topological manifold, represented as a collection of charts, as the target space of neural network based representation learning tasks. This is achieved by a simple adjustment to the output of an encoder's network architecture plus the addition of a maximal mean discrepancy (MMD) based loss function for regularization. Most algorithms in representation and metric learning are easily adaptable to our framework and we demonstrate its effectiveness by adjusting SimCLR (for representation learning) and standard triplet loss training (for metric learning) to have manifold encoding spaces. Our experiments show that we obtain a substantial performance boost over the baseline for low dimensional encodings. In the case of triplet training, we also find, independent of the manifold setup, that the MMD loss alone (i.e. keeping a flat, euclidean target space but using an MMD loss to regularize it) increases performance over the baseline in the typical, high-dimensional Euclidean target spaces. Code for reproducing experiments is provided at https://github.com/ekorman/neurve .

READ FULL TEXT

page 12

page 13

page 14

research
02/05/2019

Metric Learning on Manifolds

Recent literature has shown that symbolic data, such as text and graphs,...
research
08/15/2016

Cross Euclidean-to-Riemannian Metric Learning with Application to Face Recognition from Video

Riemannian manifolds have been widely employed for video representations...
research
02/13/2018

TVAE: Triplet-Based Variational Autoencoder using Metric Learning

Deep metric learning has been demonstrated to be highly effective in lea...
research
04/01/2018

The Structure Transfer Machine Theory and Applications

Representation learning is a fundamental but challenging problem, especi...
research
03/21/2022

Hyperbolic Vision Transformers: Combining Improvements in Metric Learning

Metric learning aims to learn a highly discriminative model encouraging ...
research
12/27/2018

Topological Constraints on Homeomorphic Auto-Encoding

When doing representation learning on data that lives on a known non-tri...
research
06/28/2018

Beyond One-hot Encoding: lower dimensional target embedding

Target encoding plays a central role when learning Convolutional Neural ...

Please sign up or login with your details

Forgot password? Click here to reset