Introducing Curvature to the Label Space
View / Open Files
Authors
Sheehan, Conor
Day, Ben
Liò, Pietro
Publication Date
2018-10-22Journal Title
CoRR
Type
Article
This Version
AM
Metadata
Show full item recordCitation
Sheehan, C., Day, B., & Liò, P. (2018). Introducing Curvature to the Label Space. CoRR https://doi.org/10.17863/CAM.40293
Abstract
One-hot encoding is a labelling system that embeds classes as standard basis
vectors in a label space. Despite seeing near-universal use in supervised
categorical classification tasks, the scheme is problematic in its geometric
implication that, as all classes are equally distant, all classes are equally
different. This is inconsistent with most, if not all, real-world tasks due to
the prevalence of ancestral and convergent relationships generating a varying
degree of morphological similarity across classes. We address this issue by
introducing curvature to the label-space using a metric tensor as a
self-regulating method that better represents these relationships as a bolt-on,
learning-algorithm agnostic solution. We propose both general constraints and
specific statistical parameterizations of the metric and identify a direction
for future research using autoencoder-based parameterizations.
Keywords
cs.LG, cs.LG, stat.ML
Embargo Lift Date
2100-01-01
Identifiers
This record's DOI: https://doi.org/10.17863/CAM.40293
This record's URL: https://www.repository.cam.ac.uk/handle/1810/293144
Rights
All rights reserved
Licence:
http://www.rioxx.net/licenses/all-rights-reserved
Statistics
Total file downloads (since January 2020). For more information on metrics see the
IRUS guide.