Every major self-supervised visual learning method is presented as a special case of a single energy model built from view invariance, spatial prediction, and explicit anti-collapse regularization, which the paper argues are all necessary.
In: International conference on machine learning
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Three Necessary Principles for Self-Supervised Visual Representation Learning
Every major self-supervised visual learning method is presented as a special case of a single energy model built from view invariance, spatial prediction, and explicit anti-collapse regularization, which the paper argues are all necessary.