Multi-modal InfoNCE minimizers with temperature optimization provably adapt to the intrinsic dimension of the shared latent variables when aligned maximal-information encoders exist.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
stat.ML 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Multi-modal contrastive learning adapts to intrinsic dimensions of shared latent variables
Multi-modal InfoNCE minimizers with temperature optimization provably adapt to the intrinsic dimension of the shared latent variables when aligned maximal-information encoders exist.