A claimed generalization bound for deep contrastive learning that is independent of negative-sample count up to logs, but the main theorems as written contain a factor error making them grow with sample size.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
stat.ML 1years
2024 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Generalization Analysis for Deep Contrastive Representation Learning
A claimed generalization bound for deep contrastive learning that is independent of negative-sample count up to logs, but the main theorems as written contain a factor error making them grow with sample size.