An adaptive margin added to InfoNCE improves time series clustering metrics but hurts linear-probe classification, exposing a disconnect between clustering scores and downstream utility.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
eMargin: Revisiting Contrastive Learning with Margin-Based Separation
An adaptive margin added to InfoNCE improves time series clustering metrics but hurts linear-probe classification, exposing a disconnect between clustering scores and downstream utility.