Decodable but Not Clusterable: Coarse Supervision, Within-Class Geometry, and a Label-Free Diagnostic for Recoverable Subclass Order

Abstract Networks trained with coarse labels often keep fine-label information that a linear probe can read but that clustering cannot recover. We give this dissociation a mechanism, a test and a remedy. Within a coarse class, the two-means optimum is captured by the leading nuisance eigendirection once the ratio of that eigenvalue to the semantic variance, the spectral dose , passes a crossover between 1 and π/2; beyond π the optimum is provably blind to the fine label. Within-class whitening removes this phase transition and recovers the Bayes semantic split. A label-free monitor, the k-means gain relative to a Gaussian with the same covariance, is bounded by 1 + π/(2·dose) and yields a repair rule that needs no fine labels. Across seven settings and 647 class snapshots, cross-entropy fine-tuning from in-domain self-supervised bases raised the dose and cut clusterability by 59–63% while the probe stayed flat; ImageNet and DINOv2 bases lowered the dose and gained. The monitor tracked clusterability changes (Spearman 0.77), and whitening recovered 48–59% of the loss on CIFAR-100 and all of it on Fashion-MNIST. The rule still helped when the number of clusters was also estimated, whereas covariance-adaptive clustering stayed trapped.

Authors

Publication Details

Journal
Research Square
Published
2026-11-13
DOI
https://doi.org/10.21203/rs.3.rs-11708078/v1
Primary Topic
Domain Adaptation and Few-Shot Learning
Type
preprint

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Decodable but Not Clusterable: Coarse Supervision, Within-Class Geometry, and a Label-Free Diagnostic for Recoverable Subclass Order

Volkan Oban
1 links
Research Square
Domain Adaptation and Few-Shot Learning
preprint

Decodable but Not Clusterable: Coarse Supervision, Within-Class Geometry, and a Label-Free Diagnostic for Recoverable Subclass Order

Volkan Oban
preprint en

Abstract

Abstract Networks trained with coarse labels often keep fine-label information that a linear probe can read but that clustering cannot recover. We give this dissociation a mechanism, a test and a remedy. Within a coarse class, the two-means optimum is captured by the leading nuisance eigendirection once the ratio of that eigenvalue to the semantic variance, the spectral dose , passes a crossover between 1 and π/2; beyond π the optimum is provably blind to the fine label. Within-class whitening removes this phase transition and recovers the Bayes semantic split. A label-free monitor, the k-means gain relative to a Gaussian with the same covariance, is bounded by 1 + π/(2·dose) and yields a repair rule that needs no fine labels. Across seven settings and 647 class snapshots, cross-entropy fine-tuning from in-domain self-supervised bases raised the dose and cut clusterability by 59–63% while the probe stayed flat; ImageNet and DINOv2 bases lowered the dose and gained. The monitor tracked clusterability changes (Spearman 0.77), and whitening recovered 48–59% of the loss on CIFAR-100 and all of it on Fashion-MNIST. The rule still helped when the number of clusters was also estimated, whereas covariance-adaptive clustering stayed trapped.

Research Square
Nvidia
Domain Adaptation and Few-Shot Learning
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon