Detection and Classification in Latent Spaces: High-Dimensional Analysis and Validation
Organizations: *NATO STO CMRE
Abstract
Foundation models produce high-dimensional latent representations in which suitably designed statistical decision rules can be effective yet analytically tractable. We study such detectors in the high-dimensional regime, where latent dimension and training sample size grow proportionally. For linear and quadratic statistics, we derive explicit large-deviation characterizations of false-alarm and miss-detection performance under the Neyman-Pearson paradigm, and show how the binary analysis extends to multi-class classification. The resulting error probabilities and corresponding decay rates exhibit non-monotonic dependence on the shape parameter, including double-descent and, in some regimes, multiple-descent. Experiments on vision, text, and audio datasets using modern foundation models show close agreement between theoretical predictions and empirical performance. In particular, the proposed framework allows us to set a prescribed false-alarm level and accurately predict the corresponding miss-detection probability. Overall, our results provide a theoretically grounded and interpretable framework with competitive empirical performance across different modalities.
Figures & tables
Appendix figures & tables15 assets
Supplementary material from the paper’s appendix.
Appendix
| Hypothesis | Tweet | Humanitarian Category |
|---|---|---|
| Flash Flooding Threatens Towns Devastated by Hurricane Maria. #Weather #PuertoRico: via @YouTube | Caution and advice | |
| Louisiana – Hurricane categories tell only part of the story — here’s the damage storms like Irma and Harvey can do | Other relevant information | |
| Community Blood Center needs donations to prepare for Hurricane Dorian via @ndn | Requests or urgent needs | |
| Drone video: St. Augustine condos destroyed by Hurricane Irma - Droneblog - #drones | Infrastructure and utility damage |
Classification
Classification maps observations to discrete labels, including familiar, unknown, hierarchical, or multiple categories.
See all 451 papers