Representation Selection via Cross-Model Agreement using Canonical Correlation Analysis

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Lewis, Dylan B., Gregor, Jens, Santos-Villalobos, Hector
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917377577320448
author Lewis, Dylan B.
Gregor, Jens
Santos-Villalobos, Hector
author_facet Lewis, Dylan B.
Gregor, Jens
Santos-Villalobos, Hector
contents Modern vision pipelines increasingly rely on pretrained image encoders whose representations are reused across tasks and models, yet these representations are often overcomplete and model-specific. We propose a simple, training-free method to improve the efficiency of image representations via a post-hoc canonical correlation analysis (CCA) operator. By leveraging the shared structure between representations produced by two pre-trained image encoders, our method finds linear projections that serve as a principled form of representation selection and dimensionality reduction, retaining shared semantic content while discarding redundant dimensions. Unlike standard dimensionality reduction techniques such as PCA, which operate on a single embedding space, our approach leverages cross-model agreement to guide representation distillation and refinement. The technique allows representations to be reduced by more than 75% in dimensionality with improved downstream performance, or enhanced at fixed dimensionality via post-hoc representation transfer from larger or fine-tuned models. Empirical results on ImageNet-1k, CIFAR-100, MNIST, and additional benchmarks show consistent improvements over both baseline and PCA-projected representations, with accuracy gains of up to 12.6%.
format Preprint
id arxiv_https___arxiv_org_abs_2604_00921
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Representation Selection via Cross-Model Agreement using Canonical Correlation Analysis
Lewis, Dylan B.
Gregor, Jens
Santos-Villalobos, Hector
Computer Vision and Pattern Recognition
Artificial Intelligence
68T01
I.4.10; I.4.2; I.2.4
Modern vision pipelines increasingly rely on pretrained image encoders whose representations are reused across tasks and models, yet these representations are often overcomplete and model-specific. We propose a simple, training-free method to improve the efficiency of image representations via a post-hoc canonical correlation analysis (CCA) operator. By leveraging the shared structure between representations produced by two pre-trained image encoders, our method finds linear projections that serve as a principled form of representation selection and dimensionality reduction, retaining shared semantic content while discarding redundant dimensions. Unlike standard dimensionality reduction techniques such as PCA, which operate on a single embedding space, our approach leverages cross-model agreement to guide representation distillation and refinement. The technique allows representations to be reduced by more than 75% in dimensionality with improved downstream performance, or enhanced at fixed dimensionality via post-hoc representation transfer from larger or fine-tuned models. Empirical results on ImageNet-1k, CIFAR-100, MNIST, and additional benchmarks show consistent improvements over both baseline and PCA-projected representations, with accuracy gains of up to 12.6%.
title Representation Selection via Cross-Model Agreement using Canonical Correlation Analysis
topic Computer Vision and Pattern Recognition
Artificial Intelligence
68T01
I.4.10; I.4.2; I.2.4
url https://arxiv.org/abs/2604.00921