Rethinking the Good Enough Embedding for Easy Few-Shot Learning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Karnes, Michael, Yilmaz, Alper
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909041408606208
author Karnes, Michael
Yilmaz, Alper
author_facet Karnes, Michael
Yilmaz, Alper
contents The field of deep visual recognition is undergoing a paradigm shift toward universal representations. The Platonic Representation Hypothesis suggests that diverse architectures trained on massive datasets are converging toward a shared, "ideal" latent space. This again raises a critical question: is a "Good Embedding All You Need?" In this paper, we leverage this convergence to demonstrate that off-the-shelf embeddings are inherently "good enough" for complex tasks, rendering intensive task-specific fine-tuning unnecessary. We explore this hypothesis within the few-shot learning framework, proposing a straightforward, non-parametric pipeline that entirely bypasses backpropagation. By utilizing a k-Nearest Neighbor classifier on frozen DINOv2-L features, we conduct a layer-wise characterization to identify an optimal feature extraction. We further demonstrate that manifold refinement via PCA and ICA provides a beneficial regularizing effect. Our results across four major benchmarks demonstrate that our approach consistently surpasses sophisticated meta-learning algorithms, achieving state-of-the-art performance.
format Preprint
id arxiv_https___arxiv_org_abs_2605_14145
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Rethinking the Good Enough Embedding for Easy Few-Shot Learning
Karnes, Michael
Yilmaz, Alper
Computer Vision and Pattern Recognition
The field of deep visual recognition is undergoing a paradigm shift toward universal representations. The Platonic Representation Hypothesis suggests that diverse architectures trained on massive datasets are converging toward a shared, "ideal" latent space. This again raises a critical question: is a "Good Embedding All You Need?" In this paper, we leverage this convergence to demonstrate that off-the-shelf embeddings are inherently "good enough" for complex tasks, rendering intensive task-specific fine-tuning unnecessary. We explore this hypothesis within the few-shot learning framework, proposing a straightforward, non-parametric pipeline that entirely bypasses backpropagation. By utilizing a k-Nearest Neighbor classifier on frozen DINOv2-L features, we conduct a layer-wise characterization to identify an optimal feature extraction. We further demonstrate that manifold refinement via PCA and ICA provides a beneficial regularizing effect. Our results across four major benchmarks demonstrate that our approach consistently surpasses sophisticated meta-learning algorithms, achieving state-of-the-art performance.
title Rethinking the Good Enough Embedding for Easy Few-Shot Learning
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2605.14145