Semi-supervised learning made simple with self-supervised clustering

Enrico Fini, Pietro Astolfi, Alahari Karteek, Xavier Alameda-Pineda, J. Mairal +2 more
2/14/2026

No official code found. Best implementation: pietroastolfi/suave-daino (63 stars, last commit 2023-06-13).

Abstract

Self-supervised learning models have been shown to learn rich visual representations without requiring human annotations. However, in many real-world scenarios, labels are partially available, motivating a recent line of work on semi-supervised methods inspired by self-supervised principles. In this paper, we propose a conceptually simple yet empirically powerful approach to turn clustering-based self-supervised methods such as SwAV or DINO into semi-supervised learners. More precisely, we introduce a multi-task framework merging a supervised objective using ground-truth labels and a self-supervised objective relying on clustering assignments with a single cross-entropy loss. This approach may be interpreted as imposing the cluster centroids to be class prototypes. Despite its simplicity, we provide empirical evidence that our approach is highly effective and achieves state-of-the-art performance on CI-FAR100 and ImageNet.

View on arXivView PDF

Code Implementations(1)

pietroastolfi/suave-dainoCommunity100%

Code release for "Semi-supervised learning made simple with self-supervised clustering"

635Shell, PythonJun 12, 20233 years agoNOASSERTION

Similar Papers

Clustering Properties of Self-Supervised Learning

86% match1 repo32025

Supercm: Revisiting Clustering for Semi-Supervised Learning

85% match2025

Rethinking Deep Clustering Paradigms: Self-Supervision Is All You Need

83% match2025

An Empirical Study into Clustering of Unseen Datasets with Self-Supervised Encoders

83% match1 repo22024

Patch-Level Kernel Alignment for Dense Self-Supervised Learning

83% match2025

The Triad of Failure Modes and a Possible Way Out

83% match2023

Cite this paper

@article{fini2026semisupervised,
  title  = {Semi-supervised learning made simple with self-supervised clustering},
  author = {Enrico Fini and Pietro Astolfi and Alahari Karteek and Xavier Alameda-Pineda and J. Mairal and Moin Nabi and E. Ricci},
  year   = {2026},
  eprint = {2306.07483},
  archivePrefix = {arXiv},
  url    = {https://arxiv.org/pdf/2306.07483},
  journal = {CVPR 2023 2023}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2306.07483.svg)](https://codeofpaper.com/paper/2306.07483)

Discussion