Joint image clustering and self-supervised representation learning through debiased contrastive loss
Bibliographic record
Abstract
Joint self-supervised representation learning and image clustering have emerged as some of the most effective techniques for visual representation learning. However, existing methods often rely on artificially balanced datasets, raising concerns about their performance on imbalanced and long-tail data distributions. To address this challenge, we propose a novel framework that combines debiased self-supervised representation learning with joint clustering. By adapting the debiased contrastive loss, our approach mitigates the under-clustering of minority classes in imbalanced datasets. Furthermore, integrating the debiased contrastive loss with a divergence clustering loss significantly improves the quality of learned representations. We conducted extensive experiments on diverse datasets, including CIFAR-10, CIFAR-100, iNaturalist-2018, ISIC-2018 (skin lesions), and two ophthalmic retina fundus glaucoma datasets. Our framework was compared against state-of-the-art methods such as SimCLR, SimSiam, Debiased, and BNN, as well as other self-supervised and clustering algorithms. The results demonstrate that our method outperforms existing deep clustering, self-supervised, and semi-supervised techniques across various classification and clustering tasks, particularly on imbalanced and clinical datasets. These findings establish the effectiveness of our framework for representation learning under challenging data distributions, offering new insights into addressing imbalances in real-world applications.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".