Pith. sign in

REVIEW 1 cited by

Supervision Accelerates Pre-training in Contrastive Semi-Supervised Learning of Visual Representations

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2006.10803 v2 pith:TQJOMHPJ submitted 2020-06-18 cs.LG cs.CVstat.ML

classification cs.LGcs.CVstat.ML
keywords contrastivelearningpre-trainingamountduringrepresentationssemi-supervisedsuncet
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

We investigate a strategy for improving the efficiency of contrastive learning of visual representations by leveraging a small amount of supervised information during pre-training. We propose a semi-supervised loss, SuNCEt, based on noise-contrastive estimation and neighbourhood component analysis, that aims to distinguish examples of different classes in addition to the self-supervised instance-wise pretext tasks. On ImageNet, we find that SuNCEt can be used to match the semi-supervised learning accuracy of previous contrastive approaches while using less than half the amount of pre-training and compute. Our main insight is that leveraging even a small amount of labeled data during pre-training, and not only during fine-tuning, provides an important signal that can significantly accelerate contrastive learning of visual representations. Our code is available online at github.com/facebookresearch/suncet.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Mitigating Degree Bias Adaptively with Hard-to-Learn Nodes in Graph Contrastive Learning

    cs.LG 2025-06 conditional novelty 5.0 of 10

    A hardness-aware reweighted contrastive loss that uses labels and similarity to upweight hard positives and negatives reduces degree bias in graph node classification.

Pith tools