pith. sign in

arxiv: 2105.03746 · v4 · pith:73ZTN7Q2new · submitted 2021-05-08 · 💻 cs.LG · cs.CV· stat.ML

Contrastive Attraction and Contrastive Repulsion for Representation Learning

classification 💻 cs.LG cs.CVstat.ML
keywords negativepositivecontrastivesamplesdatasetswhenattractioncacr
0
0 comments X
read the original abstract

Contrastive learning (CL) methods effectively learn data representations in a self-supervision manner, where the encoder contrasts each positive sample over multiple negative samples via a one-vs-many softmax cross-entropy loss. By leveraging large amounts of unlabeled image data, recent CL methods have achieved promising results when pretrained on large-scale datasets, such as ImageNet. However, most of them consider the augmented views from the same instance are positive pairs, while views from other instances are negative ones. Such binary partition insufficiently considers the relation between samples and tends to yield worse performance when generalized on images in the wild. In this paper, to further improve the performance of CL and enhance its robustness on various datasets, {we propose a doubly CL strategy that separately compares positive and negative samples within their own groups, and then proceeds with a contrast between positive and negative groups}. We realize this strategy with contrastive attraction and contrastive repulsion (CACR), which makes the query not only exert a greater force to attract more distant positive samples but also do so to repel closer negative samples. Theoretical analysis reveals that CACR generalizes CL's behavior by positive attraction and negative repulsion, and it further considers the intra-contrastive relation within the positive and negative pairs to narrow the gap between the sampled and true distribution, which is important when datasets are less curated. With our extensive experiments, CACR not only demonstrates good performance on CL benchmarks, but also shows better robustness when generalized on imbalanced image datasets. Code and pre-trained checkpoints are available at https://github.com/JegZheng/CACR-SSL.

This paper has not been read by Pith yet.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 2 Pith papers

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

  1. FuTCR: Future-Targeted Contrast and Repulsion for Continual Panoptic Segmentation

    cs.CV 2026-05 unverdicted novelty 7.0

    FuTCR boosts new-class panoptic quality by up to 28% in continual settings by discovering future-like background regions and applying targeted contrast and repulsion to reserve space for incoming categories.

  2. FuTCR: Future-Targeted Contrast and Repulsion for Continual Panoptic Segmentation

    cs.CV 2026-05 unverdicted novelty 6.0

    FuTCR improves new-class panoptic quality by up to 28% in continual panoptic segmentation by discovering future-like regions in background areas and applying targeted contrast and repulsion to restructure representations.