Dataset Awareness is not Enough: Implementing Sample-level Tail Encouragement in Long-tailed Self-supervised Learning

TL;DR

This paper introduces TASE, a novel method that uses pseudo-labels and dynamic temperature adjustments at the sample level to improve self-supervised learning on long-tailed datasets, enhancing recognition and robustness.

Contribution

The paper proposes TASE, a new approach that assigns optimal, sample-specific temperature parameters and re-weighting strategies to better guide training in long-tailed SSL.

Findings

Significant improvements in long-tail recognition across six benchmarks.

Enhanced robustness in self-supervised learning on imbalanced datasets.

Outperforms existing methods in handling long-tailed distributions.

Abstract

Self-supervised learning (SSL) has shown remarkable data representation capabilities across a wide range of datasets. However, when applied to real-world datasets with long-tailed distributions, performance on multiple downstream tasks degrades significantly. Recently, the community has begun to focus more on self-supervised long-tailed learning. Some works attempt to transfer temperature mechanisms to self-supervised learning or use category-space uniformity constraints to balance the representation of different categories in the embedding space to fight against long-tail distributions. However, most of these approaches focus on the joint optimization of all samples in the dataset or on constraining the category distribution, with little attention given to whether each individual sample is optimally guided during training. To address this issue, we propose Temperature Auxiliary Sample-level Encouragement (TASE). We introduce pseudo-labels into self-supervised long-tailed learning, utilizing pseudo-label information to drive a dynamic temperature and re-weighting strategy. Specifically, We assign an optimal temperature parameter to each sample. Additionally, we analyze the lack of quantity awareness in the temperature parameter and use re-weighting to compensate for this deficiency, thereby achieving optimal training patterns at the sample level. Comprehensive experimental results on six benchmarks across three datasets demonstrate that our method achieves outstanding performance in improving long-tail recognition, while also exhibiting high robustness.