Tailoring Self-Supervision for Supervised Learning

TL;DR

This paper proposes a new self-supervision task called Localizable Rotation (LoRot) designed to improve supervised learning by guiding feature learning, maintaining data distribution, and being lightweight and versatile.

Contribution

It introduces the LoRot task, tailored for supervised learning, fulfilling key properties for effective auxiliary self-supervision, and demonstrates its benefits through extensive experiments.

Findings

LoRot enhances robustness and generalization in supervised models.

LoRot outperforms existing self-supervision tasks in experiments.

The proposed method is lightweight and broadly applicable.

Abstract

Recently, it is shown that deploying a proper self-supervision is a prospective way to enhance the performance of supervised learning. Yet, the benefits of self-supervision are not fully exploited as previous pretext tasks are specialized for unsupervised representation learning. To this end, we begin by presenting three desirable properties for such auxiliary tasks to assist the supervised objective. First, the tasks need to guide the model to learn rich features. Second, the transformations involved in the self-supervision should not significantly alter the training distribution. Third, the tasks are preferred to be light and generic for high applicability to prior arts. Subsequently, to show how existing pretext tasks can fulfill these and be tailored for supervised learning, we propose a simple auxiliary self-supervision task, predicting localizable rotation (LoRot). Our exhaustive experiments validate the merits of LoRot as a pretext task tailored for supervised learning in terms of robustness and generalization capability. Our code is available at https://github.com/wjun0830/Localizable-Rotation.