Timezone: »

 
Poster
Self-supervised Label Augmentation via Input Transformations
Hankook Lee · Sung Ju Hwang · Jinwoo Shin

Wed Jul 15 10:00 AM -- 10:45 AM & Wed Jul 15 11:00 PM -- 11:45 PM (PDT) @ Virtual

Self-supervised learning, which learns by constructing artificial labels given only the input signals, has recently gained considerable attention for learning representations with unlabeled datasets, i.e., learning without any human-annotated supervision. In this paper, we show that such a technique can be used to significantly improve the model accuracy even under fully-labeled datasets. Our scheme trains the model to learn both original and self-supervised tasks, but is different from conventional multi-task learning frameworks that optimize the summation of their corresponding losses. Our main idea is to learn a single unified task with respect to the joint distribution of the original and self-supervised labels, i.e., we augment original labels via self-supervision. This simple, yet effective approach allows to train models easier by relaxing a certain invariant constraint during learning the original and self-supervised tasks simultaneously. It also enables an aggregated inference which combines the predictions from different augmentations to improve the prediction accuracy. Furthermore, we propose a novel knowledge transfer technique, which we refer to as self-distillation, that has the effect of the aggregated inference in a single (faster) inference. We demonstrate the large accuracy improvement and wide applicability of our framework on various fully-supervised settings, e.g., the few-shot and imbalanced classification scenarios.

Author Information

Hankook Lee (KAIST)
Sung Ju Hwang (KAIST, AITRICS)
Jinwoo Shin (KAIST)

More from the Same Authors