CompRess: Self-Supervised Learning by Compressing Representations
This addresses the challenge of making self-supervised learning more efficient for resource-constrained applications, though it is incremental as it builds on existing compression and self-supervised techniques.
The paper tackles the problem of compressing deep self-supervised models to smaller ones, achieving a self-supervised AlexNet that outperforms the supervised version on ImageNet classification with 59.0% vs. 56.5% in linear evaluation and 50.7% vs. 41.4% in nearest neighbor evaluation.
Self-supervised learning aims to learn good representations with unlabeled data. Recent works have shown that larger models benefit more from self-supervised learning than smaller models. As a result, the gap between supervised and self-supervised learning has been greatly reduced for larger models. In this work, instead of designing a new pseudo task for self-supervised learning, we develop a model compression method to compress an already learned, deep self-supervised model (teacher) to a smaller one (student). We train the student model so that it mimics the relative similarity between the data points in the teacher's embedding space. For AlexNet, our method outperforms all previous methods including the fully supervised model on ImageNet linear evaluation (59.0% compared to 56.5%) and on nearest neighbor evaluation (50.7% compared to 41.4%). To the best of our knowledge, this is the first time a self-supervised AlexNet has outperformed supervised one on ImageNet classification. Our code is available here: https://github.com/UMBCvision/CompRess