作者
Roland S Zimmermann, Yash Sharma, Steffen Schneider, Matthias Bethge, Wieland Brendel
发表日期
2021/2/17
期刊
International Conference on Machine Learning (ICML 2021)
简介
Contrastive learning has recently seen tremendous success in self-supervised learning. So far, however, it is largely unclear why the learned representations generalize so effectively to a large variety of downstream tasks. We here prove that feedforward models trained with objectives belonging to the commonly used InfoNCE family learn to implicitly invert the underlying generative model of the observed data. While the proofs make certain statistical assumptions about the generative model, we observe empirically that our findings hold even if these assumptions are severely violated. Our theory highlights a fundamental connection between contrastive learning, generative modeling, and nonlinear independent component analysis, thereby furthering our understanding of the learned representations as well as providing a theoretical foundation to derive more effective contrastive losses.
引用总数
20202021202220232024115637843
学术搜索中的文章
RS Zimmermann, Y Sharma, S Schneider, M Bethge… - International Conference on Machine Learning, 2021