English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Contrastive learning has recently seen tremendous success in self-supervisedlearning. So far, however, it is largely unclear why the learnedrepresentations generalize so effectively to a large variety of downstreamtasks. We here prove that feedforward models trained with objectives belongingto the commonly used InfoNCE family learn to implicitly invert the underlyinggenerative model of the observed data. While the proofs make certainstatistical assumptions about the generative model, we observe empirically thatour findings hold even if these assumptions are severely violated. Our theoryhighlights a fundamental connection between contrastive learning, generativemodeling, and nonlinear independent component analysis, thereby furthering ourunderstanding of the learned representations as well as providing a theoreticalfoundation to derive more effective contrastive losses.

Contrastive Learning Inverts the Data Generating Process