Representation Learning of Visual Features Through Self-Supervision: Recipes to learn representations from unlabeled data

Presenter: Thalles Silva, Ph.D candidate and Research Scientist, University of Campinas, São Paulo, Brazil

This presentation dives into state-of-the-art techniques for learning general-purpose representations from unlabeled images. We will explore the primary paradigms of self-supervised learning (SSL) methods, detailing their core components and limitations. Additionally, we will review proposed techniques aimed at enhancing current methods. By the conclusion of this talk, viewers will have a fair understanding of contemporary SSL techniques for training deep learning models using unlabeled data, leveraging self-supervision as the source of the training signal.

In compliance with GDPR consent requirements, presentations given in a Visual Intelligence context may be recorded with the consent of the speaker. All recordings are edited to remove all faces, names and voices of other participants. Questions and comments by the audience will hence be removed and will not appear in the recording.  With the freely given consent from the speaker, recorded presentation may be posted on the Visual Intelligence YouTube channel.

‍