CRITICAL ANALYSIS OF SELF-SUPERVISION: Difference between revisions
Jump to navigation
Jump to search
| Line 3: | Line 3: | ||
== Introduction == | == Introduction == | ||
This paper aims to learn the deep features of CNNs without manual labels by using supervision techniques. | |||
The main idea of self-supervision approaches is to learn from unlabeled data and pre-train networks via pretext tasks that can be automatically generated from the data itself. | |||
== Previous Work == | == Previous Work == | ||
Revision as of 01:35, 25 November 2020
Presented by
Maral Rasoolijaberi
Introduction
This paper aims to learn the deep features of CNNs without manual labels by using supervision techniques. The main idea of self-supervision approaches is to learn from unlabeled data and pre-train networks via pretext tasks that can be automatically generated from the data itself.