Self-supervised learning is a machine learning paradigm that uses unlabeled data to obtain useful representations for downstream tasks. It consists of two steps: generating pseudo-labels and then performing supervised or unsupervised learning. It has been used in audio processing and speech recognition, and more closely imitates the way humans learn to classify objects.
Stanford University
Spring 2022
This is a deep-dive into the details of deep learning architectures for visual recognition tasks. The course provides students with the ability to implement, train their own neural networks and understand state-of-the-art computer vision research. It requires Python proficiency and familiarity with calculus, linear algebra, probability, and statistics.
No concepts data
+ 55 more concepts