In recent years, artificial intelligence has made remarkable progress in developing algorithms that can learn from vast amounts of carefully labeled data. This paradigm of supervised learning has made great success in training specialist models that perform extremely well on the task they were trained to do, but there’s a limit to building more intelligent, generalized models without a massive amount of labeled data.
Self-supervised learning is an unsupervised learning method where the supervised learning task is created out of the unlabeled input data. It enables machines to understand new concepts quickly after seeing only a few examples that are labeled.
In this seminar, we will briefly review the basic concepts of deep learning and computer vision, and explain the recent development of self-supervised learning methods, focusing on representation learning in the field of computer vision.