Machine Learning Coffee seminar: Alexander Ilin "Sample-Efficient Deep Learning"
Sample-Efficient Deep Learning
Professor of Computer Science, Aalto University
The talk will cover deep learning methods designed to learn efficiently from few labelled examples. I will present Ladder networks and its extensions for semi-supervised classification, recurrent Ladder networks for unsupervised and semi-supervised learning of video and algorithms of few-shot classification: learning to learn new classification tasks from few examples.