University of Cambridge > Talks.cam > Machine Learning Reading Group @ CUED > Active Learning

Active Learning

Add to your list(s) Download to your calendar using vCal

If you have a question about this talk, please contact jg801.

In many real-world problems such as medical diagnosis, labelling data is expensive and labelled datasets are scarce. Active learning describes a set of machine learning techniques that choose their own training data. The key idea behind active learning is to achieve greater accuracy with fewer training labels by allowing the ML algorithm to choose the data from which it learns. To this end, the active learning algorithm poses queries (i.e. unlabelled data instances) to be labelled by an oracle (e.g., a human annotator).

In this reading group session, we aim to provide a broad overview over both the foundations and frontiers of active learning. We will first define active learning, contrast it against other formulations of machine learning, and describe the standard techniques used to select queries. We will then survey recent advances in active learning, often in the context of (Bayesian) deep neural networks. We hope to conclude with an open-ended discussion on active learning.

This talk is part of the Machine Learning Reading Group @ CUED series.

Tell a friend about this talk:

This talk is included in these lists:

Note that ex-directory lists are not shown.

 

© 2006-2020 Talks.cam, University of Cambridge. Contact Us | Help and Documentation | Privacy and Publicity