Alexander Obiednikov
21 Sep 2019
15:00 - 15:45
Data Science and Machine Learning

Active Learning

As Wikipedia says, «active learning» is a specific case of machine learning in which a learning algorithm can interactively query the user (or some other information source) to obtain the desired outputs at new data points. In other words imagine that your model could tell you something like «hey, I’m not sure what’s that; could you please help me to specify?» or imagine that a model could select what specific data it needs additionally to improve its performance. Active learning approaches study how to build such algorithms, and I’m going to give an overview of the methods, where and how they can be applied in practice.