Amos Storkey (University of Edinburgh)
Wed 02 Nov 2016, 16:00 - 17:00
JCMB 5327

If you have a question about this talk, please contact: Kostas Zygalakis (kzygalak)

When tackling common forms of real world problems, we quickly meet a number of constraints that make machine learning an interesting challenge. Here I give three very different examples:

Handling non-analyticity of finite-time transitions in stochastic processes.
Learning and inference in generative models and fair representations.
Small shot learning through learning transfer.

I introduce the problems and the methods that can be used, and discuss the open practical and mathematical challenges.