Timezone: »

Learning To Optimize
Nando de Freitas

Thu Dec 08 11:30 PM -- 12:00 AM (PST) @ None

The move from hand-designed features to learned features in machine learning has been wildly successful. In spite of this, optimization algorithms are still designed by hand. In this talk I describe how the design of an optimization algorithm can be cast as a learning problem, allowing the algorithm to learn to exploit structure in the problems of interest in an automatic way. The learned algorithms, implemented by LSTMs, outperform generic, hand-designed competitors on the tasks for which they are trained, and also generalize well to new tasks with similar structure.

Author Information

Nando de Freitas (DeepMind)

More from the Same Authors