AI Talk: Black box Optimization for Machine Learning

3e-4! Some of you might know this as the magic learning rate for optimizing deep neural networks.

Making design choices about NNs and training them is an art and can be frustrating with a lot of trial and error.

So why not just automate it? In recent years black box optimization tool for hyperparameters in deep learning has become popular and helps engineers save time and compute resources.

As a part of the deep learning course at the Technical University of Denmark (DTU) Tobias Andreassen will present how to get started with black box optimization and showcase a recent toolbox on deep neural networks.

About Tobias Andreassen
https://www.linkedin.com/in/tobias-andreasen-84364091/
Tobias Andreassen has working multiple years in Silicon Valley doing black box optimization with some of the worlds largest tech companies.