Best Practices in Deep Learning and the Art of Research Management
Best Practices in Deep Learning and the Art of Research Management

Abstract: 

The participant will be provided with insights and guidance on how to use experiment management from research to production.

In this workshop, we share our experience from numerous deep learning projects. We will discuss topics such as hyperparameter optimization search, data biases, diminishing returns of annotated data and productive use of an experiment management platform.
In addition, we will dedicate the majority of the workshop to hands-on training using Trains - a free open-source experiment manager for machine learning and deep learning projects.

In this section, we will cover the following topics:
*Introduction to Trains - Installing and using trains and trains-agent,
*Advanced use of Trains - Experiments artifacts and hyperparameter search,
*From research to production with Trains - Managing the full life cycle of your project.

Setup:
Allegro Trains (Allegro Trains, open-source & free)
https://allegro.ai

Bio: 

Moses is a computer vision & deep learning guru and a serial entrepreneur. Moses has more than 20 years of experience making visionary technologies a reality. Prior to Allegro AI, Moses co-founded several start-ups in the computer vision and embedded processing spaces. Moses is an alum of the IDF’s most elite technology unit, has 40 patent and patent applications and 5 academic publications in his name. Moses holds an M.Sc. Cum Laude from Tel Aviv University in Computer Science.