6 PyData Athens Meetup: Automated Product Categorization using Deep Learning

My fellow PyDataneers,

We would like to wish you all a happy new year! May all your wishes and dreams come true in 2019!

For the first PyData of this year, we have a special guest in store for you! We feel honoured to host Andreas Loupasakis, who will be talking to us about his creative work on Automated Product Categorization at Skroutz.

6:00-6:30 Networking time

6:30-7:30 Automated Product Categorization At Skroutz, Andreas Loupasakis, Senior Software Engineer at Skroutz.

(https://www.kaggle.com/andreaslup,
https://github.com/alup,
https://twitter.com/andy_lupo
https://www.linkedin.com/in/andreas-loupasakis-06399a47 )

Abstract

A brief description of how we use Deep Learning at Skroutz to categorize thousands of products on a daily basis. Skroutz indexes millions of products in its search engine, dispatched by thousands of e-shops. It is mission critical to ensure these products are matched to the most relevant categories in the category tree.

Last year we released a new version of our category classifier called "Megatron" which significantly improved our automated categorization accuracy.

We are going to take a look at the building blocks of our categorization service. Infrastructure, open source libraries, training procedures, and neural network architectures are some of the key points of this presentation.

7.30-8.15 Presentation to be announced.

All skill levels are welcome!