Migrating AngularJS to Angular, Graph neural networks to the rescue

DESCRIPTION
We are going to explore the possibilities to automate the migration from AngularJS to Angular. Starting out with a simple regular expression approach, we will quickly run into limitations and discover the advantages of manipulating abstract syntax trees. Despite its advantages, writing migration rules for abstract syntax trees is hard work, which leaves us wondering if it is possible to leverage artificial intelligence to learn the migration rules automatically. This leads us to a hands-on session, explaining the failure of common encode-decoder translation networks when applied to code, and building an intuition for graph-convolutional neural networks that can tackle the challenges of code migration.

AGENDA
18:30 - Doors open and Networking
19:00 - Migrating AngularJS to Angular, Graph neural networks to the rescue by Lucas Fievet
20:00 - Optional networking
21:00 - Doors closing

BIO
Lucas holds a PhD in Machine Learning from ETH and has over 16 years of programming experience. He is the co-founder of Ahead Solutions GmbH, where he worked on multiple AngularJS projects. In the past years he has been focusing on applications of AI to improve software development productivity.