Data-driven ecosystems in BMW Group & Autonomous Intelligent Control

Join us in our first meetup of the season where Tobias Bürger, who leads the Platform and Architecture Group within the Big Data, Machine Learning, and Artificial Intelligence Department at BMW Group, will detail the path taken, the established technology stack and the challenges they faced while using data-driven technologies, forming a data-centric culture and establishing a globally available and scalable data lake equipped with data science tools within the BMW Group. Afterward, Nikolaos Passalis, Ph.D. candidate in the Department of Informatics, AUTh, will review the recent developments on autonomous control and self-driving cars using Deep Learning and will discuss the use of Reinforcement Learning algorithms for developing end-to-end trainable models for autonomous control.

Thursday 20/09 - 19:00, at OK!Thess.

Agenda:
19.15: 'Data-driven ecosystems in the automotive industry', Tobias Bürger, BMW Group
20.30: 'Deep Learning for Autonomous Intelligent Control', Nikolaos Passalis, AUTh
21:00: Networking and socializing

------

'Data-driven ecosystems in the automotive industry', Tobias Bürger

Abstract: As data-driven solutions based on machine and deep learning are gaining more and more momentum, workflows to build and deploy such services in a reliable and flexible fashion are of utmost importance. The BMW Group IT team drives the usage of data-driven technologies and forms the nucleus of a data-centric culture inside of the organization. Part of the team’s mission is the establishment of a globally available and scalable data lake equipped with data science tools that help efficiently process and analyze data. Tobias Bürger details the path the BMW Group has taken and the established technology stack and explore the challenges faced along the way. Tobias also offers an overview of novel machine learning use cases, such as those based on gradient boosting and convolutional neural nets, that have been applied and deployed in real-world environments. Along the way, you’ll learn the value that has been created for domains including but not limited to connected vehicles, vehicle development, and aftersales.

Bio: Tobias Bürger leads the Platform and Architecture Group within the Big Data, Machine Learning, and Artificial Intelligence Department at BMW Group, where he is responsible for the global big data platform that is the core technical pillar of the BMW data lake and is used across different divisions inside the BMW Group, spanning areas such as production, aftersales, and ConnectedDrive.

--
'Deep Learning for Autonomous Intelligent Control', Nikolaos Passalis

Abstract: Deep Learning (DL) has revolutionized the development of self-driving cars and provided powerful methods for autonomous intelligent control of various vehicles, e.g., drones. However, successfully developing embedded DL solutions is not straightforward and several challenges are posed (safety constraints, energy and computational power restrictions). In this talk, we will review the recent developments on autonomous control using DL and discuss the use of Reinforcement Learning (RL) algorithms for developing end-to-end trainable models for this task. Finally, the session will get practical with a hands-on demonstration of developing and training a Reinforcement Learning agent for performing a fine-grained control task: landing a spaceship on the moon!

Bio: Nikolaos Passalis is a Ph.D. candidate in the Department of Informatics, Aristotle University of Thessaloniki. He obtained his B.Sc. in Informatics in 2013 and his M.Sc. in Information Systems in 2015 from Aristotle University of Thessaloniki, Greece. His research interests include deep learning, computational intelligence and information retrieval. His doctoral studies are supported by the General Secretariat for Research and Technology (GSRT) and the Hellenic Foundation for Research and Innovation (HFRI) since 2017.

The presentations will be in English.