Autonomous Driving Munich .1
Details
It is time for the second edition of Autonomous Driving Munich. This time, we are hosted by Valtech at Friedenheimer Brücke. Looking forward to see you all around, particularly for some great discussions!
___ Agenda ___
_ Valeo - AI in a nutshell: A small Deep Learning review from a Self-Driving Car Engineer
Dr. Stefan Milz, Technical Lead - Deep Learning (DAR Germany) - Valeo Expert, Valeo
_ Dinner Break
_ Processing autonomous driving data with Apache Spark and TensorFlow
Adrian Achihăei, Lead Data Scientist, Valtech
_ 3D ultrasonic sensors: non-optical vision for autonomous vehicles
Tobias Bahnemann, Managing Director, Toposens
_ Socializing
___ Sessions ___
_ Valeo - AI in a nutshell: A small Deep Learning review from a Self-Driving Car Engineer
Dr. Stefan Milz, Technical Lead - Deep Learning (DAR Germany) - Valeo Expert, Valeo
http://www.valeo.de/
Session Abstract:
Valeo is the leading sensor supplier for Autonomous Driving and strongly pushes effort in the research area of AI. This workshop simply gives an overview of different Deep Learning techniques in the field of Autonomous Driving. The talk investigates varying use cases: starting from camera image analysis up to lidar based point cloud processing. What are the present challenges? What are current bottlenecks? Lets start an AI discussion for Autonomous Driving.
Speaker Bio:
Stefan Milz is Technical Lead and AI Enthusiast at the Valeo in Kronach. Stefan holds a PhD in physics (TUM) and is Valeo Expert for Machine Learning
_ Processing autonomous driving data with Apache Spark and TensorFlow
Adrian Achihăei, Lead Data Scientist, Valtech
https://www.valtech.de/
Session Abstract
The development of self driving car algorithms are driven by large amount of data from car sensors. Besides lidar, radar and traces, the cars are equipped with 4k cameras that generate gigabytes of data per minute. Neuronal models getting bigger and more computation power is needed to increase the model performance and reduce the error rate. Processing large amount of data on a cluster requires higher computation and I/O workload. The industry adopted tools are Hadoop for distributed storage, Apache Spark for parallel processing and TensorFlow for deep learning. This hand-on session will focus on how to build a distributed data pipeline to analyse, train and evaluate self-driving car algorithms. Speaker
Speaker Bio
Adrian Achihăei is Lead Data Scientist for Autonomous Driving and Machine Learning at Valtech. Adrian is a computer science engineer with a passion for data processing, visualisation, algorithms, distributed systems, fluent in several programming languages and an early adopter of Hadoop ecosystem, using it in production since almost a decade ago.
_ 3D ultrasonic sensors: non-optical vision for autonomous vehicles
Tobias Bahnemann, Managing Director, Toposens
https://www.toposens.com/
Session Abstract
Toposens is building innovative 3D sensor technology. We are developing a completely new method of localizing 3D positions through ultrasound for the precise 3D detection of objects in real time. The sensor system of Toposens generates new ways of providing non-optical vision for autonomous vehicles bringing a new way of 3D sensing to the market. The systems are energy efficient, small, lightweight, robust and also available at a competitive price. In our talk we will explain you the basic principles of the system and show you the functionality in a small demo.
Speaker Bio
Tobias Bahnemann has been interested in startups since high school. He founded his first one man e-commerce company at the age of 19 followed by studying business administration in Münster. Working for startups in Berlin, like p2p payment platform Lendico of Rocket Internet he continued to dive further into the startup scene, before he cofounded the 3D sensor startup Toposens in Munich at the end of 2014.
___ Further Details ___
> Sponsor: Valtech
> The talks are all held in english and will take ~30 minutes each
> Details about the talks and speakers will be added soon!