International VDI Conference - Future of AI in Automotive

Chair’s Welcome & Opening Address

Keynotes

AI Across the Automotive Domains

• Software & AI in autonomous driving, cockpit, e-powertrain & connected services
• Ingredients for success
• War for AI talent

Joachim Langenwalter, Sr. VP SW & HW, Stellantis

Unsupervised Learning Empowers AI – Where Do We Empower Trust & Ethical Deployment?

• Traditional supervised learning is far from intelligence – What is ahead of us?
• Technology needs to be beneficial to its users, to society, to the environment - such care must be extended to so-called AI methodologies
• What is AI?

Prof. Dr. Patrick van der Smagt, Director of AI Research, Volkswagen AG, Germany

Coffee Break

I. Building the AI-Powered Organisation

Leveraging the Potential of AI: The Project AI at BMW Group

• Hub- & spokes-approach enabling the organization to leverage the full technological potential
• AI applications relevant & under development along the value chain
• D³ use case portfolio accelerating technology scaling across the organization

Daniela Rittmeier, AI Portfolio Manager, BMW Group, Germany

Increasing AI Maturity at Large Organizations: A Story from the Real World

• A framework for choosing organizational structures to boost ML capabilities
• Assessing the current status & existing challenges with ML
• Supplementing the framework with real-world examples

Jan Zawadzki, Head of Artificial Intelligence (acting), Car.Software Org, Germany

Anchoring AI in an Automotive Supplier

• Where AI can bring value to automotive supplier products
• Towards a product-centric AI reference architecture from supplier perspective
• AI as a challenge & opportunity for automotive suppliers in the context of automated & autonomous driving

Dr. Jürgen Bohn, Head of Innovation Cluster Artificial Intelligence & Data Science, Schaeffler AG, Germany

Lunch Break

II. AI for Autonomous Driving

AI & Big Data Management for Autonomous Driving

• Safer driving today begins in the data center
• Data management & artificial intelligence
• AV workflow challenges & pain points

Frank Kraemer, IBM Systems Architect, IBM, Germany

Deep Teaching: A Scalable AI Approach to Autonomous Driving

• Corner case scenarios make developing safety critical AI for AD slow & costly
• Deep Teaching reduces the cost & time of development by allowing large scale training of neural networks without the bottleneck of human annotation & by tackling training in the small data regime

Vladislav Voroninski, CEO, Helm.ai, USA

An Optimal Sensor Set for Automated Driving Using AI

• Overview Valeo & sensor portfolio
• Auomated driving sensor set & processing chain
• Valeo AI algos for cameras, LIDAR and ultrasonic sensors

Jörg Schrepfer, Head of Driving Asistance Research Germany, Martin Simon, Harald Barth, all: VALEO Schalter & Sensoren GmbH, Germany

Coffee Break

III. Safety & Security of AI

Must AI Have Common Sense in Order to Allow Guarantees?

• AI systems based on DL cannot be trusted with anything that has not been precisely anticipated by their designers
• AI needs to be able to deal with problems that are novel & variations that have not been seen before
• Some notion of common sense in AI is required to allow guaranteed behavior; promising research directions are explainable AI & hybrid AI

Dr. Christian Müller, Principal Researcher, DFKI Research Fellow, Head of Competence Center Autonomous Driving, German Research Center for Artificial Intelligence DFKI, Germany

Determination & Establishment of Development Life Cycle for AI based on Automotive Cybersecurity Application

• Adapting a secured AI architecture should be encouraged as the basic AI architectures do not provide provision against security risks
• Performing adversarial attacks on AI should be mandatory activity as they showcase the security vulnerabilities while training and testing the AI model
• A development life cycle with security objectives helps in mitigating security vulnerabilities in AI

Philipp Veronesi, Managing Director, Narasimha Deshpande both: CYRES Consulting Services GmbH, Germany

Panel Discussion

Data Security by Design for AI

• Dr. Tobias Große-Puppendahl, Head of Platform Incubation Emerging Technologies, Porsche AG, Germany
• Dr. Klaudius Kalcher, Chief Data Scientist, Mostly AI, Germany
• Dr. Felix Schuster, Co-Founder, Edgeless Systems GmbH, Germany
• Jan Zawadzki, Head of Artificial Intelligence (acting), Car.Software Org, Germany

Chair‘s Closing Remarks & End of Day 1

IV. Data Management & Data Architecture

Opening Keynote

Data & AI Organization in Groupe PSA & Data Sharing Initiatives

• Presentation of Data & AI organization in Groupe PSA
• Skills to set up Data Projects
• Data projects initiatives launched: Predictive maintenance,connected cars, parts conformity...

Guillaume Gruel, Head of Engineering & Quality Data & AI Department, Matthieu Donain, Head of Openlab AI, both: Groupe PSA, France

PROMETHEA Project: Data Sharing for AI exploitation

• PROMETHEA : A virtuous loop for the product Life Cycle
• Definition of a standard data model leading to solutions homologation by GALIA
• Development of prototype exploiting measurements in production according to 3 use cases (Product/Process feaseability, convergence and quality monitoring)
• PROMETHEA Common data model based on STEP AP 242 xml overview

Alexandre Loire, Engineering Standardization Projects Director, GALIA, France

AI at the Edge: Building an Automotive Edge Computing Platform

• Fully automated setup of clusters on on-premise hardware & operating concept for a large number of clusters
• Self-service for platform services, multi-cluster management & centralized logging and monitoring

Dr. Tobias Bürger, Lead Big Data & AI Platform, BMW Group, Germany

Coffee Break

Breakout Sessions Directly engage with international experts in these highly interactive sessions. Discuss current challenges, needs, and solutions with your peers!

• Manufacturing: When can we trust AI to make better decisions than humans – and what happens next? Pierre Garrot, AI Projects Director, Faurecia, France
• Physical vs. AI (Empirical) Models, Dr. Marc Hilbert, Team Lead – Machine Learning for Engineering & Production, Volkswagen Software Innovation Center – Data: Lab Munich, Germany
• Explainable AI: Theory and Industrial Applications, Dr. Tarek R. Besold, Head of Strategic AI at DEKRA DIGITAL

Lunch Break

V. AI for Development & Engineering

Using AI to Accelerate & Automate Automotive Development Processes

• Unused potential for further automation in automotive R&D
• AI can help automate similar but not identical tasks significantly
• Example: Plausibility checks of data prior to actual task performed by a human

Dr. Mirko Knaak, Expert in Artificial Intelligence, IAV GmbH, Germany

Accelerating Product Development with AI-based Virtual Testing

• Introduction to our company & vision
• A platform leveraging AI for engineers
• Virtual testing applied to engine calibration

Dr. Joël Henry, Principal Engineer, Monolith AI, UK

VI. AI for Manufacturing & Production

Leveraging Artificial Intelligence in Automotive Manufacturing

• Insights into how Audi uses AI in manufacturing to increase vehicle quality
• Requirements for scaling AI solutions
• Success factors for bringing AI projects from proof-of-concepts into production

Johannes Schniertshauer, Lead Machine Learning Engineer, Audi AG, Germany

Coffee Break

Collaborative Learning as Enabler for Smart Automotive Services

• Automotive edge computing as fast-developing automotive technology bringing AI technologies into & facilitating intelligent services between vehicles
• Cars as platforms for local smart automotive services that support local communication between vehicles, but also interactions between drivers, passengers & external actors
• Collaborative learning as extension of federated learning supporting training & usage of AI-based services locally in cars

Prof. Dr. Wolfgang Maaß, Scientific Director & Head of Smart Service Engineering, German Research Center for Artificial Intelligence DFKI, Germany

VII. AI for HMI & UX

What It Takes to Put Conversational AI Systems Into Cars – Lessons Learned

• Building Conversational AI assistants in cars
• AI technology is (only) the basis
• The crucial role of UX Research in the process
• Experiences gained from real projects

Dr. Nils Lenke, VP & GM Apps BU, Vanessa Tobisch, both: Cerence, Germany

Explainability in Human-AI-Interaction for Automotive Systems

• Interaction model for transparent communication
• Customer empowerment and positive UX with explanations
• From Explainable AI to system information for endusers

Dr.-Ing. Joana Hois, Artificial Intelligence Research, Mercedes-Benz AG, Germany

Chair‘s Closing Remarks & End of Conference