The IEEE Intelligent Transportation Systems Society (ITSS) advances the theoretical, experimental, and operational aspects of intelligent transportation systems, including autonomous driving. The Intelligent Vehicles Symposium (IV 2022) is a premier forum sponsored by the IEEE ITSS in its 33rd edition of the conference this year. This workshop is part of IV 2022 and a follow-up to the first workshop held at IV 2021.
The workshop will take place on Sunday, 05. June 2022 in Room K5 in the Eurogress Aachen from 08:30 CEST to 12:30 CEST.
For detailed information, check the agenda below.
| Time in CEST | Talk | Speaker |
|---|---|---|
| 08:30 | Begin | |
| 08:30 | Opening and Introduction | Larissa Triess (Mercedes-Benz AG) Saqib Bukhari (ZF Group) Cristobal Curio (Hochschule Reutlingen) |
| 08:35 | Introduction of KI Delta Learning Project | Amin Hosseini (Mercedes-Benz AG) |
| 08:45 | Invited Talk: Domain Generalization and (Continual) Unsupervised Domain Adaptation: A Birds Eye View Synopsis for Semantic Segmentation | Prof. Tim Fingscheidt (Technische Universität Braunschweig) |
| 09:15 | Contributed Paper: Capsule Networks for Hierarchical Novelty Detection in Object Classification | Thies de Graaff (German Aerospace Center) |
| 09:35 | Invited Talk: Supervised and Unsupervised Semantic Image Synthesis | George Eskandar (University of Stuttgart) |
| 10:10 | Break | |
| 10:30 | Invited Talk: Neural Rendering for Automated Driving | Prof. Matthias Niessner (Technical University of Munich) |
| 11:00 | Contributed Paper: BackboneAnalysis: Structured Insights into Compute Platforms from CNN Inference Latency | Matthias Zeller (CARIAD) |
| 11:20 | Contributed Paper: MEAT: Maneuver Extraction from Agent Trajectories | Julian Jordan (Mercedes-Benz AG) |
| 11:40 | Invited Talk: Automated Detection of Labeling Errors in Semantic Segmentation Datasets | Dr. Matthias Rottmann (University of Wuppertal) |
| 12:10 | Wrap Up and Closing | Larissa Triess (Mercedes-Benz AG) Saqib Bukhari (ZF Group) Cristobal Curio (Hochschule Reutlingen) |
| 12:30 | End |
Highly automated vehicles are an integral part of everyday traffic and will be used even more in the future with further autonomy. The world is dynamically evolving with an "open-world" nature, creating new domains for highly/fully automated vehicles to learn. The environmental changes and developments are not only important for in time and location aspects, but also the constant developments in the field of hardware, such as sensors, software, and development states, pose great challenges. Up until now, artificial intelligence (AI) modules in autonomous driving applications are only scalable to a limited extent. The reason lies in the training strategies applied, in which algorithms have to be trained for each new domain. This results in enormous development costs.
This workshop addresses the need for new training methods for AI systems and researches. Disruptive methods of "effective" machine learning can enable a more efficient and unrestricted use of AI and thus “scaling of autonomy” in new domains. With this approach, automated vehicles can access new markets faster and respond agile to new demands.
The scaling of autonomy can be achieved by using the gained knowledge from known domains and focusing on learning gaps to the target domain. This is also called Delta Learning, which can be achieved by applying methods from transfer learning and incremental learning.
 | Dr. rer. nat. Matthias Rottmann (University of Wuppertal) Topic: Automated Detection of Labeling Errors in Semantic Segmentation Datasets |
 | George Eskandar, M.Sc. (University of Stuttgart) Topic: Supervised and Unsupervised Semantic Image Synthesis |
 | Prof. Dr.-Ing. Tim Fingscheidt (Technische Universität Braunschweig) Topic: Domain Generalization and (Continual) Unsupervised Domain Adaptation: A Birds Eye View Synopsis for Semantic Segmentation |
 | Prof. Dr.-Ing. Matthias Nießner (Technical University of Munich) Topic: The Revolution of Neural Rendering |
The focus of this workshop is on methods and approaches that contribute to the scaling of autonomy solutions. The goal of these methods is to reduce the dependency on high data volumes, to use different data sources and to be able to use knowledge that has already been gained from other domains:
The application areas for the above methods are different sources of domain shifts and changes. These changes can be caused by:
The application areas may be applied to different sensor technologies (e.g. camera, LiDAR, and RADAR) and also be used in different functional modules such as perception, fusion, behavior and trajectory planning.
Thies de Graaff and Arthur Ribeiro de Menezes, “Capsule Networks for Hierarchical Novelty Detection in Object Classification”
Frank M. Hafner, Matthias Zeller, Mark Schutera, Jochen Abhau, and Julian Francisco Pieter Kooij, “BackboneAnalysis: Structured Insights into Compute Platforms from CNN Inference Latency”
Julian Schmidt, Julian Jordan, David Raba, Tobias Welz, and Klaus Dietmayer, “MEAT: Maneuver Extraction from Agent Trajectories”
The workshop is organized by members of the German publicly funded project of KI Delta Learning together with great partners such as Mercedes-Benz, BMW, CARIAD, Porsche Engineering, Valeo, Bosch, ZF, DLR, FZI, Uni Wuppertal, Hochschule Reutlingen, TUM, Uni Freiburg, Uni Stuttgart, Uni Tübingen, and others.
 | Larissa Triess from Mercedes-Benz AG |
 | Prof. Cristobal Curio from University of Reutlingen |
 | Dr.-Ing. Saqib Bukhari from ZF |
