290 pixel image width
Newsletter #8 10/2023
600 pixels wide image

Hello to friends and family out there!

Already in the last quarter of 2023, we are rallying towards various finishing lines for 2023 and beyond. Project No 3 of our KI FAMILIE – KI Data Tooling – will be completed by the end of the year. This means another final event is coming up. In a 2-days-event on December 5 and 6, 2023, the project’s results will be presented at the BMW World in Munich. Continue reading this Newsletter and you we will get a glimpse of what the project has achieved – fairly exciting, to be honest.

Our fourth and last KI FAMILIE project KI Wissen will finish next next spring and close the KI FAMILIE umbrella. Originally planned for a project duration of 36 months, a project extension of three months is currently under progress. Accordingly, peparations for the forth and last final event within the KI FAMILIE are already in the pipeline - see the Save-the-Date further below.

Related to this, we also want to bring your attention to KI FAMILIE’s sister family PEGASUS and its project VVM (Verification Validation Methods). VVM will also reach the finishing line by December 31, 2023. Presentation of the project results are scheduled for November 21 and 22, 2023 at the Carl-Benz-Arena in Stuttgart. More information can also found in this Newsletter.

It is obvious, that the way to automated driving will be enriched with further important results and experiences of large pre-commercial R&D projects in the next months. Knowledge transfer and information exchange will become even more important in the future to support the automotive industry and strengthen Germany’s and Europe’s industrial base.

For the 8th time, enjoy reading our Newsletter.

The KI Familie Editorial Team
600 pixels wide image

KI Data Tooling Final Event at the BMW World in Munich
600 pixels wide image
KI Data Tooling, a BMWK funded research project, has taken the important step from a Data Set to the Data Kit, a complete data solution for the training and validation of AI functions for highly automated driving. Join us for presentations about our significant improvements on data, methods and tools for different sensor modalities. The conference also offers the opportunity, to discuss challenges on the KI Data Tooling journey and opportunities ahead.

Project Achievement: KI Data Tooling Introduces Resilient ISO 8800-Oriented Data Kit
600 pixels wide image
Picture: Data Kit Framework Architecture

Starting in 2024, the ISO 8800 standard for road vehicle safety in the realm of artificial intelligence (AI) will take effect. This standard has significant implications for all AI-based automotive applications. Companies involved in AI development for such applications face the challenge of meeting these new requirements. Approaching the end of the project period in December 2023, the publicly funded collaboration project, KI Data Tooling, is unveiling a reference architecture and accompanying methods for a data kit designed to generate, process, and evaluate sensor data used for automated driving perception. This data kit has been influenced by the currently known ISO 8800 specifications and creates a basis to pass future safety audits with a focus on data quality and Machine Learning Operations (ML-Ops) tooling.
600 pixels wide image

KI Wissen: Save the Date for Final Event
600 pixels wide image
Prepare for an extraordinary journey in AI for automated driving at the KI Wissen Final Event.
The event will be a dynamic live experience featuring presentations, demonstrations, a poster session and lively discussions on project results.
600 pixels wide image

KI Delta Learning: Final Report available
The project period of KI Delta Learning ended on 31. March 2023 . The project presented the results to the public at its final event on 9. March 2023. Since then, the Final Report has been compiled and is available on the project's website.
600 pixels wide image
VVM Final Event offers presentations, exhibitions, demonstrations and international panel discussions.This occasion marks the culmination of the VVM project, celebrating the collective achievements and setting the course for the future of safety verification in automated vehicles.

Events 2023

190 pixel image width

34th British Machine Vision Conference 

20.-24. November
Aberdeen, UK
Find out more

 

Pegasus Family: VVM Final Event
21.-22. November
Stuttgart, Germany
Find out more


7th ICSRS
International Conference on System Reliability and Safety
22.-24. November
Bologna, Italy
Find out more

KI Familie: KI Data Tooling Final Event
05.-06. December
Munich, Germany
Find out more

 

Expert Conference „Forschung und Technologie für autonomes Fahren"
12. December
Berlin, Germany

Academic Corner - KI Familie Publications
190 pixel image width

Localized Semantic Feature Mixers for Efficient Pedestrian Detection in Autonomous Driving
Abdul Hannan Khan
Read more

LDFA: Latent Diffusion Face Anonymization for Self-driving Applications
Marvin Klemp, Kevin Rösch, Royden Wagner, Martin Lauer 
Read more

Principles of Forgetting in Domain-Incremental Semantic Segmentation in Adverse Weather Conditions
Tobias Kalb
Read more

Sensor Equivariance by LiDAR Projection Image
Hannes Reichert, Manuel Hetzel, Steven Schreck, Konrad Doll, Bernhard Sick 
Read more

The IMPTC Dataset: An Infrastructural Multi-Person Trajectory and Context Dataset
Manuel Hetzel, Hannes Reichert, Günther Reitberger, Erich Fuchs, Konrad Doll, Bernhard Sick 
Read more

Survey on LiDAR Perception in Adverse Weather Conditions
Mariella Dreissig, Florian Piewak, Dominik Scheuble, Joschka Boedecker
Read more

Perception Datasets for Anomaly Detection in Autonomous Driving: A Survey
Daniel Bogdoll, Svenja Uhlemeyer, Kamil Kowol 
Read more

Effects of Architectures on Continual Semantic Segmentation
Tobias Kalb
Read more

Severity of Catastrophic Forgetting in Object Detection for Autonomous Driving
Christian Witte, Saqib Bukhari 
Read more

Auxiliary Task-Guided CycleGAN for Black-Box Model Domain Adaptation
Michael Brunner, Markus Rehmann, Cristóbal Curio
Read more

Focus on the Challenges: Analysis of a User-friendly Data Search Approach with CLIP in the Automotive Domain
Philipp Rigoll, Patrick Petersen, Hanno Stage, Lennart Ries, Eric Sax
Read more

Sparsely-Gated Mixture-Of-Expert Layers for CNN Interpretability
Svetlana Pavlitska, Christian Hubschneider, Lukas Struppek, J. Marius Zöllner
Read more

Category Differences Matter: A Broad Analysis of Inter-Category Error in Semantic Segmentation
Jingxing Zhou 
Read more

LMR: Lane Distance-Based Metric for Trajectory Prediction
Julian Schmidt , Thomas Monninger, Julian Jordan, Klaus Dietmayer
Read more

RESET: Revisiting Trajectory Sets for Conditional Behavior Prediction
Julian Schmidt, Pascal Huissel, Julian Wiederer, Julian Jordan, Vasileios Belagiannis, Klaus Dietmayer
Read more

Improving Cross-Domain Semi-Supervised Object Detection with Adversarial Domain Adaptation
Maximilian Menke, Thomas Wenzel, Andreas Schwung
Read more

Exploring Navigation Maps for Learning-Based Motion Prediction
Julian Schmidt, Julian Jordan, Franz Gritschneder, Thomas Monninger, Klaus Dietmayer
Read more

survAIval: Survival Analysis with the Eyes of AI
Kamil Kowol, Stefan Bracke, Hanno Gottschalk
Read more

An Empirical Bayes Analysis of Object Trajectory Representation Models
Yue Yao, Daniel Goehring, Joerg Reichardt
Read more

A Semi-Automated Corner Case Detection and Evaluation Pipeline
Isabelle Tulleners, Tobias Moers, Thomas Schulik, Martin Sedlacek
Read more
Contact us for comments & feedback 

ki-kommunikation@eict.de
Follow us
X/Twitter Nutzername LinkedIn Firmenname
600 pixels wide image
If the message is not displayed correctly, please click here.
European Center for Information and Communication Technologies – EICT GmbH
Torgauer Straße 12-15, Haus 13
10829 Berlin
Deutschland

+49 30 3670235-000
ki-kommunikation@eict.de
Wenn Sie diese E-Mail (an: unknown@noemail.com) nicht mehr empfangen möchten, können Sie diese hier kostenlos abbestellen.