Results

Booklet - Posters - Presentations - Video

Booklet

The poster booklet contains general KI Wissen information, illustrates the project's trajectory, and presents research findings.
>>> Download Booklet: KI Wissen - Powered by Knowledge <<<

Posters for Download

No.Title
1.2Model-Agnostic Body Part Relevance Assessment for Pedestrian Detection
1.3Investigating the Effects of Augmentation Techniques on Interpretability and Performance of Pedestrian Detection Models
1.4Static Scene Knowledge for Pedestrian Detection
1.5Method to Fuse Map Layer and Pedestrian State Space to Improve 3D Detector Output
1.7Knowledge Aware Pedestrian Detection
1.8Geo-Informed Conformity Check of Pedestrian Detection Models
1.9Localized Semantic Feature Mixers for Efficient Pedestrian Detection in Autonomous Driving
1.12Plausibility Verification for 3D Object Detectors Using Energy-Based Optimization
1.13Object Detection Plausibility with Concept-Bottleneck Models
1.14Digital Twin Creation and Evaluation of Color-Dependent Object Detection
1.15Multi-Modal Sensor Fusion for Robust Obstacle Detection
1.16Leveraging Knowledge for Traffic Sign Detection
2.1Informed Motion Planning
2.2Knowledge Integrated Plausible Motion Forecasting
2.3An Empirical Bayes Analysis of Object Trajectory Representation Models
2.4An Efficient and Robust Multi-Modal Trajectory Predictor Baseline for Autonomous Driving
2.6Integration of a Priori Knowledge Using a Causal Model of Vehicle Trajectories
2.7Causality-Driven Checks of the Physical Conformity of Vehicle Trajectories
2.8Tracking and Trajectory Prediction Using Scene Semantics
2.9Learning-Aided Warmstart of Model Predictive Control in Uncertain Fast-Changing Traffic
2.10Motion Planning Under Temporal Logic Specifications
2.11An Outlier-Robust and Efficient Bayesian Filter and Smoother
2.12Behavior-Conditioned Driving Policies
3.1Harnessing Symbolic Knowledge Extraction and Utilization for Informed Decision-Making
3.2Model Agnostic Local Analysis with Latent Attacks (MALALA)
3.3A Concept to Support AI Models by Using Ontologies
3.4Knowledge-Based Comparison of CNNs
3.5AI Net Observer
3.6Legal Norm Formalization – An Overview
3.7From Legal Documents to Formalized Rules with Large Language Models
3.9Traffic Sequence Charts for Knowledge Formalization and AI Application
3.10Risk Assessment and Adaptive Decision-Making for Autonomous Vehicle Safety
3.11Monitoring Traffic Rule Conformance and Integrated Knowledge Impact
3.12Runtime Monitoring Using TSC-Based Knowledge Building Blocks
3.15Consistent Joint Action and Explanation Prediction for Autonomous Driving
3.16Generating Training Data from Formalized Traffic Dynamics
3.17TSC-Based Generation of Concrete Scenarios for Synthetic Data
3.18Local Concept-Based Explanations for Object Detectors
3.19The Corpus of Traffic Rules

Presentations

Presentations of the excellent KI Wissen Final Event:

Keynote: From Road to Cloud. From Virtual to Real. | Gilles Mabire, CTO Continental
KI Wissen: Project Overview | Simon Heinz, Jörg Dietrich, Continental
Knowledge Integration | Jörg Reichardt, Continental
Knowledge Extraction | Antje Loyal, Continental
Knowledge Conformity | Johannes Link, Valeo
Highlight Session 1
     Interpretable Model-Agnostic Plausibility Verification for 2D Object Detectors | Mert Keser, Continental
     Efficient Pedestrian Detection with Inter-stage Knowledge Integration | Abdul Hannan Khan, DFKI
     Knowledge-augmented Pedestrian Detection | Julian Wörmann, fortiss
Highlight Session 2
     A Causal Model of Vehicle Trajectories for Integration of a Priori Knowledge | Christian Brunner, e:fs
     Advancements of Local and Global Explanation Methods for Failure Case Detection | Christian Hellert, Continental
     Knowledge Formalization, Integration & Monitoring with Traffic Sequence Charts (TSCs) | Dominik Grundt, DLR
     The Normative Body of Traffic Rules for Autonomous Driving | Georg Borges, Universität des Saarlands
Highlight Session 3
     Model Predictive Control under Temporal Logic Specifications | Etienne Bührle, FZI

     Knowledge Integration in Tracking, Prediction and Planning | Mohamed-Khalil Bouzidi, Yue Yao, Jonas Neuhöfer, Christian Schlauch, Continental AG

Enabeling, Integration and Demonstration | Federica Paolicelli, AVL
Demonstrator | Thomas Knorr, AVL
Demonstration of Causality-driven Physical Conformity Checks of Vehicle Trajectories | Himanshu Agarwal, e:fs
Demonstration of Trajectory Prediction and Planning | Yue Yao, Mohamed-Khalil Bouzidi, Continental
Demonstration of the Interactive Visual Analytics System | Kostadin Cvejoski, Fraunhofer IAIS

Video

>>> Download Video: KI Wissen Final Event (720p/mp4) <<<