This week we introduce next AIthena consortium partner: BUW – Institute for Technologies and Management of Digital Transformation (TMDT), University of Wuppertal (BUW)
In its research areas, the TMDT deals with issues related to the realization of the capabilities of intelligent technical systems. It pursues a holistic approach, which not only investigates questions regarding the technical implementation and feasibility of the systems in industry, but also their sustainable and economic implementation and use. For this purpose, the TMDT is active in application-oriented research projects, which are characterized by close cooperation with partners from research and industry. The research results are made available to the scientific community through publications in relevant conferences and journals.
What is the role of TMDT and the team in the project, and what are you currently working on in the project?
The role of the TMDT is the research on explainability and transparency of learning models that are implemented in CCAM (Connected, Cooperative and Automated Mobility) solutions. Specifically, we investigate learned model representations about trajectory predictions and environment perception.
We use a variety of methods to dissect individual neural networks and determine the role of individual network parameters for the overall learning task, i.e., predicting the trajectory of objects such as pedestrians and vehicles. In case of model failure, we make the faulty model decisions transparent and explainable to address the model limitations and to facilitate trustworthiness of their applications within the CCAM context.
From your perspective, how do you see the contribution of the AIthena project to building trustworthy, explainable, and accountable AI-based CCAM?
In the rapidly evolving landscape of Connected, Cooperative and Automated Mobility, research projects that prioritize transparency and explainability of AI methods within autonomous driving systems play a pivotal role. By developing and implementing AI technologies that can be easily understood and scrutinized, these projects bolster the overall trustworthiness and accountability of AI-powered vehicles.
Ensuring that the decision-making processes of autonomous cars are transparent to both users and regulators not only enhances safety but also fosters public confidence in this transformative technology. In the pursuit of a future where CCAM becomes a seamless and reliable mode of transportation, transparent AI methods serve as the cornerstone, promoting responsible and accountable AI deployment. The AIthena project serves as an exemplary research project that contributes to this important development.
You can read more about the TMDT at Homepage: Institute for TMDT (uni-wuppertal.de)