Some of my past work:

AI Projects

• Security of AI Systems: an in-depth study on the security risks of transfer learning, including backdoor and adversarial attacks; in collaboration with the German Federal Office of Information Security (BSI) publication

• BerDiBa (Berlin Digital Rail Operations): a holistic digital twin for rail operations using innovative signal processing methods and Artificial Intelligence. My work focused on robustness assessment of AI modules in autonomous railway systems.

• Localized Classmaps: a new Explainable AI method that enables an interpretation of the outputs of classification algorithms. code, publication

Talks

• WeAreDevelopers 2023: Machine Learning: Promising, but Perilous
• Automotive Software Development Meetup: Building Safety Cases for AI-based perception, May 2023
• PyData Yerevan 2022: The Explainability Problem summary, video

Publications

• (2024) Structuring a Training Strategy to Robustify Perception Models with Realistic Image Augmentations arXiv preprint. Ahmed Hammam, Bharathwaj Krishnaswami Sreedhar, Nura Kawa, Tim Patzelt, Oliver De Candido.

• (2022) Security of AI-Systems: Fundamentals - Provision or use of external data or trained models Bundesamt für Sicherheit in der Informationstechnik.

• (2022) Explainable AI (XAI) with Class Maps Towards Data Science

• (2022) Defining Hydrogeological Site Similarity with Hierarchical Agglomerative Clustering. Groundwater. Kawa, N., Cucchi, K., Rubin, Y., Attinger, S. and Heße, F.

• (2021) exPrior: An R Package for the Formulation of Ex-Situ Priors. The R Journal. Falk Heße, Karina Cucchi, Nura Kawa and Yoram Rubin.

• (2019) Ex-situ priors: A Bayesian hierarchical framework for defining informative prior distributions in hydrogeology. Advances in Water Resources. Cucchi, K., Heße, F., Kawa, N., Wang, C., & Rubin, Y.