Aktuelles
Eingeladener Vortrag am Lehrstuhl für Künstliche Intelligenz am 28. Oktober 2026 von Jun.-Prof. Dr. Gesina Schwalbe
[28.10.2026]Im Rahmen des Oberseminars des Lehrstuhls für Künstliche Intelligenz hält am Mittwoch, den 28.10.2026, um 10:00 Uhr, Frau Jun.-Prof. Dr. Gesina Schwalbe (Universität Ulm) einen Vortrag zum Thema:
"Concept-Based Explainability for Verification and Integration of Symbolic Knowledge in Neural Networks"
Abstract:
Deep neural networks (DNNs) achieve unbeaten performance in automated processing of non-symbolic sensory inputs like images. This includes safety-critical application areas such as automated driving or diagnosis assistant systems in healthcare. Such domains come along with a rich set of prior, humandefined symbolic (expert) knowledge, like known object relations (e.g., a head typically indicates a pedestrian). System safety certification now demands strict compliance to these prior rules, making it a compelling goal to use available knowledge directly during DNN training and verification. Conceptbased explainability (C-XAI), a subfield in explainable Artificial Intelligence, poses a promising pathway to access DNN-learned knowledge, associating human-readable specifications with latent encodings in DNNs. Particularly, it can be used to inspect, verify, and even insert prior symbolic knowledge into otherwise opaque deep neural networks. This talk introduces basic techniques from C-XAI, and discusses concurrent research on applications for model verification and correction.
Interessierte sind herzlich eingeladen, online im Rahmen eines Zoom Meetings teilzunehmen:
https://fernuni-hagen.zoom.us/j/62388524887?pwd=WmJRRUpDZWZ4WlJoeFlGQTBESWplQT09
