The talk by Prof. Dr. Zeynep Akata from the Technical University of Munich (TUM) on 18 March 2025, focused on “Explainability in Multimodal Large Language Models”. The talk was part of the IUC Applied Research Talks series and was co-hosted by the Predictive Town Hall Meetings.

Some key research areas include interpretability and biases in Vision Language Models (VLMs), compositionality, and model modularity and reusability. The talk highlighted research on rare image generation, text-to-image synthesis, few-shot learning, and anomaly detection using attributes. Finishing with possible use cases of multimodal AI in the SAP world.

The discussion round between SAP and TUM included the insight of several researchers from Prof. Dr. Akata’s chair and continued in a dynamic discussion between the attendees.