OFA-X: Explainable Visual Question Answering

The project „OFA-X“ deals with the interpretability of Neuronal Nets for the multimodal image processing. The objective is to compile a research project based on the seminar work of the applicants „Visual Explanation of Multimodal Transformer Models“ and to submit it at the CVPR 2023 (Conference on Computer Vision and Pattern Recognition). This is one of the most renowned conference in the field of image processing neuronal nets. we will train a new AI-model, which answers any questions on images and offers a textual explanation for its output by using a highly up to date approach. Further details on the relevance of the topic can be found in the detailed project description.

Research group

• Lukas Braach
• Björn Plüster

Mentor

• Prof. Dr. Stefan Wermter
• Dr. Jae Hee Lee