Almut Sophia Koepke

Dr. Almut Sophia Koepke is a junior research group leader and TUM Junior Fellow at the Technical University of Munich (TUM) and University of Tübingen. She leads the multi-modal learning research group focusing on video understanding through sound, vision, and text integration.

• University: Technical University of Munich
• School: TUM School of Computation, Information and Technology
• Department: Informatics 9
• Academic Rank: Researcher

Her research spans multi-modal learning, audio-visual foundation models, and cross-modal attention mechanisms. Key themes include:

• Advancing zero-shot learning through language-guided audio-visual models
• Developing explainable AI systems via attention pattern translation in VQA
• Exploring temporal understanding in video-adverb retrieval
• Building robust multi-modal representations for self-driving applications

Recent publications analyze foundation model capabilities in audio-visual tasks (ICCV 2025), temporal reasoning (ACMMM 2024), and cross-modal attention frameworks (ECCV 2022). She co-organizes CVPR workshops on foundation model evaluations and serves as area chair/reviewer for major conferences.