Matthias Neuwirth-Trapp

Doctoral candidate at ETH Zurich and industry research scientist at Bosch Research, working on continual learning and adaptation for vision-language and multimodal foundation models.

My research focuses on lifelong multimodal intelligence, including incremental learning for object detection, multimodal image generation across camera views and domains, and representation learning for computer vision.

Contact

Current Roles

ETH Zurich

Doctoral Candidate, April 2024-present

Dissertation topic: “Towards Lifelong Multimodal Intelligence: Continual Learning for Vision-Language and Multimodal Foundation Models”

Supervisors: Luc Van Gool, Konrad Schindler, Christos Sakaridis, Danda Paudel, and Maarten Bieshaar.

Expected graduation: September 2027.

Bosch Research

Industry Research Scientist and Doctoral Candidate, April 2024-present

  • Continual learning and adaptation for vision-language and multimodal foundation models.
  • Multimodal image generation across camera views, sensor modalities, and visual domains.
  • Representation learning and keypoint refinement for computer vision.

Education

ETH Zurich

Doctoral Candidate, April 2024-present

TU Braunschweig

M.Sc. Physics, December 2020-April 2024

Thesis: “Solving Quantum Mechanical Differential Equations with Physics-Informed Neural Networks”

B.Sc. Physics, September 2017-December 2020

Thesis: “Computer Vision for the Morphological Characterization of Soot Particles”