Building 41/200, Room 041/0211 | |
089/6004-4153 | |
philipp.roesch@unibw.de |
Dr. rer. nat. Philipp J. Rösch
Since December 2023, I am Research Head Artificial Intelligence at the Institute of Distributed Intelligent Systems (VIS) at the University of the Bundeswehr Munich. From 2020 to 2023, I was a research assistant at the same institute, where I helped to manage the Monacum One GPU cluster.
I received my doctorate (Dr. rer. nat.) in Computer Science (Artificial Intelligence), where my research focused on Vision-Language Models. I obtained my Master's Degree (M.Sc.) in Statistics from Ludwig-Maximilians-Universität München in 2016 and my Bachelor's Degree (B.Sc.) in Economics from University of Mannheim in 2013. After my Master's, I worked in the private sector for several years.
Machine Learning Interest Group (MLIG)
I organize the Machine Learning Interest Group (MLIG), a university-wide platform for researchers at UniBw M and HSU focusing on machine learning and AI. MLIG hosts a variety of events including the AI LUNCH, AI TALK, AI STAMMTISCH, AI SALON and the UniBw AI Retreat (last both in cooperation with GCMC) to strengthen exchange and collaboration in that field.
Please visit the website at https://go.unibw.de/mlig.
If you work at UniBw M or HSU and would like to join the ILIAS group, please use https://go.unibw.de/JoinMLIG.
Teaching (German)
Herbsttrimester 2025
Logik (Bachelor) |
Di: 13:00 - 14:30 Mi: 09:45 - 11:00 |
Künstliche Intelligenz I (Bachelor) |
Mo: 09:45 - 11:15 Do: 13:15 - 14:45 |
Kognitive Systeme (Master) |
Mo: 13:15 - 14:45 Do: 15:00 - 16:30 |
Consultation hours
By appointment
Research Interests
- Deep Learning, Computer Vision
- Vision-Language, Multimodal Deep Learning
- AI for Defense (write me if you want to chat about this topic)
Publications
Datasets
- dacl10k (Toolkit, arXiv, Space, WACV Workshop, Voxel51) – in collaboration with the Institute for Structural Engineering: Multi-label semantic segmentation dataset for 19 classes (13 damages and 6 objects) present on bridges
- InpaintCOCO (arXiv, ALVR@ACL) – in collaboration with Charles University Prague: Vision-Language dataset for fine-grained understanding and retrieval
GPU cluster "Monacum One"
- Please use the (internal) Ilias group to sign up for cluster usage and/or for more details.
Student thesis/student assistant
- Write to me if you want to write a bachelor's/master's thesis in the field of Deep Learning, Vision-Language, or AI for defense.
- We are looking for student assistants (with or without a bachelor's degree) to work about 8 hours per week. Required skills include programming knowledge, a structured approach to work, reliability, and logical thinking.