Simon Schrodi PhD Student Computer Vision Lab Department of Computer Science University of Freiburg Germany Office location: Georges-Köhler-Allee, building 052, 2nd floor, room 01-041 79110 Freiburg Phone: +49 761 203 8217 E-Mail: schrodi@cs.uni-freiburg.de

Find me on:

Homepage

GitHub

Google Scholar

arXiv

DBLP

ResearchGate

Linkedin

Student thesis/project supervision

• Concept Bottleneck Models Without Predefined Concepts
• Investigating the Impact of the Stem on Model Robustness
• Probabilistic networks for climate modeling with aleatoric uncertainty
• Latent Diffusion Counterfactual Explanations (GCPR 2024)

Teaching

Summer term 2024

• Lecture: Image Processing and Computer Graphics - Teaching assistant
• Seminar: Current Works in Computer Vision - Supervisor
  Paper: Robust agents learn causal world models
• Seminar: Block-Seminar on Deep Learning - Supervisor
  Paper: Intriguing properties of generative classifiers

Winter term 2023/2024

• Lecture: Optimization - Teaching assistant
• Seminar: Current Works in Computer Vision - Supervisor
  Paper: Linear Spaces of Meanings: Compositional Structures in Vision-Language Models
• Seminar: Current Works in Computer Vision - Supervisor
  Paper: Rosetta Neurons: Mining the Common Units in a Model Zoo

Summer term 2023

• Lecture: Image Processing and Computer Graphics - Teaching assistant
• Seminar: Current Works in Computer Vision - Supervisor
  Paper: Vision Models Are More Robust And Fair When Pretrained On Uncurated Images Without Supervision
• Seminar: Block-Seminar on Deep Learning - Supervisor
  Paper: DINOv2: Learning Robust Visual Features without Supervision

Winter term 2022/2023

• Lecture: Optimization - Teaching assistant
• Seminar: Current Works in Computer Vision - Supervisor
  Paper: ComPhy: Compositional Physical Reasoning of Objects and Events from Videos
• Seminar: Block-Seminar on Deep Learning - Supervisor
  Paper: Data Distributional Properties Drive Emergent In-Context Learning in Transformers

Summer term 2022

• Seminar: Current Works in Computer Vision - Supervisor
  Paper: Self-Supervised Learning with Data Augmentations Provably Isolates Content from Style
• Seminar: Block-Seminar on Deep Learning - Supervisor
  Paper: Denoising Diffusion Implicit Models