Scholar
Markus Marks
Google Scholar ID: XkVkcywAAAAJ
Research Scientist at FAIR (Meta AI)
Citations & Impact
Citations
390
H-index
8
i10-index
8
Publications
20
Co-authors
22
list available
Publications
Resume (English only)
Academic Achievements
- Publications:
- - Diffusion-Based Action Recognition Generalizes to Untrained Domains
- - SAVeD: Learning to Denoise Low-SNR Video for Improved Downstream Performance
- - Probing the Mid-level Vision Capabilities of Self-Supervised Learning
- - Learning Keypoints for Multi-Agent Behavior Analysis using Self-Supervision
- - A Closer Look at Benchmarking Self-Supervised Pre-training with Image Classification
- - Text-image Alignment for Diffusion-based Perception
- - A Foundation Model for Cell Segmentation
- - MABe22: A Multi-Species Multi-Task Benchmark for Learned Representations of Behavior
- - Deep-learning-based identification, tracking, pose estimation and behaviour classification of interacting primates and mice in complex environments
- - Robust Disentanglement of a Few Factors at a Time using rPU-VAE
Research Experience
- Currently a Research Scientist at FAIR (AI at Meta) working on fundamental computer vision. Previously, a postdoctoral scholar in the Vision Group at Caltech, working with Pietro Perona.
Education
- Ph.D. from ETH Zurich, supervised by Mehmet Yanik, Valerio Mante, Angelika Steger, and Benjamin Grewe.
Background
- Research Interests: Machine learning and computer vision, particularly in their applications to scientific discovery in biology and medicine. Leveraging large-scale models and self-supervision to manage and interpret the rapidly growing volume of (biomedical) data. Focuses on reducing human effort in data annotation, minimizing bias, revealing hidden patterns in biomedical data, and developing new methods along the way.
