Jonas Geiping

Jonas Geiping

Postdoctoral Researcher

University of Maryland, College Park

Biography

Hello, I’m Jonas. I work in computer science as postdoctoral researcher at the University of Maryland, College Park. My background is in Mathematics, more specifically in mathematical optimization and I am interested in research that intersects current deep learning and mathematical optimization in general, but also in the implications of optimization in machine learning for the design of secure and private ML systems.

As such I’ve done work ranging from understanding the impact of optimization on fundamental phenomena behind generalization in deep learning, to practical optimization strategies that show vulnerabilities in user privacy for federated learning.

You can find code for our research here: github.com/JonasGeiping

For recent publications, check here: scholar.google.com/citations?user=206vNCEAAAAJ


If you have research questions, feel free to reach out anytime via email or twitter!

Interests

  • Safety, Security and Privacy in Machine Learning
  • Deep Learning as-a-Science
  • Optimization in Machine Learning
  • Trustworthy AI
  • Theory of Mathematical Optimization

Education

  • Dr. rer. nat. in Computer Science, 2016-2021

    University of Siegen

  • M.Sc. in Mathematics, 2014-2016

    University of Münster (WWU)

  • B.Sc. in Mathematics, 2011-2014

    University of Münster (WWU)

Experience

 
 
 
 
 

Post-Doctoral Associate

University of Maryland, College Park

Sep 2021 – Present
Postdoctoral research into optimization in machine learning and its implications for security and privacy with the group of Tom Goldstein.
 
 
 
 
 

Visiting Researcher

University of Maryland, College Park

Aug 2019 – Dec 2019
Visited the group of Tom Goldstein at UMD for research collaborations regarding the theory of machine learning and machine learning security.
 
 
 
 
 

Research Associate

University of Siegen - Computer Vision Group

Oct 2016 – Aug 2021
PhD Student in the group of Michael Moeller. Research into topics related to mathematical optimization in computer vision, variational methods, nonconvex optimization theory, efficient learning of energy models, bilevel optimization.
 
 
 
 
 

Research Assistant

University of Münster (WWU) - Cells-in-Motion (CiM) Cluster of Excellence

Oct 2014 – Jun 2016
Development of mathematical tools to track neural migration for the CiM flexible fund project FF-2014-06 - ”Analysis of cell-cell interactions during neuronal migration in the developing cortex by live cell imaging and cell shape quantification”
 
 
 
 
 

Student Assistant

University of Münster (WWU) - Imaging Workgroup

Feb 2014 – Aug 2014
Programming for methods in Biomedical Imaging, related to Grah, Joana Sarah et al., ”Mathematical imaging methods for mitosis analysis in live-cell phase contrast microscopy.” Methods 115 (2017): 91-99

Recent Publications

A Simple Strategy to Provable Invariance via Orbit Mapping
Autoregressive Perturbations for Data Poisoning
Diffusion Art or Digital Forgery? Investigating Data Replication in Diffusion Models
Panning for Gold in Federated Learning: Targeted Text Extraction under Arbitrarily Large-Scale Aggregation

Contact