Mohamed Amine Ketata

Mohamed Amine Ketata

PhD Student in Machine Learning
Technical University of Munich
a.ketata@tum.de


About Me

I am a PhD student in machine learning at the Technical University of Munich, advised by Stephan Günnemann in the Data Analytics and Machine Learning group. My PhD is supported by relAI.

My research focuses on generative modeling and foundation models for structured data, particularly tabular data, relational databases, and graphs. I’m currently working toward diffusion-based tabular foundation models with in-context generation and native support for both generative and predictive tasks.

I collaborate with SAP Business AI Research on my PhD research. Prior to my PhD, I interned at QuantCo.

I am looking for research internships in 2027 and am open to opportunities across machine learning. Feel free to get in touch.

News

Publications

* Equal contribution.

  1. Preprint Schema-conditioned diffusion architecture for numerical and categorical tabular features.
    CDMD: A Cross-Dataset Mixed-Type Diffusion Model for Tabular Data
    Mohamed Amine Ketata, Maximilian Schambach, Stephan Günnemann
    Preprint, 2026 (under review).
    Accepted at the BeNTo Workshop at NeurIPS 2026.
    PDF
  2. NeurIPS Comparison of autoregressive table generation and joint relational database generation.
    Joint Relational Database Generation via Graph-Conditional Diffusion Models
    Mohamed Amine Ketata, David Lüdke, Leo Schwinn, Stephan Günnemann
    Conference on Neural Information Processing Systems (NeurIPS), 2025
    PDF Code
  3. ICLR Molecular graphs encoded into a latent point cloud for forward and reverse diffusion.
    Lift Your Molecules: Molecular Graph Generation in Latent Euclidean Space
    Mohamed Amine Ketata*, Nicholas Gao*, Johanna Sommer*, Tom Wollschläger, Stephan Günnemann
    International Conference on Learning Representations (ICLR), 2025
    Oral presentation and Best Paper Runner-up at the GRaM Workshop at ICML 2024.
    PDF Talk
  4. ICML Molecular uncertainty model combining a graph neural network, latent space, and Gaussian process.
    Uncertainty Estimation for Molecules: Desiderata and Methods
    Tom Wollschläger, Nicholas Gao, Bertrand Charpentier, Mohamed Amine Ketata, Stephan Günnemann
    International Conference on Machine Learning (ICML), 2023
    PDF Code
  5. ICLR Workshop Protein docking through reverse diffusion, followed by confidence-based pose selection.
    DiffDock-PP: Rigid Protein-Protein Docking with Diffusion Models
    Mohamed Amine Ketata*, Cedrik Laue*, Ruslan Mammadov*, Hannes Stärk, Menghua Wu, Gabriele Corso, Céline Marquet, Regina Barzilay, Tommi S. Jaakkola
    Machine Learning for Drug Discovery Workshop at the International Conference on Learning Representations (ICLR), 2023
    PDF Code

Teaching & Supervision

At TUM, I have led exercise sessions and co-organize a research seminar:

I also supervise theses and guided research on generative models and machine learning for tabular, relational, and graph data.

Academic Service

Conference Reviewer

Workshop Reviewer


Powered by Jekyll and Minimal Light theme.