About Me

I’m Michal Balcerak, a PhD Candidate in AI at the University of Zurich and ETH AI Center, advised by Prof. Bjoern Menze. My research focuses on probabilistic inference for generative modeling and inverse problems, combining optimal transport, energy-based models, and physics-informed priors. I was a Fellow at Harvard University (2023-2024, Prof. Petros Koumoutsakos Lab), and previously worked at CERN (ATLAS) building a machine learning framework for collision topology identification at the LHC.

I’m currently based in Zürich, Switzerland.

Email: m1balcerak[at]gmail.com

Google Scholar · GitHub · LinkedIn · Twitter/X

CV · Research Statement

News

Featured talk at Frontiers of Flows for Generative AI

MFAI Workshop, Spring 2026, Carnegie Mellon University / Workshop

Energy Matching crossed 200+ GitHub stars

Highlight / Code

Research Highlights

Generative Modeling

Energy-based generative models and optimal-transport training objectives for controllable sampling, learned priors, and composable generation.

Scientific AI / Inverse Problems

Physics-informed inverse methods for multimodal medical imaging and cancer modeling, including PDE- and elasticity-based priors.

Selected Publications

Graph Energy Matching: Transport-Aligned Energy-Based Modeling for Graph Generation

Michal Balcerak, Suprosanna Shit, Chinmay Prabhakar, Sebastian Kaltenbach, Michael S. Albergo, Yilun Du^†, Bjoern Menze^†

^† equal advising.

Preprint 2026 / Project / Paper / Code coming soon

Energy Matching: Unifying Flow Matching and Energy-Based Models for Generative Modeling

Michal Balcerak, Tamaz Amiranashvili, Antonio Terpin, Suprosanna Shit, Lea Bogensperger, Sebastian Kaltenbach, Petros Koumoutsakos, Bjoern Menze

NeurIPS 2025 / Paper / Code

Individualizing Glioma Radiotherapy Planning by Optimization of a Data- and Physics-Informed Discrete Loss

Michal Balcerak, Jonas Weidner, Petr Karnakov, Ivan Ezhov, Sergey Litvinov, Petros Koumoutsakos, Tamaz Amiranashvili, Ray Zirui Zhang, John S. Lowengrub, Igor Yakushev, Benedikt Wiestler^†, Bjoern Menze^†

^† equal advising.

Nature Communications 2025 / Paper / Code

Physics-Regularized Multi-Modal Image Assimilation for Brain Tumor Localization

Michal Balcerak, Tamaz Amiranashvili, Andreas Wagner, Jonas Weidner, Petr Karnakov, Johannes C. Paetzold, Ivan Ezhov, Petros Koumoutsakos, Benedikt Wiestler^†, Bjoern Menze^†

^† equal advising.

NeurIPS 2024 / Paper / Code