Research Associate · Bavarian Center for Battery Technology (BayBatt)
University of Bayreuth, Germany
Research Associate at the University of Bayreuth (BayBatt), working on how battery storage should be scheduled across reserve, day-ahead and intraday markets under price uncertainty and degradation.
I hold an MSc in Power Engineering from TUM. My thesis on graph generative models for synthetic distribution grids was published at the NeurIPS 2025 workshop Tackling Climate Change with Machine Learning. Before moving into research, I spent three years at Munich Electrification building data and test infrastructure for battery management systems.
A variational graph autoencoder with a self-designed iterative GCN decoder for generating synthetic distribution-grid topologies.
LP least-cost dispatch for a three-zone system in PyPSA, showing how interconnector capacity drives price divergence and congestion.
Linear regression, Random Forest, XGBoost, LSTM and GRU compared on SMARD data with 10-fold walk-forward validation.
Exploring Variational Graph Autoencoders for Distribution Grid Data Generation
To address the lack of public power system data for machine learning research in energy networks, we investigate the use of variational graph autoencoders (VGAEs) for synthetic distribution grid generation. Using two open-source datasets, ENGAGE and DINGO, we evaluate four decoder variants and compare generated networks against the original grids using structural and spectral metrics. Results indicate that simple decoders fail to capture realistic topologies, while GCN-based approaches achieve strong fidelity on ENGAGE but struggle on the more complex DINGO dataset, producing artifacts such as disconnected components and repeated motifs. These findings highlight both the promise and limitations of VGAEs for grid synthesis, underscoring the need for more expressive generative models and robust evaluation. We release our models and analysis as open source to support benchmarking and accelerate progress in ML-driven power system research.
@inproceedings{abbas2025vgae,
title = {Exploring Variational Graph Autoencoders for Distribution Grid Data Generation},
author = {Abbas, Syed Zain and Okoyomon, Ehimare},
booktitle = {NeurIPS 2025 Workshop on Tackling Climate Change with Machine Learning},
year = {2025},
eprint = {2509.02469},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2509.02469}
}Joined BayBatt, University of Bayreuth, as a Research Associate working on ancillary services from batteries.
NEWCompleted the MicroMasters in Data Science, UC San Diego (edX).
Presented my MSc thesis research at the NeurIPS 2025 workshop Tackling Climate Change with Machine Learning, San Diego.
Preprint on arXiv: variational graph autoencoders for distribution grid data generation.
Achieved 97% in UC San Diego’s Machine Learning Fundamentals course (edX).