news

Apr 05, 2026 Paper accepted in Medical Physics: A parallel algorithm for generating Pareto-optimal radiosurgery treatment plans. This is a result of our collaboration with Elekta.
Jan 22, 2026 Two papers accepted at L4DC 2026: Warm-starting active-set solvers using graph neural networks (led by Ella J. Schmidtobreick) and Learning to accelerate distributed ADMM using graph neural networks (led by Henri Doerks and Paul Häusner).
Nov 01, 2025 Received a VINNOVA planning grant for Automate Sweden, a cluster of excellence on robotics bringing together leading Swedish universities and industry partners under a shared Physical AI vision.
Mar 01, 2025 New paper in npj Computational Materials: Navigating chemical design spaces for metal-ion batteries via ML-guided phase-field simulations. Combines phase-field modelling with Bayesian optimization to study electrode/electrolyte interface reactions. Work led by Liam (Hamed) Taghavian.
Feb 01, 2025 New paper in Cell Reports Physical Science: Accelerating aqueous electrolyte design with automated full-cell battery experimentation and Bayesian optimization. Demonstrates a self-driving lab for high-throughput battery electrolyte screening. Work led by Jackie Yik.
Jan 15, 2025 Two review articles published in Philosophical Transactions of the Royal Society A: Conditional sampling within generative diffusion models and Taming diffusion models for image restoration: a review. Work led by Zheng Zhao and Ziwei Luo.
Dec 01, 2024 New paper at NeurIPS 2024: Entropy-regularized diffusion policy with Q-ensembles for offline reinforcement learning. Combines diffusion models with entropy regularization for offline RL. Work led by Ruoqi Zhang.
Sep 15, 2024 New paper at ICLR 2024: Controlling vision-language models for multi-task image restoration. DA-CLIP adapts pretrained vision-language models for unified image restoration. Work led by Ziwei Luo.
Sep 01, 2024 New paper in TMLR: Neural incomplete factorization: learning preconditioners for the conjugate gradient method. A GNN that learns sparse matrix factorizations to precondition iterative solvers. Work led by Paul Häusner.
Jun 15, 2024 Received a Swedish Research Council (VR) starting grant for the project Accelerating sparse linear algebra with graph neural networks.
Jun 01, 2024 Awarded the Göran Gustafsson prize for young researchers in natural sciences and technology.