Publications

You can also find my articles on my Google Scholar profile.

Thesis

Deep Learning for the Atomic Scale: Graph Neural Networks and Deep Generative Models with Some Applications to Materials and Molecules

Published in Linköping University Electronic Press, 2025

PhD thesis on the topic of graph neural networks and deep generative models

Recommended citation: Ekström Kelvinius, F. (2025). Deep Learning for the Atomic Scale : Graph Neural Networks and Deep Generative Models with Some Applications to Materials and Molecules (PhD dissertation, Linköping University Electronic Press). https://doi.org/10.3384/9789181181852
Download Paper

Conference Papers

WyckoffDiff – A Generative Diffusion Model for Crystal Symmetry

Published in ICML, 2025

Discrete diffusion model for generating materials as descriptions of their symmetry properties

Recommended citation: Ekström Kelvinius, F., Andersson, O.B., Parackal, A.S., Qian, D., Armiento, R. & Lindsten, F.. (2025). WyckoffDiff – A Generative Diffusion Model for Crystal Symmetry. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:15130-15147 Available from https://proceedings.mlr.press/v267/ekstrom-kelvinius25a.html.
Download Paper

Solving Linear-Gaussian Bayesian Inverse Problems with Decoupled Diffusion Sequential Monte Carlo

Published in ICML, 2025

Sequential Monte Carlo method for solving linear-Gaussian inverse problems with diffusion priors

Recommended citation: Ekström Kelvinius, F., Zhao, Z. & Lindsten, F.. (2025). Solving Linear-Gaussian Bayesian Inverse Problems with Decoupled Diffusion Sequential Monte Carlo. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:15148-15181 Available from https://proceedings.mlr.press/v267/ekstrom-kelvinius25b.html.
Download Paper

Discriminator Guidance for Autoregressive Diffusion Models

Published in The 27th International Conference on Artificial Intelligence and Statistics (AISTATS), 2024

Three versions of discriminator guidance for autoregressive (diffusion) models

Recommended citation: Ekström Kelvinius, F. & Lindsten, F.. (2024). Discriminator Guidance for Autoregressive Diffusion Models. Proceedings of The 27th International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 238:3403-3411 Available from https://proceedings.mlr.press/v238/ekstrom-kelvinius24a.html.
Download Paper

Accelerating Molecular Graph Neural Networks via Knowledge Distillation

Published in Thirty-Seventh Annual Conference on Neural Information Processing Systems (NeurIPS 2023), 2023

Using knowledge distillation for improving performance of molecular graph neural networks

Recommended citation: Ekström Kelvinius, F., Georgiev, D., Toshev, A., & Gasteiger, J. (2024). Accelerating molecular graph neural networks via knowledge distillation. Advances in Neural Information Processing Systems, 36.
Download Paper | Download Slides

Journal Articles

Graph-based machine learning beyond stable materials and relaxed crystal structures

Published in Physical Review Materials, 2022

Evaluating the performance of a GNN for crystals in a high-throughput search setting, and proposing an ML-assisted workflow

Recommended citation: Ekström Kelvinius, F., Armiento, R., & Lindsten, F. (2022). Graph-based machine learning beyond stable materials and relaxed crystal structures. Physical Review Materials, 6(3), 033801.
Download Paper

Workshop Papers

Autoregressive Diffusion Models with non-Uniform Generation Order

Published in ICML 2023 Workshop on Structured Probabilistic Inference & Generative Modeling, 2023

Investigating the generation order in autoregressive (diffusion) models for graph generation

Recommended citation: Kelvinius, F. E., & Lindsten, F. (2023). Autoregressive Diffusion Models with non-Uniform Generation Order. In ICML 2023 Workshop on Structured Probabilistic Inference & Generative Modeling.
Download Paper

Filip Ekström Kelvinius