filek51

Filip Ekström Kelvinius

PhD student

My research is in the field of machine learning were I have mainly focused on deep learning methods for graph data, with discovery of new materials as the motivating application.

Presentation

I am a PhD student at the Division of Statistics and Machine Learning. My research has focused on method development of machine learning methods, in particular deep learning, for graph-based data. The research is often interdisciplinary, with the search for new materials as the main application. My research has involved working with discriminative models like graph neural networks for prediction of material and molecular properties, but lately I have focused mostly on generative models, in particular so called diffusion models.

For more information about me, see my personal website. For a complete list of publications, see my Google Scholar.

Publications

2024

Filip Ekström Kelvinius, Fredrik Lindsten (2024) Discriminator Guidance for Autoregressive Diffusion Models Proceedings of The 27th International Conference on Artificial Intelligence and Statistics, p. 3403-3411 (Conference paper)

2023

Filip Ekström Kelvinius, Dimitar Georgiev, Artur Petrov Toshev, Johannes Gasteiger (2023) Accelerating Molecular Graph Neural Networks via Knowledge Distillation Advances in Neural Information Processing Systems 36 (NeurIPS 2023) (Conference paper)

2022

Filip Ekström Kelvinius, Rickard Armiento, Fredrik Lindsten (2022) Graph-based machine learning beyond stable materials and relaxed crystal structures Physical Review Materials, Vol. 6, Article 033801 (Article in journal) Continue to DOI

Research

About the division

Colleagues at STIMA

About the department