SMBP Group Meeting: Isak (John) Falk

Tuesday Apr 2, 2024, 2:00 PM → 3:00 PM America/New_York

3rd Floor Classroom/3-Flatiron Institute (162 5th Avenue)

Speaker: Isak (John) Falk

Topic: Transfer learning for atomistic simulations using GNNs and kernel mean embeddings

Abstract: Interatomic potentials learned using machine learning methods have been successfully applied to atomistic simulations. However, accurate models require large training datasets, while generating reference calculations is computationally demanding. To bypass this difficulty, we propose a transfer learning algorithm that leverages the ability of graph neural networks to represent chemical environments together with kernel mean embeddings. We extract a feature map from GNNs pre-trained on the OC20 dataset and use it to learn the potential energy surface from system-specific datasets of catalytic processes. Our method is further enhanced by incorporating into the kernel the chemical species information, resulting in improved performance and interpretability. We test our approach on a series of realistic datasets of increasing complexity, showing excellent generalization and transferability performance, and improving on methods that rely on GNNs or ridge regression alone, as well as similar fine-tuning approaches.Bio: Isak (John) Falk recently got his PhD in machine learning at UCL where he worked on transfer and meta-learning, in particular understanding fundamental questions in few-shot learning. Towards the end of his PhD he has collaborated with other scientists in applying ML to science, in particular genetics and atomistic systems.

Bio: Isak (John) Falk recently got his PhD in machine learning at UCL where he was supervised by Massimiliano Pontil and Carlo Ciliberto and worked on transfer and meta-learning, in particular understanding fundamental questions in few-shot learning. Towards the end of his PhD he collaborated with other scientists in applying ML to science, in particular genetics and atomistic systems.