Machine Learning and Science: It's All Fun and Games Until CUDA Crashes (Constructor Technology)

30.08.2023 10:00

Recently we published a paper where we developed a machine learning algorithm for predicting properties of 2D materials. The core idea of the paper was invented and tested in about a week. All the remaining 1.5 years went into writing the code to generate the data, writing the code to process it, writing the code to implement baselines, writing the code to do hyperparameter search (which didn’t change the conclusions in the slightest, we still won by a large margin). And then running, debugging and re-running on several different HPC platforms.

We'll discuss the technical challenges of conducting computational research labour-efficiently, and the ways to address them, followed by the demonstration of our project on the Constructor Research Platform.

Lieu

Bâtiment: Conseil Général 7-9

Room 1-15, Séminaire d'analyse numérique

entrée libre