Publications

Description of the modular machine-learning anomaly-detection framework developed for ITER’s Tokamak Systems Monitor.

Proof of concept for detecting anomalous gyrotron pulses in ITER’s Tokamak Systems Monitor using dimensionality reduction and clustering.

A machine-learning-based calibration strategy for the ITER Tokamak Systems Monitor, using Bayesian optimization and Gaussian-process regression to efficiently update finite-element models.

This study focuses on the use of dimensionality reduction techniques to assess the structural heterogeneity of glassy binary mixtures. This work is in collaboration with Daniele Coslovich and Robert L. Jack.

partycls is a Python package for the analysis of systems of interacting particles using unsupervised machine learning methods. A paper on the code, written in collaboration with Daniele Coslovich, was just published in The Journal of Open Source Software.