RESEARCH
We develop machine learning and quantum dynamics methods to explore light-matter interactions with applications in quantum spectroscopy and technologies.
Machine Learning Quantum Light-Matter​
​
Do we understand what a neural network learns? More importantly, what does it mean to understand in this context. Despite unprecedented success in predictive performance and broad applicability across disciplines, this question remains largely unexplored.
​
We explore light-matter interactions by treating both light and matter quantum mechanically. In several regimes, such a description is essential and gives rise to phenomena that are fundamentally distinct from semiclassical methods. ​Our research leverages machine learning (ML) approaches for modeling the dynamics efficiently. Beyond predictive performance, we place particular emphasis on the interpretability of ML models, with a goal of understanding what neural networks learn and how their internal representations connect to the fundamental principles.
​​"The purpose of models is not to fit the data but to sharpen the questions."
— Samuel Karlin
Nanoscale Quantum Light-Matter
​​​
Noble metal nanoparticles provide a potential platform for manipulating light-matter interactions at the nanoscale, owing to their wavelength tunable light absorption through surface plasmon resonances. We explore the quantum dynamics that emerge at the interfaces of these plasmonic structures, where plasmons can strongly interact with nearby systems. Our research aims to uncover the quantum aspects of such light-matter interactions - with a particular emphasis on the understanding of the influence of vacuum fluctuations, quantum coherence and entanglement on light induced processes. To model these effects, we develop and apply methods from open quantum dynamics and machine learning.