报告题目：Designing molecules for high-harmonic generation with machine learning
报告人：Prof. Camilo Ruiz
单 位：University of Salamanca
主持人：倪宏程 研究员
时间：2026年11月14日（周六）14:00至15:00
地点：闵行校区光学大楼B225会议室
报告摘要：
High-harmonic generation (HHG) in molecules depends strongly on the emitting orbital, yet chemical space is discrete and each spectrum is expensive. We present a workflow that searches small C/N/O molecules for bright harmonic emission and explains why they are bright. A SELFIES variational autoencoder trained on QM9 provides a continuous 32-dimensional latent space, explored by trust-region Bayesian optimisation. Spectra come from an analytic strong-field approximation with transverse-momentum averaging, validated against TDSE references. An antenna picture, the band strength of the projected orbital dipole, predicts the plateau and symmetry-forbidden orientations. The first optimisation campaign already finds brighter long-chain molecules, notably heptanal.
报告人简介：
Camilo Ruiz is Associate Professor at the Universidad de Salamanca and PI of the EMC3 research group. He was a postdoc at the Max Planck Institute in Dresden and Imperial College London. His interests span strong-field and attosecond physics, laser-plasma interaction and climate-change education. He is currently interested in machine learning for strong-field physics.