QUANTUM PHYSICS × MACHINE LEARNING

Yuan Wang

Google Scholar • ORCID iD iconORCiD: 0000-0002-0688-3276 • github.com/ywang-phy
ywang.org • yuan.wang.phy@gmail.com • linkedin.com/in/ywang-phy

I am passionate about quantum light-matter interaction, parallel computing, and AI4Science: I develop GPU-based computational tools for quantum systems, apply machine learning to accelerate scientific discovery, and leverage physical principles to enhance machine-learning architectures (Science4AI).

Research

GPU-accelerated quantum simulation

CPU- and GPU-based Gross-Pitaevskii equation (GPE) solvers in C++ and CUDA, built around the split-step Fourier method (SSFM), for fast and accurate simulation of quantum fluids of light such as exciton-polariton condensates.

CUDAC++SSFMGPE

Neural operators for condensate physics

A Fourier neural operator approach, trained on numerical and experimental data, that solves the polariton GPE coupled with exciton rate equations almost three orders of magnitude faster than CUDA-based solvers while maintaining high accuracy.

AI4ScienceFNOPolariton Condensation

Photonics-enhanced machine learning

Physics-informed positional embeddings (PEs) for graph neural networks (GNN) are derived from light propagation on synthetic frequency lattices whose couplings match the input graph. These internode intensity correlations supply GCNs with global structural information beyond local message passing, supporting the optical acceleration of graph ML.

Science4AIGNNPEsSynthetic Lattices

Generative modelling with polariton condensates

Room-temperature exciton-polariton condensates in organic microcavities are employed as a physical stochastic nonlinear layer within a conditional generative adversarial network (GAN). Their nonlinear many-body dynamics and shot-to-shot fluctuations supply structured variability that stabilises adversarial training and surpasses digital sampling baselines on conditional digit-to-image translation.

Science4AIGANPolariton Condensation

Exciton-polariton condensates

Theory of enhancing and focusing polariton condensates in semiconductor microcavities via tailored non-resonant excitation: increasing spatial coherence and interaction strength toward all-optical transistors and large-scale condensate networks.

MicrocavitiesSemiconductorsPolariton Condensation

Experience

Education

Internships

Talks & Activities

Skills & Languages

Workshop

Journal/Preprint

†corresponding author