Synthetic Quantum Electrodynamics via Generative AI

AI in Science tutorial given at AIMS, Cape Town, South Africa, 2025

Mikael Mieskolainen
Department of Physics and I-X Centre for AI in Science, Imperial College London
[email protected]

---

Environment setup

conda env create -f environment.yml
conda activate aims25
pip install -r requirements.txt

Notebooks

Part 1: Physics Theory

Part 2: AI Theory

Part 3: Surrogate Simulator

Part 4: Amortized Bayesian Posterior

Part 5: Deep Inverse Problems

See environment.yml and requirements.txt for Python and library versions.

Python files

xaims/qedgen.py     - QED 2->2 Monte Carlo event generator

xaims/flows.py      - Coupling based Normalizing Flows
xaims/splines.py    - Splines for Spline Couplings
xaims/ddpm.py       - Discrete-time DDPM diffusion
xaims/sde.py        - Continuous-time SDE diffusion

xaims/transforms.py - Pre-processing transforms
xaims/aux.py        - Auxialiary functions
xaims/visualize.py  - Plotting functions
xaims/coolplots.py  - Illustrative simulations

Reference

If you use this in your work, you can cite the repository:

@software{aims25,
  author  = "{Mikael Mieskolainen}",
  title   = "Synthetic Quantum Electrodynamics via Generative AI",
  url     = "https://github.com/mieskolainen/aims25",
  version = {0.1},
  date    = {2025-XX-YY},
}