The real story here is that ANL needs to fact check their press release. For a review of (previous) uses of machine learning in particle physics, see:

- https://home.cern/news/news/physics/speeding-machine-learnin...

- https://www.nature.com/articles/s42254-022-00455-1

- https://www.nature.com/articles/s41586-018-0361-2

Among many many others.

The paper in this post might have been the first to use an autoencoder trained on collider data, but that qualification is buried further down.