Depends heavily on the kind (and intensity) of radiation. Beta (electron/positron) and gamma (photon) generally won’t, but neutron and alpha can. Many of the atoms that become radioactive will rapidly decay, and that’s one of the mechanisms behind the impact to structural integrity.

Exactly.

The general approach is to use interpretable models where you can understand how the model works and what features it uses to discriminate, but that doesn’t work for all ML approaches (and even when it does our understanding is incomplete.)

Maybe not the hardest, but still challenging. Unknown biases in training data are a challenge in any experimental design. Opaque ML frequently makes them more challenging to discover.

Love the cheese name! I think he looks like more like a Muenster to me, but Colby works! https://en.m.wikipedia.org/wiki/Muenster_cheese