I think the most significant scientific result in which machine learning is playing a major role is computational protein folding, and that at least doesn't see

I think the most significant scientific result in which machine learning is playing a major role is computational protein folding, and that at least doesn't seem to have these reproducibility problems: https://alphafold.ebi.ac.uk/ https://alphafold.ebi.ac.uk/ It's a very well-defined problem and the datasets it uses are very well-characterized (the protein crystallography database, maybe some NMR structures as well), so perhaps that helps.