I am working on an ML-based problem for my undergrad thesis, based on an experimental dataset and prediction results published in a paper two years ago by an associate prof. in a very high ranking journal - the paper has 200+ citations. For months, I'd been trying to replicate the paper's results...