Model editing for distribution shifts in uranium oxide morphological analysis
This addresses a domain-specific issue for materials science researchers dealing with instrument-induced distribution shifts, but it is incremental as it applies an existing method to a new dataset.
The paper tackled the problem of deep learning models struggling with distribution shifts in scientific data, specifically for classifying uranium ore concentrate synthesis conditions, and found that model editing outperformed finetuning on datasets with humidity-induced aging and different microscope instruments.
Deep learning still struggles with certain kinds of scientific data. Notably, pretraining data may not provide coverage of relevant distribution shifts (e.g., shifts induced via the use of different measurement instruments). We consider deep learning models trained to classify the synthesis conditions of uranium ore concentrates (UOCs) and show that model editing is particularly effective for improving generalization to distribution shifts common in this domain. In particular, model editing outperforms finetuning on two curated datasets comprising of micrographs taken of U$_{3}$O$_{8}$ aged in humidity chambers and micrographs acquired with different scanning electron microscopes, respectively.