Researchers at PNNL have developed an interpretable, lightweight AI model that can easily predict weld microstructure features using only basic machine sensor inputs.

Spatial proteomics enables researchers to link protein measurements to features in the image of a tissue sample, which are lost using standard approaches.