Global explainability of a deep abstaining classifier

no code implementations • 1 Apr 2025 • Sayera Dhaubhadel, Jamaludin Mohd-Yusof, Benjamin H. McMahon, Trilce Estrada, Kumkum Ganguly, Adam Spannaus, John P. Gounley, Xiao-Cheng Wu, Eric B. Durbin, Heidi A. Hanson, Tanmoy Bhattacharya

We present a global explainability method to characterize sources of errors in the histology prediction task of our real-world multitask convolutional neural network (MTCNN)-based deep abstaining classifier (DAC), for automated annotation of cancer pathology reports from NCI-SEER registries.

Dimensionality Reduction