A Comprehensive and Reliable Feature Attribution Method: Double-sided Remove and Reconstruct (DoRaR)

1 code implementation • 27 Oct 2023 • Dong Qin, George Amariucai, Daji Qiao, Yong Guan, Shen Fu

While avoiding the artifacts problem, this new category suffers from the Encoding Prediction in the Explanation (EPITE) problem, in which the predictor's decisions rely not on the features, but on the masks that selects those features.

Decision Making