HarsanyiNet: Computing Accurate Shapley Values in a Single Forward Propagation
Lu Chen, Siyu Lou, Keyan Zhang, Jin Huang, Quanshi Zhang
Abstract
The Shapley value is widely regarded as a trustworthy attribution metric. However, when people use Shapley values to explain the attribution of input variables of a deep neural network (DNN), it usually requires a very high computational cost to approximate relatively accurate Shapley values in real-world applications. Therefore, we propose a novel network architecture, the HarsanyiNet, which makes inferences on the input sample and simultaneously computes the exact Shapley values of the input variables in a single forward propagation. The HarsanyiNet is designed on the theoretical foundation that the Shapley value can be reformulated as the redistribution of Harsanyi interactions encoded by the network.
BibTeX
@inproceedings{icml2023_harsanyinetcompu,
title = {HarsanyiNet: Computing Accurate Shapley Values in a Single Forward Propagation},
author = {Lu Chen and Siyu Lou and Keyan Zhang and Jin Huang and Quanshi Zhang},
booktitle = {ICML 2023},
year = {2023}
}