English

Shapley Values with Uncertain Value Functions

Machine Learning 2023-04-03 v1 Machine Learning

Abstract

We propose a novel definition of Shapley values with uncertain value functions based on first principles using probability theory. Such uncertain value functions can arise in the context of explainable machine learning as a result of non-deterministic algorithms. We show that random effects can in fact be absorbed into a Shapley value with a noiseless but shifted value function. Hence, Shapley values with uncertain value functions can be used in analogy to regular Shapley values. However, their reliable evaluation typically requires more computational effort.

Keywords

Cite

@article{arxiv.2301.08086,
  title  = {Shapley Values with Uncertain Value Functions},
  author = {Raoul Heese and Sascha Mücke and Matthias Jakobs and Thore Gerlach and Nico Piatkowski},
  journal= {arXiv preprint arXiv:2301.08086},
  year   = {2023}
}

Comments

12 pages, 1 figure, 1 table

R2 v1 2026-06-28T08:15:23.783Z