Pith. sign in

REVIEW 4 cited by

Problems with Shapley-value-based explanations as feature importance measures

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2002.11097 v2 pith:HDVHCXPS submitted 2020-02-25 cs.AI cs.LGstat.ML

classification cs.AIcs.LGstat.ML
keywords explanationsfeatureimportanceshapleyvaluesgamemathematicalmethods
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Game-theoretic formulations of feature importance have become popular as a way to "explain" machine learning models. These methods define a cooperative game between the features of a model and distribute influence among these input elements using some form of the game's unique Shapley values. Justification for these methods rests on two pillars: their desirable mathematical properties, and their applicability to specific motivations for explanations. We show that mathematical problems arise when Shapley values are used for feature importance and that the solutions to mitigate these necessarily induce further complexity, such as the need for causal reasoning. We also draw on additional literature to argue that Shapley values do not provide explanations which suit human-centric goals of explainability.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 4 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Rule of Thumb: Explaining Artificial Intelligence Systems using Partial Information

    cs.AI 2026-08 conditional novelty 6.0 of 10

    ROT explains individual AI predictions by fitting a single additive model with feature dropout to observed input-output pairs, yielding feature importances based on predictiveness rather than perturbation.

  2. Unifying Attribution-Based Explanations Using Functional Decomposition

    cs.LG 2024-12 reject novelty 6.0 of 10

    A unification framework for XAI attribution methods whose core canonical decomposition theorem fails because the components sum to the fully removed function rather than to the original function.

  3. On Model Extrapolation in Marginal Shapley Values

    stat.ML 2024-12 conditional novelty 5.0 of 10

    Stratifying marginal Shapley by feature region avoids off-manifold extrapolation and, with a causal direction and a chosen reference constant, reproduces causal Shapley values on a linear spline and an insurance example.

  4. A Comprehensive Guide to Explainable AI: From Classical Models to LLMs

    cs.LG 2024-12 unverdicted novelty 1.0 of 10

    A survey-style XAI book with code examples, covering standard interpretability methods and models, but no new scientific contributions.

Pith tools