New taxonomy unifies explainable AI feature attribution methods

A new survey paper published on arXiv introduces a mathematical taxonomy for local additive feature attribution methods, which are crucial for explainable AI. The paper organizes various methods, including Shapley, path-based, and gradient-based approaches, under a common framework based on five key specification choices. It also details common failure modes and proposes a ten-item reporting checklist to ensure transparency and comparability of attribution results. AI

IMPACT Standardizes reporting for explainable AI methods, potentially improving reproducibility and trust in AI systems.

RANK_REASON The cluster contains a single academic paper detailing a new taxonomy for AI methods. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• alphaXiv
• arXiv
• CatalyzeX
• DagsHub
• Gotit.pub
• Hugging Face
• IArxiv
• Influence Flower
• Local Additive Feature Attribution
• ScienceCast

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →