Quadrature-TreeSHAP offers faster, more stable AI model explanations

Researchers have developed Quadrature-TreeSHAP, a novel method for explaining tree ensemble predictions that is depth-independent and more numerically stable than existing approaches. This new technique extends naturally to higher-order Shapley interaction values and utilizes a quadrature-based reformulation for efficient computation. Empirical evaluations show Quadrature-TreeSHAP significantly outperforms TreeSHAP and GPUTreeSHAP in speed for both Shapley values and interaction calculations. AI

IMPACT Introduces a more efficient and stable method for explaining tree-based models, potentially improving interpretability in ML applications.

RANK_REASON This is a research paper introducing a new method for explaining machine learning model predictions. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• Banzhaf interaction polynomial
• GPUTreeSHAP
• Quadrature-TreeSHAP
• TreeSHAP
• XGBoost
• Shapley values

• paper
• other

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