[論文レビュー] Enhancing LIME using Neural Decision Trees

Mohamed Aymen Bouyahia, Argyris Kalogeratos|arXiv (Cornell University)|Mar 21, 2026

Explainable Artificial Intelligence (XAI)被引用数 0

ひとこと要約

The paper introduces NDT-LIME, a LIME variant that uses Neural Decision Trees as local surrogates to improve fidelity and interpretability for tabular data. It shows higher local fidelity than traditional LIME surrogates across benchmark datasets.

ABSTRACT

Interpreting complex machine learning models is a critical challenge, especially for tabular data where model transparency is paramount. Local Interpretable Model-Agnostic Explanations (LIME) has been a very popular framework for interpretable machine learning, also inspiring many extensions. While traditional surrogate models used in LIME variants (e.g. linear regression and decision trees) offer a degree of stability, they can struggle to faithfully capture the complex non-linear decision boundaries that are inherent in many sophisticated black-box models. This work contributes toward bridging the gap between high predictive performance and interpretable decision-making. Specifically, we propose the NDT-LIME variant that integrates Neural Decision Trees (NDTs) as surrogate models. By leveraging the structured, hierarchical nature of NDTs, our approach aims at providing more accurate and meaningful local explanations. We evaluate its effectiveness on several benchmark tabular datasets, showing consistent improvements in explanation fidelity over traditional LIME surrogates.

研究の動機と目的

• Bridge the gap between high predictive performance and interpretable local explanations for tabular data.
• Integrate Neural Decision Trees (NDTs) as local surrogate models within the LIME framework.
• Demonstrate improvements in explanation fidelity and stability over traditional LIME surrogates on benchmark datasets.
• Provide theoretical and empirical evidence of smoother, more faithful local explanations.
• Offer open-source implementation to facilitate adoption and further research.

提案手法

• Replace standard surrogates in LIME with Neural Decision Trees (NDTs) as local explainers.
• Initialize NDTs from a conventional decision tree trained on local neighborhoods (warm-start).
• Fine-tune NDTs with gradient-based optimization to minimize a local fidelity loss using perturbed samples: L_fidelity = sum_i pi(x_i, x) ||f(x_i) - g(x_i)||^2, where pi is a Gaussian kernel in input space.
• Represent DTs as NDTs via a three-layer neural network with two hidden layers and one output layer, enabling differentiable splits.
• Extract explanations from the trained NDT (e.g., feature importance or rules) by tracing paths to leaves and computing gradients.
• Compare NDT-LIME against LR-LIME and DT-LIME across eight benchmark datasets using fidelity, stability, and regularity metrics.

実験結果

リサーチクエスチョン

• RQ1 Can Neural Decision Trees as LIME surrogates improve local fidelity compared to linear or greedy DT surrogates?
• RQ2 Do NDT-based explanations provide more stable and regular local explanations under perturbations?
• RQ3 How does NDT-LIME perform across varied tabular datasets in terms of fidelity, stability, and regularity?
• RQ4 What is the impact of initializing NDTs from traditional decision trees on explanation quality?
• RQ5 Is the added computational cost justified by gains in explanation quality?

主な発見

Dataset	Stability LR-LIME	Stability DT-LIME	Stability NDT-LIME	Fidelity LR-LIME	Fidelity DT-LIME	Fidelity NDT-LIME	Regularity LR-LIME	Regularity DT-LIME	Regularity NDT-LIME
Breast Cancer	0.997 ± 0.003	0.986 ± 0.103	0.991 ± 0.004	0.527 ± 0.085	0.686 ± 0.049	0.785 ± 0.031	0.812 ± 0.014	0.873 ± 0.005	0.915 ± 0.024
Iris	0.994 ± 0.006	0.997 ± 0.003	0.943 ± 0.010	0.554 ± 0.150	0.777 ± 0.037	0.860 ± 0.021	0.743 ± 0.017	0.813 ± 0.028	0.820 ± 0.039
Wine	0.999 ± 0.001	0.997 ± 0.001	0.998 ± 0.002	0.321 ± 0.133	0.395 ± 0.183	0.518 ± 0.131	0.830 ± 0.019	0.864 ± 0.042	0.910 ± 0.023
Digits	0.980 ± 0.009	0.961 ± 0.091	0.816 ± 0.023	0.243 ± 0.141	0.440 ± 0.076	0.577 ± 0.106	0.654 ± 0.021	0.512 ± 0.084	0.563 ± 0.034
Covtype	0.984 ± 0.005	0.983 ± 0.108	0.931 ± 0.007	0.362 ± 0.052	0.556 ± 0.110	0.632 ± 0.067	0.890 ± 0.012	0.926 ± 0.020	0.849 ± 0.072
California Housing	0.999 ± 0.000	0.999 ± 0.000	0.973 ± 0.001	0.297 ± 0.148	0.890 ± 0.022	0.960 ± 0.014	0.834 ± 0.009	0.864 ± 0.011	0.794 ± 0.033
Diabetes	0.999 ± 0.001	0.999 ± 0.001	0.998 ± 0.002	0.886 ± 0.012	0.562 ± 0.124	0.920 ± 0.035	0.956 ± 0.010	0.794 ± 0.102	0.978 ± 0.017
Ames Housing	0.998 ± 0.000	0.996 ± 0.041	0.990 ± 0.013	0.865 ± 0.030	0.506 ± 0.062	0.713 ± 0.054	0.976 ± 0.008	0.903 ± 0.012	0.893 ± 0.027

• NDT-LIME generally achieves higher fidelity (R^2) than LR-LIME and DT-LIME across most datasets, with notable gains on California Housing and Diabetes.
• NDT-LIME delivers strong stability comparable to baselines, with some datasets showing slightly lower stability due to greater expressive power.
• NDT-LIME achieves higher regularity, indicating smoother explanations across nearby data points, especially relative to linear surrogates.
• Across datasets, NDT-LIME provides a favourable balance between interpretability and fidelity by leveraging the expressive power of NDTs.
• The authors provide open-source code for NDT-LIME, facilitating replication and further experimentation.