[论文解读] Locally Interpretable Models and Effects based on Supervised Partitioning (LIME-SUP)

Linwei Hu, Jie Chen+2 more|arXiv (Cornell University)|Jun 2, 2018

Explainable Artificial Intelligence (XAI)参考文献 16被引用 41

一句话总结

本论文介绍 LIME-SUP，一种本地可解释的建模框架，利用监督分区（树）来通过对拟合响应及其导数进行建模来解释拟合的机器学习模型，并通过仿真实验和真实数据与 KLIME 进行比较。

ABSTRACT

Supervised Machine Learning (SML) algorithms such as Gradient Boosting, Random Forest, and Neural Networks have become popular in recent years due to their increased predictive performance over traditional statistical methods. This is especially true with large data sets (millions or more observations and hundreds to thousands of predictors). However, the complexity of the SML models makes them opaque and hard to interpret without additional tools. There has been a lot of interest recently in developing global and local diagnostics for interpreting and explaining SML models. In this paper, we propose locally interpretable models and effects based on supervised partitioning (trees) referred to as LIME-SUP. This is in contrast with the KLIME approach that is based on clustering the predictor space. We describe LIME-SUP based on fitting trees to the fitted response (LIM-SUP-R) as well as the derivatives of the fitted response (LIME-SUP-D). We compare the results with KLIME and describe its advantages using simulation and real data.

研究动机与目标

• 解决对大规模数据集上高性能监督学习模型（GBM、随机森林、神经网络）的不透明性。
• 通过监督分区（树）开发本地可解释的模型和效应。
• 提供两种变体：LIM-SUP-R（拟合响应）和 LIME-SUP-D（拟合响应的导数）。
• 将 LIME-SUP 与 KLIME 进行比较，并通过仿真和真实数据评估其益处。

提出的方法

• 对拟合响应拟合树以获得本地可解释的解释（LIM-SUP-R）。
• 对拟合响应的导数拟合树以捕捉局部效应（LIME-SUP-D）。
• 将 LIME-SUP 与依赖于将预测变量空间聚类的 KLIME 方法进行对比。
• 通过仿真研究和真实数据实验进行评估。

实验结果

研究问题

• RQ1 监督分区是否能够为复杂的 SML 模型提供本地可解释的解释？
• RQ2 就可解释性和性能而言，LIM-SUP-R 与 LIME-SUP-D 与 KLIME 方法有何比较？
• RQ3 在拟合响应与其导数上使用基于树的分区有哪些优点？
• RQ4 仿真和真实数据实验是否显示出 LIME-SUP 相对于 KLIME 的实际收益？

主要发现

• LIME-SUP 通过基于拟合响应及其导数的树对预测变量空间进行分区，从而提供本地可解释的模型。
• 将 LIME-SUP 与 KLIME 进行对比，突出在仿真和真实数据实验中发现的优势。