ITL-LIME: Instance-Based Transfer Learning for Enhancing Local Explanations in Low-Resource Data Settings

TL;DR

This paper introduces ITL-LIME, a transfer learning-enhanced local explanation method that improves stability and fidelity of explanations in low-resource settings by leveraging relevant source domain instances.

Contribution

ITL-LIME integrates instance transfer learning into LIME, utilizing source domain prototypes and contrastive learning to generate more accurate and stable local explanations with limited data.

Findings

Enhanced explanation stability in low-resource scenarios

Improved fidelity of surrogate models

Effective use of source domain instances for explanations

Abstract

Explainable Artificial Intelligence (XAI) methods, such as Local Interpretable Model-Agnostic Explanations (LIME), have advanced the interpretability of black-box machine learning models by approximating their behavior locally using interpretable surrogate models. However, LIME's inherent randomness in perturbation and sampling can lead to locality and instability issues, especially in scenarios with limited training data. In such cases, data scarcity can result in the generation of unrealistic variations and samples that deviate from the true data manifold. Consequently, the surrogate model may fail to accurately approximate the complex decision boundary of the original model. To address these challenges, we propose a novel Instance-based Transfer Learning LIME framework (ITL-LIME) that enhances explanation fidelity and stability in data-constrained environments. ITL-LIME introduces instance transfer learning into the LIME framework by leveraging relevant real instances from a related source domain to aid the explanation process in the target domain. Specifically, we employ clustering to partition the source domain into clusters with representative prototypes. Instead of generating random perturbations, our method retrieves pertinent real source instances from the source cluster whose prototype is most similar to the target instance. These are then combined with the target instance's neighboring real instances. To define a compact locality, we further construct a contrastive learning-based encoder as a weighting mechanism to assign weights to the instances from the combined set based on their proximity to the target instance. Finally, these weighted source and target instances are used to train the surrogate model for explanation purposes.