Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A Comparative Study of Explainable AI Methods: Model-Agnostic vs. Model-Specific Approaches

Keerthi Devireddy

TL;DR

This paper tackles the problem of interpreting deep learning image classifiers by comparing model-agnostic (LIME, SHAP) and model-specific (Grad-CAM, Guided Backpropagation) XAI methods. The authors use ResNet50 across seven diverse species to evaluate how different explanations reflect the model's decision-making. They find that each method reveals distinct aspects of predictions, with broad attributions from agnostic methods and precise localization from specific methods, and conclude that a hybrid approach yields the most robust explanations. The work provides practical guidance for selecting interpretability techniques under computational constraints and lays out future directions for scalable, human-centered XAI across architectures.

Abstract

This paper compares model-agnostic and model-specific approaches to explainable AI (XAI) in deep learning image classification. I examine how LIME and SHAP (model-agnostic methods) differ from Grad-CAM and Guided Backpropagation (model-specific methods) when interpreting ResNet50 predictions across diverse image categories. Through extensive testing with various species from dogs and birds to insects I found that each method reveals different aspects of the models decision-making process. Model-agnostic techniques provide broader feature attribution that works across different architectures, while model-specific approaches excel at highlighting precise activation regions with greater computational efficiency. My analysis shows there is no "one-size-fits-all" solution for model interpretability. Instead, combining multiple XAI methods offers the most comprehensive understanding of complex models particularly valuable in high-stakes domains like healthcare, autonomous vehicles, and financial services where transparency is crucial. This comparative framework provides practical guidance for selecting appropriate interpretability techniques based on specific application needs and computational constraints.

A Comparative Study of Explainable AI Methods: Model-Agnostic vs. Model-Specific Approaches

TL;DR

This paper tackles the problem of interpreting deep learning image classifiers by comparing model-agnostic (LIME, SHAP) and model-specific (Grad-CAM, Guided Backpropagation) XAI methods. The authors use ResNet50 across seven diverse species to evaluate how different explanations reflect the model's decision-making. They find that each method reveals distinct aspects of predictions, with broad attributions from agnostic methods and precise localization from specific methods, and conclude that a hybrid approach yields the most robust explanations. The work provides practical guidance for selecting interpretability techniques under computational constraints and lays out future directions for scalable, human-centered XAI across architectures.

Abstract

This paper compares model-agnostic and model-specific approaches to explainable AI (XAI) in deep learning image classification. I examine how LIME and SHAP (model-agnostic methods) differ from Grad-CAM and Guided Backpropagation (model-specific methods) when interpreting ResNet50 predictions across diverse image categories. Through extensive testing with various species from dogs and birds to insects I found that each method reveals different aspects of the models decision-making process. Model-agnostic techniques provide broader feature attribution that works across different architectures, while model-specific approaches excel at highlighting precise activation regions with greater computational efficiency. My analysis shows there is no "one-size-fits-all" solution for model interpretability. Instead, combining multiple XAI methods offers the most comprehensive understanding of complex models particularly valuable in high-stakes domains like healthcare, autonomous vehicles, and financial services where transparency is crucial. This comparative framework provides practical guidance for selecting appropriate interpretability techniques based on specific application needs and computational constraints.

Paper Structure

This paper contains 4 sections, 7 equations, 1 figure, 2 tables.

Table of Contents

  1. Introduction to Explainable AI (XAI)
  2. Model-Agnostic vs. Model-Specific XAI
  3. Comparative Analysis of XAI Methods
  4. Conclusion & Future Work

Figures (1)

  • Figure 1: Comparative visualizations of XAI methods applied to different species. Each row represents a species, and each column represents a different XAI method.