Characterizing Disparity Between Edge Models and High-Accuracy Base Models for Vision Tasks

TL;DR

This paper introduces XDELTA, an explainable AI tool that characterizes and explains differences between high-accuracy base models and lower-accuracy edge models for vision tasks, aiding understanding of model disparity.

Contribution

It presents a novel learning-based framework, DELTA network, with a sparsity optimization and negative correlation learning, to effectively explain model differences in edge AI applications.

Findings

XDELTA accurately explains model differences across diverse models and datasets.

The framework is effective in real-world edge deployment scenarios.

It provides geometric and concept-level insights into model disparities.

Abstract

Edge devices, with their widely varying capabilities, support a diverse range of edge AI models. This raises the question: how does an edge model differ from a high-accuracy (base) model for the same task? We introduce XDELTA, a novel explainable AI tool that explains differences between a high-accuracy base model and a computationally efficient but lower-accuracy edge model. To achieve this, we propose a learning-based approach to characterize the model difference, named the DELTA network, which complements the feature representation capability of the edge network in a compact form. To construct DELTA, we propose a sparsity optimization framework that extracts the essence of the base model to ensure compactness and sufficient feature representation capability of DELTA, and implement a negative correlation learning approach to ensure it complements the edge model. We conduct a comprehensive evaluation to test XDELTA's ability to explain model discrepancies, using over 1.2 million images and 24 models, and assessing real-world deployments with six participants. XDELTA excels in explaining differences between base and edge models (arbitrary pairs as well as compressed base models) through geometric and concept-level analysis, proving effective in real-world applications.