Forward Learning for Gradient-based Black-box Saliency Map Generation

TL;DR

This paper introduces a unified framework for estimating gradients in black-box models to generate saliency maps, enabling interpretability of complex, closed-source neural networks like GPT-Vision.

Contribution

It presents a novel likelihood ratio-based gradient estimation method with blockwise techniques for improved accuracy in black-box settings.

Findings

Effective gradient estimation in black-box models

Generation of accurate saliency maps for interpretability

Scalability demonstrated on GPT-Vision

Abstract

Gradient-based saliency maps are widely used to explain deep neural network decisions. However, as models become deeper and more black-box, such as in closed-source APIs like ChatGPT, computing gradients become challenging, hindering conventional explanation methods. In this work, we introduce a novel unified framework for estimating gradients in black-box settings and generating saliency maps to interpret model decisions. We employ the likelihood ratio method to estimate output-to-input gradients and utilize them for saliency map generation. Additionally, we propose blockwise computation techniques to enhance estimation accuracy. Extensive experiments in black-box settings validate the effectiveness of our method, demonstrating accurate gradient estimation and explainability of generated saliency maps. Furthermore, we showcase the scalability of our approach by applying it to explain GPT-Vision, revealing the continued relevance of gradient-based explanation methods in the era of large, closed-source, and black-box models.