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

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zhang, Zeliang, Feng, Mingqian, Jiang, Jinyang, Zhu, Rongyi, Peng, Yijie, Xu, Chenliang
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929405306077184
author Zhang, Zeliang
Feng, Mingqian
Jiang, Jinyang
Zhu, Rongyi
Peng, Yijie
Xu, Chenliang
author_facet Zhang, Zeliang
Feng, Mingqian
Jiang, Jinyang
Zhu, Rongyi
Peng, Yijie
Xu, Chenliang
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2403_15603
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Forward Learning for Gradient-based Black-box Saliency Map Generation
Zhang, Zeliang
Feng, Mingqian
Jiang, Jinyang
Zhu, Rongyi
Peng, Yijie
Xu, Chenliang
Computer Vision and Pattern Recognition
Artificial Intelligence
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.
title Forward Learning for Gradient-based Black-box Saliency Map Generation
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2403.15603