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

arXiv:2403.15603v1 Announce Type: new
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. …

abstract apis arxiv become box chatgpt computing cs.ai cs.cv decisions deep neural network framework gradient however map maps network neural network novel type work