ONNXExplainer: an ONNX Based Generic Framework to Explain Neural Networks Using Shapley Values

TL;DR

The paper introduces ONNXExplainer, a cross-platform, efficient framework for explaining neural network decisions using Shapley values, enabling one-shot deployment and significantly reducing explanation latency.

Contribution

It presents a novel automatic differentiation and optimization approach within ONNXExplainer, improving explanation efficiency and enabling cross-platform deployment compared to existing methods.

Findings

Explanation latency improved by up to 500% for several neural networks.

Supports one-shot deployment of inference and explanations.

Outperforms existing SHAP implementation in efficiency and cross-platform compatibility.

Abstract

Understanding why a neural network model makes certain decisions can be as important as the inference performance. Various methods have been proposed to help practitioners explain the prediction of a neural network model, of which Shapley values are most popular. SHAP package is a leading implementation of Shapley values to explain neural networks implemented in TensorFlow or PyTorch but lacks cross-platform support, one-shot deployment and is highly inefficient. To address these problems, we present the ONNXExplainer, which is a generic framework to explain neural networks using Shapley values in the ONNX ecosystem. In ONNXExplainer, we develop its own automatic differentiation and optimization approach, which not only enables One-Shot Deployment of neural networks inference and explanations, but also significantly improves the efficiency to compute explanation with less memory consumption. For fair comparison purposes, we also implement the same optimization in TensorFlow and PyTorch and measure its performance against the current state of the art open-source counterpart, SHAP. Extensive benchmarks demonstrate that the proposed optimization approach improves the explanation latency of VGG19, ResNet50, DenseNet201, and EfficientNetB0 by as much as 500%.