让深度学习的显著区域分析具备统计可信度
si4onnx: A Python package for Selective Inference in Deep Learning Models
- 通过选择性推断实现对模型显著区域的假设检验
- 可控制第一类错误率，提升AI解释结果的可靠性
- 兼容PyTorch/TensorFlow，适合可解释AI研究者使用
本文介绍si4onnx，一个用于深度学习模型选择性推断的Python工具包。像XAI中的CAM或基于变分自编码器（VAE）的重建异常检测等技术，常被用来识别输入图像中的显著区域。然而，这些识别出的区域未必具有实际意义。因此，评估这些区域的统计显著性，成为提升AI系统可靠性的重要挑战。si4onnx提供了一种简便方法，可在保持第一类错误率受控的前提下进行假设检验，支持主流框架如PyTorch和TensorFlow构建的深度学习模型。
原文摘要 · Abstract (English)
In this paper, we introduce si4onnx, a package for performing selective inference on deep learning models. Techniques such as CAM in XAI and reconstruction-based anomaly detection using VAE can be interpreted as methods for identifying significant regions within input images. However, the identified regions may not always carry meaningful significance. Therefore, evaluating the statistical significance of these regions represents a crucial challenge in establishing the reliability of AI systems. si4onnx is a Python package that enables straightforward implementation of hypothesis testing with controlled type I error rates through selective inference. It is compatible with deep learning models constructed using common frameworks such as PyTorch and TensorFlow.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。