paperbot · PL 论文追踪

全部论文 待读 收藏

P2 Repairing confusion and bias errors for DNN-based image classifiers

修复基于 DNN 的图像分类器的混淆与偏差错误
FSE Proceedings of the 28th ACM Joint Meeting on European Software Engineering Conference and Symposium on the Foundations of Software Engineering2020匹配度 2/10 · 该论文关注 DNN 图像分类器的混淆与偏差错误修复，属于机器学习测试与修复方向，与用户核心关注的 PL→AI 静态分析、AI→PL 形式化验证、浮点/贝叶斯程序分析及 LLM/Agent 交叉均无直接关联。虽带有“机器学习与程序语言”标签，但方法不涉及程序分析或 LLM/Agent 技术。引用 3 inbox
Yuchi Tian
测试、调试与修复机器学习与程序语言
crossref: 2026-09-12

AI 速览

一句话：提出与混淆错误高度相关的可微距离度量，并在重训练中拉大类间距离，以修复 DNN 图像分类器的混淆和偏差错误。

基于 DNN 的图像分类器容易产生混淆与偏差错误：即使经过鲁棒训练，模型也可能在特定类别对之间高度混淆，或对某些类别存在偏向。本文提出一种与混淆错误高度相关的可微距离度量，并据此设计修复方法：在重训练过程中增大两个类别之间的距离，以降低混淆错误。作者在单标签和多标签分类模型及数据集上进行了评估。结果表明，该方法能有效减少混淆错误，同时仅带来轻微的准确率下降。

问题
DNN 图像分类器即使鲁棒训练后仍会在特定类别对之间发生混淆，或对某些类别产生偏差。
方法
提出与混淆错误高度相关的可微距离度量，并通过重训练时增大类别间距离来修复混淆错误。
评估
在单标签和多标签分类模型及数据集上评估，结果显示混淆错误显著减少且准确率仅轻微下降。
原文摘要（Abstract）

Recent works in DNN testing show that DNN based image classifiers are susceptible to confusion and bias errors. A DNN model, even robust trained model can be highly confused between certain pair of objects or highly bias towards some object than others. In this paper, we propose a differentiable distance metric, which is highly correlated with confusion errors. We propose a repairing approach by increasing the distance between two classes during retraining the model to reduce the confusion errors. We evaluate our approaches on both single-label and multi-label classification models and datasets. Our results show that our approach effectively reduce confusion errors with very slight accuracy reduce.

链接与引用

DOI 原文 · PDF（开放获取） ·

BibTeX 
@inproceedings{paperbot5726,
  title = {Repairing confusion and bias errors for DNN-based image classifiers},
  author = {Yuchi Tian},
  booktitle = {FSE},
  year = {2020},
  doi = {10.1145/3368089.3418776}
}