Scholarlink

In the paper, we are interested in explainable deep learning (XDL) systems. In general, deep learning models have a sequential structure of layers. Therefore, we propose to use rough set flow graphs to model the information flow between layers. Information consists of artifacts generated by individual layers. Artifacts can be interpreted (if possible) visually or linguistically. The operation of the proposed approach is illustrated with an example of a deep learning classification model for image recognition of handwritten digits from the MNIST data set.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Framework for Explainable Deep Learning Systems Based on Rough Set Flow Graphs

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Krzysztof Pancerz,
  • Piotr Kulicki,
  • Andrzej Burda,
  • Jaromir Sarzyński

摘要

In the paper, we are interested in explainable deep learning (XDL) systems. In general, deep learning models have a sequential structure of layers. Therefore, we propose to use rough set flow graphs to model the information flow between layers. Information consists of artifacts generated by individual layers. Artifacts can be interpreted (if possible) visually or linguistically. The operation of the proposed approach is illustrated with an example of a deep learning classification model for image recognition of handwritten digits from the MNIST data set.