Saved in:
Bibliographic Details
Main Authors: Pancerz, Krzysztof, Kulicki, Piotr, Kalisz, Michał, Burda, Andrzej, Stanisławski, Maciej, Sarzyński, Jaromir
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2504.13150
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909583318974464
author Pancerz, Krzysztof
Kulicki, Piotr
Kalisz, Michał
Burda, Andrzej
Stanisławski, Maciej
Sarzyński, Jaromir
author_facet Pancerz, Krzysztof
Kulicki, Piotr
Kalisz, Michał
Burda, Andrzej
Stanisławski, Maciej
Sarzyński, Jaromir
contents Creating responsible artificial intelligence (AI) systems is an important issue in contemporary research and development of works on AI. One of the characteristics of responsible AI systems is their explainability. In the paper, we are interested in explainable deep learning (XDL) systems. On the basis of the creation of digital twins of physical objects, we introduce the idea of creating readable twins (in the form of imprecise information flow models) for unreadable deep learning models. The complete procedure for switching from the deep learning model (DLM) to the imprecise information flow model (IIFM) is presented. 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.
format Preprint
id arxiv_https___arxiv_org_abs_2504_13150
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Readable Twins of Unreadable Models
Pancerz, Krzysztof
Kulicki, Piotr
Kalisz, Michał
Burda, Andrzej
Stanisławski, Maciej
Sarzyński, Jaromir
Artificial Intelligence
Computer Vision and Pattern Recognition
Creating responsible artificial intelligence (AI) systems is an important issue in contemporary research and development of works on AI. One of the characteristics of responsible AI systems is their explainability. In the paper, we are interested in explainable deep learning (XDL) systems. On the basis of the creation of digital twins of physical objects, we introduce the idea of creating readable twins (in the form of imprecise information flow models) for unreadable deep learning models. The complete procedure for switching from the deep learning model (DLM) to the imprecise information flow model (IIFM) is presented. 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.
title Readable Twins of Unreadable Models
topic Artificial Intelligence
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2504.13150