<?xml version="1.0" encoding="UTF-8"?>
<article article-type="Research Article">
  <front>
    <journal-meta>
      <journal-id journal-id-type="publisher">journal-of-interdisciplinary-mathematics</journal-id>
      <journal-title-group>
        <journal-title>Journal of Interdisciplinary Mathematics</journal-title>
      </journal-title-group>
      <issn publication-format="electronic">2169-012X</issn>
      <issn publication-format="print">0972-0502</issn>
      <publisher>
        <publisher-name>Taru Publications</publisher-name>
      </publisher>
    </journal-meta>
    <article-meta>
      <article-id pub-id-type="doi">10.47974/JIM-2504</article-id>
      <title-group>
        <article-title>Fractional calculus applications in deep learning architectures</article-title>
      </title-group>
      <contrib-group>
        <contrib contrib-type="author">
          <name>
            <surname>Kokane</surname>
            <given-names>Chandrakant Deelip</given-names>
          </name>
          <aff>Department of Computer Science and Engineering (Artificial Intelligence), Vishwakarma Institute of Technology, Pune, Maharashtra, 411037, India</aff>
        </contrib>
        <contrib contrib-type="author">
          <name>
            <surname>More</surname>
            <given-names>Anjali S.</given-names>
          </name>
          <aff>Department of Artificial Intelligence and Machine Learning, Symbiosis International (Deemed University), Symbiosis Institute of Technology, Pune, Maharashtra, 412115, India</aff>
        </contrib>
        <contrib contrib-type="author" corresp="yes">
          <name>
            <surname>Sarode</surname>
            <given-names>Harsha Jitendra</given-names>
          </name>
          <aff>Department of Electronics and Telecommunications Engineering, Nutan Maharashtra Institute of Engineering and Technology, Pune, Maharashtra, 410507, India</aff>
        </contrib>
        <contrib contrib-type="author">
          <name>
            <surname>Mokashi</surname>
            <given-names>Mandar K.</given-names>
          </name>
          <aff>Department of Computer Science and Engineering, School of Computing, (Artificial Intelligence and Analytics), MIT-ADT University, Pune, Maharashtra, 412201, India</aff>
        </contrib>
        <contrib contrib-type="author">
          <name>
            <surname>Bhoite</surname>
            <given-names>Sonali Prashant</given-names>
          </name>
          <aff>Department of Information Technology, Vishwakarma Institute of Technology, Pune, Maharashtra, 411037, India</aff>
        </contrib>
        <contrib contrib-type="author">
          <name>
            <surname>Bhandari</surname>
            <given-names>Mahesh Ashok</given-names>
          </name>
          <aff>Department of Information Technology, Vishwakarma Institute of Technology, Pune, Maharashtra, 411037, India</aff>
        </contrib>
      </contrib-group>
      <volume>29</volume>
      <issue>3</issue>
      <fpage>693</fpage>
      <lpage>700</lpage>
      <pub-date date-type="pub">
        <day>18</day>
        <month>03</month>
        <year>2026</year>
      </pub-date>
      <abstract>
        <p>The study investigates the application of the concept of the addition of a deep learning system to fractional calculus (FC) in order to accelerate and generalize the learning process. We propose a different form of neural network FGNN, Fractional Gradient Neural Network. In this network, the derivatives of fractional order are the gradient dynamics in control resulting in the optimisation being more fluid and the extraction of features being improved. The approach is a hybrid between the ease of use and memory capabilities of FC and the adaptation and data-driven capabilities of DL. The picture and signaling datasets experimental results indicate that these networks are more precise, more robust and less prone to overfitting as compared to standard networks. These findings indicate that FC is an excellent mathematical foundation to make next-generation deep learning models improved.</p>
      </abstract>
      <kwd-group>
        <kwd>Fractional calculus</kwd>
        <kwd>Deep learning</kwd>
        <kwd>Fractional gradient neural network (FGNN)</kwd>
        <kwd>Optimization</kwd>
        <kwd>Generalization</kwd>
        <kwd>Convergence stability</kwd>
      </kwd-group>
      <custom-meta-group>
        <custom-meta>
          <meta-name>access</meta-name>
          <meta-value>open</meta-value>
        </custom-meta>
        <custom-meta>
          <meta-name>retracted</meta-name>
          <meta-value>no</meta-value>
        </custom-meta>
      </custom-meta-group>
    </article-meta>
  </front>
</article>
