Abstract :

Distributed Machine Learning (DML) Is A Foundational Technology In The Field Of Artificial Intelligence (AI). However, Current DML Frameworks Often Overlook The Importance Of Data Integrity. When Network Attackers Tamper With, Forge, Or Corrupt Training Data, The Performance And Reliability Of The Learning Model Can Be Severely Compromised, Leading To Inaccurate Results. To Address This Challenge, We Propose A Data Integrity Verification Scheme For Distributed Machine Learning (DML-DIV) To Ensure The Trustworthiness Of Training Data.First, We Incorporate The Provable Data Possession (PDP) Sampling-based Auditing Technique To Detect And Defend Against Data Forgery And Tampering. Second, To Protect Data Privacy During The Third Party Auditor (TPA) Verification Process, We Introduce A Randomly Generated Blinding Factor And Leverage The Hardness Of The Discrete Logarithm Problem (DLP) To Construct Secure Proofs. Third, Our Scheme Utilizes Identity-based Cryptography Along With A Two-step Key Generation Mechanism To Eliminate The Key Escrow Issue And Reduce Certificate Management Overhead.Theoretical Security Analysis And Experimental Evaluations Demonstrate That The Proposed DML-DIV Scheme Is Both Secure And Efficient. Index Terms:- Distributed Machine Learning, Data Integrity, Provable Data Possession, Identity-Based Cryptography, Discrete Logarithm Problem, Privacy Protection.

Published:

29-10-2025

Issue:

Vol. 25 No. 10 (2025)

Page Nos:

407-416

Section:

Articles

License:

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

How to Cite