CROSS-LANGUAGE PLAGIARISM DETECTION SYSTEM

Read full text Summarise with VIRAT
Authors: Lagudu Lahari, Kalla Akhila, Mohammad Mahaboob Yezdani, Koppada Krishna Thirumala Hemanth, Mrs. D. Hima Bindu
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

The exponential growth of multilingual digital content has intensified the challenge of identifying plagiarism across languages and writing styles. Traditional plagiarism detection systems exhibit significant limitations when confronted with cross-lingual cases, particularly translation-based copying between linguistically diverse languages such as Hindi and English. This research presents a comprehensive deep learning framework for multilingual plagiarism detection that operates without dependency on computationally expensive Transformer-based models. The proposed system employs classical deep learning architectures including Convolutional Neural Networks for local pattern extraction, Bidirectional Long Short-Term Memory networks for sequential context modeling, and Siamese neural networks for semantic similarity learning. Cross-lingual detection capability is achieved through MUSE pre-aligned FastText embeddings, which map Hindi and English text into a shared 300-dimensional semantic space with demonstrated alignment accuracy exceeding 85 percent. The framework was trained and evaluated on a synthetically generated dataset comprising 1,584 text pairs encompassing four distinct plagiarism categories across 909 source documents. Experimental results demonstrate that the hybrid CNN-BiLSTM architecture achieves 95.40 percent overall accuracy with particularly strong performance in semantic plagiarism detection at 100 percent accuracy and cross-lingual Hindi-English detection at 87.50 percent accuracy. The system has been deployed as an interactive web application providing real-time detection with detailed explanations, demonstrating practical viability for academic institutions and content verification systems operating in multilingual environments.

Fields of Study & Subjects

Academic integrity and plagiarism Topic Modeling Software Engineering Research

Publication details

Journal / source
Zenodo (CERN European Organization for Nuclear Research)
Publication date
1 January 2026
Language
EN