Abstract:
PLAGIARISM DETECTION USING DEEP LEARNING, is a detection System for
detecting plagiarism using deep learning techniques. Plagiarism is the act of using or
presenting someone else’s work, ideas, or words as your own without giving proper credit
or acknowledgment to the source. It is considered a serious academic and ethical offense
in most settings, including academia, journalism, and creative industries. The procedure of
human-based plagiarism detection is time-consuming, inaccurate, and difficult. The proposed
system for plagiarism detection using deep learning aims to detect instances of plagiarism
and identify paraphrased content.This Project proposes a plagiarism detection System based
on two deep learning models: A combination of Long Short-Term Memory (LSTM) and
Convolutional Neural Network (CNN), and a Transformer-based model. This project compares
the effectiveness of a transformer-based model, and the combination of the CNN-LSTM model
and also identifies various types of plagiarism including paraphrasing. In terms of outcomes,
research has demonstrated that deep learning models can accurately identify plagiarism. One
model is the combination of the CNN-LSTM model to detect plagiarism and had a 99.51%
success rate. Another model such as a transformer-based model detects plagiarism with a high
accuracy rate such as 99.61%
