A study on the efficacy of natural language generation techniques for similar writing personalities

This thesis is submitted in partial fulfillment of the requirements for the degree of Bachelor of Science in Computer Science, 2024.

Détails bibliographiques
Auteurs principaux: Anwar, Ahmed, Tanzir, Hasan Mohammod, Hosain, Abdul Halim, Islam, Taslima, Raya, Nusrat Zaman
Autres auteurs: Sadeque, Farig Yousuf
Format: Thèse
Langue:English
Publié: Brac University 2024
Sujets:
Accès en ligne:http://hdl.handle.net/10361/22884
id 10361-22884
record_format dspace
spelling 10361-228842024-05-20T21:02:28Z A study on the efficacy of natural language generation techniques for similar writing personalities Anwar, Ahmed Tanzir, Hasan Mohammod Hosain, Abdul Halim Islam, Taslima Raya, Nusrat Zaman Sadeque, Farig Yousuf Department of Computer Science and Engineering, Brac University Natural language processing LSTM BanglaBERT BanglaT5 Writing personality T5 model Adversarial attacks Natural language processing (Computer science) This thesis is submitted in partial fulfillment of the requirements for the degree of Bachelor of Science in Computer Science, 2024. Cataloged from PDF version of thesis. Includes bibliographical references (pages 100-101). Over the past two decades, the domain of Natural Language Processing has undergone a remarkable transformation enabling machines to generate text, summarize content, paraphrase and analyze sentiment. The captivating idea of analyzing and copying someone’s writing style is no longer impossible. Pushing boundaries further, we have embarked on a journey to implement such a model for the Bengali language by utilizing the approach of style transfer through the application of deep learning using LLM. One’s writing personality can be identified by training the model by imputing a set of documents (notes, books, etc) authored by the writers only. The system will be able to extract important information to recreate sentences with similar structural properties used by the author. Additionally, it will also be able to detect whether a particular sentence structure synchronizes with that author’s distinctive style. The outcome of our model aims to fulfill the need of a particular writing taste of an author, as requested by the user. In essence, our model blends technology and art to write in a way that is reminiscent of their favorite Bengali author. Our proposed model not only skillfully excels in authorship classification and mimicking their style, but also stands resilient against potential adversarial attacks, making it a strong and unyielding system that aligns well with our research objective. Ahmed Anwar Hasan Mohammod Tanzir Abdul Halim Hosain Taslima Islam Nusrat Zaman Raya B.Sc in Computer Science 2024-05-20T06:11:33Z 2024-05-20T06:11:33Z ©2024 2024-01 Thesis ID: 20301077 ID: 20101598 ID: 20101300 ID: 20101603 ID: 23241034 http://hdl.handle.net/10361/22884 en Brac University theses are protected by copyright. They may be viewed from this source for any purpose, but reproduction or distribution in any format is prohibited without written permission. 108 pages application/pdf Brac University
institution Brac University
collection Institutional Repository
language English
topic Natural language processing
LSTM
BanglaBERT
BanglaT5
Writing personality
T5 model
Adversarial attacks
Natural language processing (Computer science)
spellingShingle Natural language processing
LSTM
BanglaBERT
BanglaT5
Writing personality
T5 model
Adversarial attacks
Natural language processing (Computer science)
Anwar, Ahmed
Tanzir, Hasan Mohammod
Hosain, Abdul Halim
Islam, Taslima
Raya, Nusrat Zaman
A study on the efficacy of natural language generation techniques for similar writing personalities
description This thesis is submitted in partial fulfillment of the requirements for the degree of Bachelor of Science in Computer Science, 2024.
author2 Sadeque, Farig Yousuf
author_facet Sadeque, Farig Yousuf
Anwar, Ahmed
Tanzir, Hasan Mohammod
Hosain, Abdul Halim
Islam, Taslima
Raya, Nusrat Zaman
format Thesis
author Anwar, Ahmed
Tanzir, Hasan Mohammod
Hosain, Abdul Halim
Islam, Taslima
Raya, Nusrat Zaman
author_sort Anwar, Ahmed
title A study on the efficacy of natural language generation techniques for similar writing personalities
title_short A study on the efficacy of natural language generation techniques for similar writing personalities
title_full A study on the efficacy of natural language generation techniques for similar writing personalities
title_fullStr A study on the efficacy of natural language generation techniques for similar writing personalities
title_full_unstemmed A study on the efficacy of natural language generation techniques for similar writing personalities
title_sort study on the efficacy of natural language generation techniques for similar writing personalities
publisher Brac University
publishDate 2024
url http://hdl.handle.net/10361/22884
work_keys_str_mv AT anwarahmed astudyontheefficacyofnaturallanguagegenerationtechniquesforsimilarwritingpersonalities
AT tanzirhasanmohammod astudyontheefficacyofnaturallanguagegenerationtechniquesforsimilarwritingpersonalities
AT hosainabdulhalim astudyontheefficacyofnaturallanguagegenerationtechniquesforsimilarwritingpersonalities
AT islamtaslima astudyontheefficacyofnaturallanguagegenerationtechniquesforsimilarwritingpersonalities
AT rayanusratzaman astudyontheefficacyofnaturallanguagegenerationtechniquesforsimilarwritingpersonalities
AT anwarahmed studyontheefficacyofnaturallanguagegenerationtechniquesforsimilarwritingpersonalities
AT tanzirhasanmohammod studyontheefficacyofnaturallanguagegenerationtechniquesforsimilarwritingpersonalities
AT hosainabdulhalim studyontheefficacyofnaturallanguagegenerationtechniquesforsimilarwritingpersonalities
AT islamtaslima studyontheefficacyofnaturallanguagegenerationtechniquesforsimilarwritingpersonalities
AT rayanusratzaman studyontheefficacyofnaturallanguagegenerationtechniquesforsimilarwritingpersonalities
_version_ 1814308413681696768