Personalised News Summarisation Using Demographic-aware BART Model

The exponential growth of digital media has led to a saturation of online news content, making it increasingly important to deliver information that resonates with readers’ specific backgrounds and interests. While traditional summarisation methods treat all audiences homogeneously, recent advances...

Full description

Bibliographic Details
Main Authors: McFadyen, Daniel M., Mahmud, Redowan, Afrin, Mahbuba, Mistry, Sajib, Krishna, Aneesh
Format: Conference Paper
Published: 2025
Online Access:http://hdl.handle.net/20.500.11937/97945
_version_ 1848766341859770368
author McFadyen, Daniel M.
Mahmud, Redowan
Afrin, Mahbuba
Mistry, Sajib
Krishna, Aneesh
author_facet McFadyen, Daniel M.
Mahmud, Redowan
Afrin, Mahbuba
Mistry, Sajib
Krishna, Aneesh
author_sort McFadyen, Daniel M.
building Curtin Institutional Repository
collection Online Access
description The exponential growth of digital media has led to a saturation of online news content, making it increasingly important to deliver information that resonates with readers’ specific backgrounds and interests. While traditional summarisation methods treat all audiences homogeneously, recent advances highlight the need for demographic-aware approaches. However, existing automated news summarising systems rarely incorporate these nuances, often defaulting to one-size-fits-all summaries. Therefore, this research introduces a novel frameworkforpersonalised newssummarisation that accounts for reader-specific factors such as age, geography, and cultural background. We developed a specialised dataset linking demographic attributes to narrative priorities and used this to finetune an open-source BART model for dynamic, audience-aligned information emphasis. Integrated into a user-friendly web application, this model enables the generation of tailored summaries that better meet individual preferences. Comparative evaluations reveal that our approach substantially surpasses traditional summarisation tools, including leading models like ChatGPT 3.5 and Gemini, by providing more contextually relevant content.
first_indexed 2025-11-14T11:49:37Z
format Conference Paper
id curtin-20.500.11937-97945
institution Curtin University Malaysia
institution_category Local University
last_indexed 2025-11-14T11:49:37Z
publishDate 2025
recordtype eprints
repository_type Digital Repository
spelling curtin-20.500.11937-979452025-07-10T08:35:11Z Personalised News Summarisation Using Demographic-aware BART Model McFadyen, Daniel M. Mahmud, Redowan Afrin, Mahbuba Mistry, Sajib Krishna, Aneesh The exponential growth of digital media has led to a saturation of online news content, making it increasingly important to deliver information that resonates with readers’ specific backgrounds and interests. While traditional summarisation methods treat all audiences homogeneously, recent advances highlight the need for demographic-aware approaches. However, existing automated news summarising systems rarely incorporate these nuances, often defaulting to one-size-fits-all summaries. Therefore, this research introduces a novel frameworkforpersonalised newssummarisation that accounts for reader-specific factors such as age, geography, and cultural background. We developed a specialised dataset linking demographic attributes to narrative priorities and used this to finetune an open-source BART model for dynamic, audience-aligned information emphasis. Integrated into a user-friendly web application, this model enables the generation of tailored summaries that better meet individual preferences. Comparative evaluations reveal that our approach substantially surpasses traditional summarisation tools, including leading models like ChatGPT 3.5 and Gemini, by providing more contextually relevant content. 2025 Conference Paper http://hdl.handle.net/20.500.11937/97945 http://creativecommons.org/licenses/by/4.0/ fulltext
spellingShingle McFadyen, Daniel M.
Mahmud, Redowan
Afrin, Mahbuba
Mistry, Sajib
Krishna, Aneesh
Personalised News Summarisation Using Demographic-aware BART Model
title Personalised News Summarisation Using Demographic-aware BART Model
title_full Personalised News Summarisation Using Demographic-aware BART Model
title_fullStr Personalised News Summarisation Using Demographic-aware BART Model
title_full_unstemmed Personalised News Summarisation Using Demographic-aware BART Model
title_short Personalised News Summarisation Using Demographic-aware BART Model
title_sort personalised news summarisation using demographic-aware bart model
url http://hdl.handle.net/20.500.11937/97945