fbpx

Platform created jointly with the Informatics Department received award

A team from the University of Minho will receive this Wednesday the Arquivo.pt 2021 Award, for having created a platform that identifies and studies minorities in online newspapers. Paulo Martins, Leandro Costa and José Carlos Ramalho will be awarded at 5.15pm, at the Lisbon Congress Center, in the presence of the President of the Republic and during the closing of the “Science Meeting 2021“. The prize is worth 10.000 euros and is promoted by the Foundation for Science and Technology and distinguishes for the fourth year projects that value the Arquivo.pt service, which stores millions of pages of the Portuguese web.

 “We are very happy for this recognition, we were not expecting it, there were many good candidates”, smile Paulo Martins and Leandro Costa, connected to the Department of Informatics of the School of Engineering of UMinho. The platform, called “Major Minors”, is now gaining projection and may even help those who study minority groups, under-represented or targets of social stigma, they add. At minors.ilch.uminho.pt one can consult 16,286 articles, with 4406 associated comments, but there are also comparative graphs and details such as the personalities and parties mentioned. The clippings are from the period 2001-2019 and belong to “Público”, the newspaper with the oldest and widest representation in Arquivo.pt, but the idea is to soon add other newspapers and year by year.

Paulo Martins and Leandro Costa began the project in January 2020, as part of two disciplines of the Master’s Degree in Computer Engineering at the School of Engineering of UMinho, supervised by Professor José Carlos Ramalho. They developed an algorithm that automatically identifies articles, comments and photos with reference to eight groups: refugees, women, homosexuals, animals, Africans, Asians, gypsies and migrants. More general gender or racial issues are also targeted. “We know that there are lots of subcategories in minorities and this is a starting point, the initial intention was to support interests of some scientific groups of the Centre for Humanistic Studies of UMinho, which address issues such as identities, multiculturalism, marginalization, computational linguistics and animal ethics,” says Paulo Martins.

The algorithm they created allows “a sea of semantic relations” for each article, such as the number of times certain words appear in it and whether it is in the title, subtitle or body of the news item. This generates levels of relevance in articles about minorities, when in a search engine like Google those could appear in the same order of importance, explains Leandro Costa.

The platform is built in layers and in a modular way, which allows it to adapt to other themes and languages, such as introduced foreignisms, the evolution of language and stigmas about certain groups or the feeling of the negative or positive charge of the news. The authors will also invite the scientific community to use “Major Minors”; on July 1 and 2 they will present the project at the 10th Symposium on Languages, Applications and Technologies (SLATE), in Esposende.

Video here and here.

Full information about the prize and the platform here.