Multilingual sentence-level bias detection in Wikipedia

Desislava Aleksandrova (desislava (point) aleksandrova <at> umontreal (point) ca)

OLST, D├ępartement de linguistique et de traduction

Le 31 juillet 2019 à 11 h 30

Salle C-9019, Pavillon Lionel-Groulx

We propose a multilingual method for the extraction of biased sentences from Wikipedia, and use it to create corpora in Bulgarian, French and English. Sifting through the revision history of the articles that at some point have been tagged as biased (by a Wikipedia editor) and later untagged, we retrieve the last tagged and the first untagged revisions as the before/after snapshots of what was deemed a violation of Wikipedia's neutral point of view policy. We extract the sentences that were removed or rewritten between the two revisions. The approach yields sufficient data even in the case of relatively small Wikipedias, such as the Bulgarian one, where 62k articles boiled down to close to 5k biased sentences. We evaluate our method by manually annotating 520 randomly sampled sentences for Bulgarian and French, and 744 for English. We assess the level of noise and analyze its sources. Finally, we exploit the data with well-known classification methods to detect biased sentences.

Pour recevoir les annonces hebdomadaires par courriel, envoyez un message à l'adresse

Liste des autres séminaires