Abstract (may include machine translation)
Cet article présente un essai d’application de l’analyse discursive (text zoning) à l’ACL Anthology. Il s’agit ainsi de mieux caractériser le contenu des articles du domaine de la linguistique informatique afin de pouvoir en faire une analyse fine par la suite. Nous montrons que des techniques récentes d’analyse discursive fondées sur l’apprentissage faiblement supervisé permettent d’obtenir de bons résultats.
This paper presents an application of Text Zoning to the ACL Anthology. Text Zoning is known to be useful to characterize the content of papers, especially in the scientific domain. We show that recent techniques based on weakly supervised learning obtain excellent results on the ACL Anthology. Although this kind of technique is known in the domain, it is the first time it is applied to the whole ACL Anthology.
| Translated title of the contribution | Argumentative analysis of the ACL Anthology |
|---|---|
| Original language | French |
| Pages | 580–585 |
| State | Published - 2014 |
| Externally published | Yes |
| Event | 21eme Traitement Automatique des Langues Naturelles, TALN 2014 - 21st Automatic Natural Language Processing, TALN 2014 - Marseille, France Duration: 1 Jul 2014 → 4 Jul 2014 |
Conference
| Conference | 21eme Traitement Automatique des Langues Naturelles, TALN 2014 - 21st Automatic Natural Language Processing, TALN 2014 |
|---|---|
| Country/Territory | France |
| City | Marseille |
| Period | 1/07/14 → 4/07/14 |
Keywords
- ACL Anthology
- Corpus of scientific texts
- Text zoning