ARTIFACTO

"Analyzing and Recognizing Time, Factuality, and Opinion in Text"

 Partecipanti

Mappa

 Word cloud

Esplora la "nuvola delle parole (Word Cloud) per avere un'idea di massima del progetto.

 Obiettivo del progetto (Objective)

'This project aims at exploring methodologies on porting natural language processing (NLP) technology from one language to another (Objective 1), while contributing specific technological infrastructure to the target language (Objective 2). Research on cross-linguistic porting of technology is increasingly popular due to the advantages of reusing NLP tools of costly production. Moreover, it is a key issue in the EU, given its multilinguality and the challenges this poses regarding resource sharing across linguistic borders. The project focuses on technology for processing time, factuality, and opinion expressed in text. Time information concerns “what happened when” and involves ordering events along the temporal axis, factuality has to do with the degree of certainty of events, and opinion deals with attitudes and their sources. These are basic levels for NLP tasks concerning text understanding, but started to receive some attention only recently. They are expressed through well-delimited grammar and lexicon fragments and share conceptual structures across languages, which makes them ideal test-beds for exploring the cross-language porting of NLP tools. The project will result in a suite of tools for analyzing the 3 levels of information in Catalan and Spanish, to be integrated as part of the information extraction system in the host. Objective 1 will be achieved through experiments exploring the feasibility of different procedures for porting lexical and parsing resources. Objective2 involves research at theoretical (setting description models for the phenomena) and applied levels (developing specific resources, to be done based on the experiments for Objective1). Most previous work on the area has been carried out in the USA -in the group of the researcher and other collaborating centers. By reintegrating her, the EU strengthens its scientific role in the area, guarantees future international collaborations, and allows her to consolidate her scientific profile.'

Introduzione (Teaser)

Language processing gets better than ever

Descrizione progetto (Article)

Effective translation of text from one European language to another, taking into consideration complex nuances and intricacies, is important to enhance communication on many levels across the continent. Natural language processing (NLP) technology is playing an important role in facilitating this ability through high-tech means. The EU-funded project 'Analyzing and recognizing time, factuality, and opinion in text' (ARTIFACTO) furthered this vision and tested it on translation between Catalan and Spanish.

The project worked on developing technology for the processing of three semantic domains: time, event factuality and opinion information in natural language text. These three conceptual systems are ideal for examining the cross-linguistic porting of NLP technology.

ARTIFACTO defined description models for these systems as expressed in Catalan and Spanish that contribute to defining cross-lingual standards. More specifically, the ARTIFACTO system represents software that effectively recognises events, sorts them by time, and identifies the different degrees of factuality and opinions.

These achievements helped the project members enhance information extraction systems and furthered research on related topics. The work done supports the design of adequate cross-lingual descriptive models for NLP applications, linking equivalent corpus resources and expanding NLP technology to other languages.

Such technology will contribute to enhancing communication in Europe among different communities, improving access of information to citizens. In the long run, the EU can become a pioneer in this important field, helping to bridge understanding between cultures and foster knowledge exchange around the globe.