Events, Anaphora, and Computer Assisted Language Learning

Date

• 15:00, Friday, January 18^th, 2013
• Room 336

Identificação e relacionamento de eventos

Speaker

• Viviana Cabrita

Abstract

As operações de análise, extração e processamento de texto destacam-se dentro do processamento de língua natural e são essenciais para o desenvolvimento e aperfeiçoamento de sistemas capazes de, entre outros, realizar sumarizações e traduções de texto sem a intervenção humana. O trabalho descrito neste documento contribui para a expansão dessa área, focando-se na identificação e relacionamento de eventos, com o objetivo permitir a ordenação temporal dos mesmos. Pretendeu-se construir um módulo, a ser integrado numa cadeia de processamento, STRING, com foco na identificação de eventos e na deteção de relações entres eles e em relação com a linha do tempo.

Syntactic Exercises

Speaker

• Pedro Figuerinha

Abstract

The introduction of computers as an assisting tool in the field of education has affected the way teachers plan their classes and in- teract with their students. The Intelligent Computer Assisted Language Learning (ICALL) field addresses this evolution in the teaching process and has fomented great interest for the last few decades in the academic community. REAP.PT is a Portuguese system from this research area and aims to teach the Portuguese language to both foreign and native speakers. This is done by employing a series of automatically generated exercises focusing on several semantic and syntactic aspects of the lan- guage, specifically tailored to the users needs. The goal of this work is to expand the REAP.PT system, namely, its syntactic exercises. A state of the art review was elaborated and it focuses on similar CALL sys- tems already available. The proposed approach is also formalized and the challenges are described. Finally, the evaluation method for assert- ing the quality of these new exercises is specified.

Anaphora Resolution

Speaker

• João Silvestre Marques

Abstract

Nowadays, due to the large amount of information available, there is a growing need of extracting information from natural language texts and processing it. Anaphora resolution is one of the most relevant NLP tasks necessary to answer such needs and has been under study for years. This research describes some systems that have tried to resolve anaphora and propose solutions to be adopted in a Portuguese NLP system.