Ivandré Paraboni

Learn More
It is often desirable that referring expressions be chosen in such a way that their referents are easy to identify. This article focuses on referring expressions in hierarchically structured domains, exploring the hypothesis that referring expressions can be improved by including logically redundant information in them if this leads to a significant(More)
Statistical approaches have recently emerged as the main paradigm in Machine Translation (MT) research. In previous work we have shown that results of a simple statistical word-based MT system may be highly comparable to those produced by a rule-based approach for closely-related languages such as Brazilian Portuguese and European Spanish. In this work we(More)
Anaphora resolution is an essential component of most NLP applications, from text understanding to Machine Translation. In this work we discuss a supervised machine learning approach to the problem, focusing on instances of anaphora ubiquitously found in a corpus of Brazilian Portuguese texts, namely, third-person pronominal references. Although still(More)
Both greedy and domain-oriented REG algorithms have significant strengths but tend to perform poorly according to humanlikeness criteria as measured by, e.g., Dice scores. In this work we describe an attempt to combine both perspectives into a single attribute selection strategy to be used as part of the Dale & Reiter Incremental algorithm in the REG(More)
In this paper we describe issues for the computational generation of references to document parts such as sections, pictures, paragraphs etc, which we call document deixis. We discuss the circumstances under which these references need to be used in place of ordinary referring expressions and a number of problems in their generation. We also sketch a(More)
This paper will focus on how a natural language generation system can produce references to parts of documents such as sections, pictures paragraphs etc., a phenomenon we call document deixis. An example is the definite description “picture 3 in section B”. We propose an algorithm to be implemented as an extension of an existing NLG system, which is able to(More)
Natural Language Generation systems usually require substantial knowledge about the structure of the target language in order to perform the final task in the generation process – the mapping from semantic representation to text known as surface realisation. Designing knowledge bases of this kind, typically represented as sets of grammar rules, may however(More)