Eckhard Bick

Floresta Sintá(c)tica: Bigger, Thicker and Easier

In this paper, we describe the resumption of activities of Floresta Sinta(c)tica, a treebank for Portuguese. We present some underlying guidelines around the project and how they influence our linguistic choices. We then describe the new texts added to the treebank, proceed to mention the new syntactic information added to the old texts, and finally describe the new user-friendly search system and the plans for its expansion.

Multi-level NER for Portuguese in a CG framework

This paper describes and evaluates a linguistically based NER system for Portuguese, based on lexico-semantical information, pattern matching and morphosyntactic, context driven Constraint Grammar rules. Preliminary Fscores for cross-domain news texts, when distinguishing six different name types, were 91.85 (raw) and 93.6 (subtyping of ready-chunked proper nouns).

Functional aspects in portuguese NER

This paper describes and evaluates an improved and modified version of the PALAVRAS-NER parser, adapted for the HAREM joint evaluation task of Named Entity Recognition (NER) in February 2005. Apart from an extension to over 40 semantic categories, the system was changed from a lexeme-based to a token-based description, defining NER categories as functional and context-based rather than stable and lexematic. The Constraint Grammar rule body was changed accordingly, adding new rules and applying existing heuristic and disambiguation rules to contextual re-mapping of also lexically known material. In the joint evaluation, PALAVRAS-NER achieved the best overall results, achieving top ranks for both the identification, classification and morphology tasks.

Semantic tagging for resolution of indirect anaphora

This paper presents an evaluation of indirect anaphor resolution which considers as lexical resource the semantic tagging provided by the PALAVRAS parser. We describe the semantic tagging process and a corpus experiment.

A Constraint Grammar Based Question Answering System for Portuguese

This paper describes a linguistic, NLP based Information Extraction (IE) system for Portuguese text. The system focuses on the natural language Question Answering task (QA) and relies on full syntactic tree parses, exploiting both lexico-semantic information, syntactic pattern matching and named entity subtyping. Context driven Constraint Grammar rules (CG) are used not only for PoS disambiguation, but also to assign word based function tags and semantic categories. Exact QA matches are supplemented by iterative textual and syntactic relaxation of seach phrases and hit sentence chunks.

Grammatical Annotation of Historical Portuguese: Generating a Corpus-Based Diachronic Dictionary

In this paper, we present an automatic system for the morphosyntactic annotation and lexicographical evaluation of historical Portuguese corpora. Using rule-based orthographical normalization, we were able to apply a standard parser (PALAVRAS) to historical data (Colonia corpus) and to achieve accurate annotation for both POS and syntax. By aligning original and standardized word forms, our method allows to create tailor-made standardization dictionaries for historical Portuguese with optional period or author frequencies.

Building interdisciplinary bridges : MUCH : The Malmö University-Chalmers Corpus of Academic Writing as a Process

This paper describes a corpus of writing as a process (MUCH), comprising English as a Foreign Language (EFL) student texts. The corpus will contain a large number of richly annotated papers in several versions from students of different performance levels. It will also include peer and instructor feedback, as well as tools for visualising the revision process, and for analysing the writing process and the peer and instructor feedback. MUCH will make it possible to study how texts develop and change in the course of the writing process and how feedback impacts the process.

Constraint Grammar-Based Swedish-Danish Machine Translation

This paper describes and evaluates a grammar-based machine translation system for the Swedish-Danish language pair. Source-language structural analysis, polysemy resolution, syntactic movement rules and target-language agreement are based on Constraint Grammar morphosyntactic tags and dependency trees. Lexical transfer rules exploit dependency links to access contextual information, such as syntactic argument function, semantic type and quantifiers, or to integrate verbal features, e.g. diathesis and auxiliaries. Out-of-vocabulary words are handled by derivational and compound analysis with a combined coverage of 99.3%, as well as systematic morpho-phonemic transliterations for the remaining cases. The system achieved BLEU scores of 0.65-0.8 depending on references and outperformed both STMT and RBMT competitors by a large margin.

The Identification and Description of Frozen Prepositional Phrases through a Corpus-Oriented Study

This research is a corpus-based analysis of Brazilian Portuguese prepositional phrases that have a frozen status. Based on a previous list, elaborated for parser PALAVRAS (Bick, 2000), we examine the PPs frozen syntactic behavior and propose three different PPs syntactic-semantic sets, which are true assets for lexicographic purpose and NLP lexical resource.

Um mundo novo na Floresta Sintá(c)tica – o treebank do Português

A Floresta Sinta(c)tica tem como objetivo criar e disponibilizar um corpus sintaticamente anotado. Neste artigo, sao apresentados dois novos materiais do projeto: Selva (300 mil palavras e parcialmente revisto) e Amazonia (3.8 milhoes de palavras, nao revisto). Para lidar com um material tao grande e variado foi construida a interface Milhafre. O artigo mostra, ainda, como vem sendo enfrentado o desafio de compatibilizar, de uma lado, o usuario linguista, que pode ter um perfil muito heterogeneo e, em geral, pouca familiaridade determinadas formalizacoes mais utilizadas em informatica e, de outro, um unico modelo de anotacao sintatica, frequentemente pouco conhecido do lado “linguistico nao-computacional” e uma interface de acesso e manipulacao de corpora capaz de lidar com um objeto tao complexo como a lingua. Palavras-chave: arvores sintaticas, corpus anotado, corpus revisto, busca em corpora.