Extend AQUA to support addition of new Languages

AQUA has been designed so as to be able to support addition of new languages. This process consists of the following main steps:

• User Interface Localization. This is in practice the translation of a text messages file.
• Spider Model Training. The training of Spider’s classifiers is facilitated using a specialized tool called Corpus Formation Tool for the collection and annotation of corpus.
• IET Model Training. The training of extraction models is supported using the BOEMIE Annotation Tool, which enables the annotation of named entities and relations.
• Topic Categorizer Configuration. AQUA’s topic categorizer employs the Automatic Ontological Concepts Extraction Tool (POKA), a general purpose tool for automatic extraction of ontological concepts. In the current implementation of AQUA MeSH is supported.

Extend AQUA to support addition of new Labeling Criteria

AQUA has been designed so as to be able to support addition of new labeling criteria. To this end, a well-defined process is used, which involves the following key steps:

• Adaptation of the LAM User Interface,
• Definition of the desired criterion as an RDF element of the supported Vocabulary,
• Specification of the methodology for the extraction of information relevant to the new criterion, e.g. use of regular expressions, heuristics or machine learning.

Extend AQUA to support addition of new Labeling Authorities

AQUA has been designed so as to be able to support addition of new labeling authorities. This is an iterative process that extends AQUA for the possible new labeling criteria and language(s) of the new labeling authority.