Pazienza M.T. (ed.) Information Extraction. Towards Scalable, Adaptable Systems

Springer, 1999, -174 p.The ever-growing interest in new approaches to information management is strictly related to the explosion of collections of documents made accessible through communication networks. The enormous amount of daily available in- formation imposes the development of IE (Information Extraction) technologies that enable one to access relevant documents only and integrate the extracted information into the user's environment.
In fact, the classic application scenario for IE foresees, for example:
a company interested in getting detailed synthetic information (related to prede_ned categories);
the documents, as sources of information, located in electronically acces- sible sites (agencies' news, web pages, companies' textual documentation, international regulations etc.);
the extracted information eventually being inserted in private data bases for further processing (e.g. data mining, summary and report generation, forms filling,.).
A key problem for a wider deployment of IE systems is in their flexibility and easy adaptation to new application frameworks. Most of the commonly available IE systems are based on specific domain-dependent methodologies for knowledge extraction (they ignore how to pass to templates related to other do- mains or different collections of documents). The need exists for more principled techniques for managing templates in a domain-independent fashion by using the general structures of language and logic. A few attempts have been made to derive templates directly from corpora. This process is similar to deriving knowledge structures and lexicons directly from corpora. This methodological approach (adaptability) could push for a rapid customization to new domains of existing IE systems.Can We Make Information Extraction More Adaptive?
Natural Language Processing and Digital Libraries
Natural Language Processing and Information Retrieval
From Speech to Knowledge
Relating Templates to Language and Logic
Inferential Information Extraction
Knowledge Extraction from Bilingual Corpora
Engineering of IE Systems: An Object-Oriented Approach