The number of research topics covered in recent approaches to Information Extraction
(IE) is continually growing as new facts are being considered. In fact,
while the user’s interest in extracting information from texts deals mainly with
the success of the entire process of locating, in document collections, facts of
interest, the process itself is dependent on several constraints (e.g. the domain,
the collection dimension and location, and the document type) and currently it
tackles composite scenarios, including free texts, semi- and structured texts such
as Web pages, e-mails, etc.
The handling of all these factors is tightly related to the continued evolution
of the underlying technologies.
In the last few years, in real-world applications we have seen the need for
scalable, adaptable IE systems (see M.T. Pazienza, “Information Extraction: Towards
Scalable Adaptable Systems”, LNAI 1714) to limit the need for human
intervention in the customization process and portability of the IE application to
new domains. Scalability and adaptability requirements are still valid impacting
features and get more relevance into a Web scenario, wherein intelligent information
agents are expected to automatically gather information from heterogeneous
sources.