Advances in Information Extraction of Elusive Information from Text
Sponsored by AFCEA. Recent research that exploits the full range of Natural Language Processing capabilities is producing results which indicate that Information Extraction can go far beyond the earlier focus on entities, per se. More subtle aspects of textual information, conveyed either explicitly or implicitly, can be recognized, categorized and included in more informative summaries of single documents, as well as collections of multiple documents on a topic or interest. Extractions include events, as well as relations between entities and events, and a range of relations amongst entities. More sophisticated Information Extraction includes temporal, certainty, and affective aspects of text – all of which can now begin to be considered. This presentation will focus mainly on recent accomplishments on these newer contributions to information exploitation, both in the field in general and in our own work on pushing the boundaries towards more human like interpretation and extraction of the important aspects of textual reports.
