Fusion of Information from Diverse, Textual Media: A Case Restoration Approach

"Fusing information in diverse text media containing case-sensitive information is explored. It is based on a core Information Extraction (IE) system capable of processing case-sensitive text. The core engine is adapted to handle diverse, case-insensitiveinformation e.g. e-mail, chat, newsgroups, broadcast transcripts, HUMINT intelligence documents. The fusion system assimilates information extracted from text with that in structured knowledge bases.Traditional IE for case-insensitive text is limited to the named entity (NE) stage, e.g. retraining an NE tagger on case insensitive text. We explore case restoration, whereby statistical models and rules are used to recover case-sensitive form. Thus,the core IE system is not modified.IE systems are fully exploited if their output is consolidated with knowledge in relational databases. This calls for natural language processing and reasoning, including entity co-reference and event co-reference. Consolidation permits database changedetection and alerts. Feedback to the core IE system exploits information in knowledge bases thereby fusing information.Final deliverables include: (i) a prototype system for case restoration for broadcast transcripts, (ii) benchmarks for NE tagging on case-restored text, and (iii) a blueprint for an information fusion system, including criteria for merging extractedinformation with knowledge in databases. Information analysts and decision makers will ben