Advanced Capabilities for Evidence Extraction (ACEE)
Abstract
The Center for Natural Language Processing (CNLP) at Syracuse University recently completed the Advanced Capabilities for Evidence Extraction (ACEE) Project, which has improved the effectiveness of its basic entity, relation, and event extraction technology and extended these capabilities in several ways. First, the Information Extraction (IE) technology can now be quickly ported to new domains by use of algorithms utilizing Transformation-Based Learning to specialize generic relation extraction to specific domains. Second, Alias Tracking has enhanced entity coalition within documents by means of more sophisticated co-reference algorithms and has enabled entity tracking across documents. Third, new Linguistic Inferencing capabilities improved cohesiveness of extractions by enabling event coalition. Fourth, significant development in Temporal Sequencing and Scenario Understanding was accomplished based on improved interpretation of temporal attributes of events (e.g., frequency, occurs, since) and temporal relations between events (e.g., before, after, concurrent) providing a richer basis for time-line analysis of events than in previous extraction work. Fifth, an innovative model-based approach to automated certainty detection and categorization was developed and tested, including level of certainty, the person who experiences certainty (e.g., reporter, witness), the abstract or factual nature of the focus of the certainty, and point in time at which the certainty is expressed. Engineering efforts have improved the speed, scalability, and portability of CNLP's IE technology. The appendix contains a hierarchy of categories, or taxonomy, that linguistic analysts use to categorize entities. (17 figures, 45 refs.)
Document Details
- Document Type
- Technical Report
- Publication Date
- Jul 01, 2004
- Accession Number
- ADA425835
Entities
People
- Eileen Allen
- Elizabeth Liddy
- Nancy Mccracken
Organizations
- Syracuse University