Semantic Analysis and Filtering of Text
Abstract
The report describes the work of the Semantic Analysis and Filtering of Text (SAFT) project at CMU and USC/ISI funded by DARPAs Deep Exploration and Filtering of Text (DEFT). The project investigated challenging questions at the heart of multilingual semantics-oriented natural language understanding of news and similar genres. The project included six major teams/groups, focusing on: (i) semantic frame analysis, (ii) entity detection and linking, (iii) event relation extraction, (iv) event mention detection/coreference, (v) inference/relation discovery, and (vi) bringing everything together in a single multilingual knowledge base supporting English, Chinese, and Spanish. The algorithms were assembled and evaluated every year in the Text Analysis Conference (TAC) KBP and Cold Start KBP evaluations organized by National Institute of Standards and Technology (NIST).
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 2018
- Accession Number
- AD1055310
Entities
People
- Eduard Hovy
Organizations
- Carnegie Mellon University