The MiTAP System for Monitoring Reports of Disease Outbreak

The MiTAP system was developed as an experimental prototype using human language technologies for monitoring infectious disease outbreaks and other global disasters. MiTAP is designed to provide timely multi-lingual information access to analysts, medical experts, health services, and individuals involved in humanitarian assistance and relief work. Every day, thousands of articles from hundreds of global information sources are automatically captured, filtered, translated, tagged, summarized, categorized by content, and made available to users via a news server and web-based search engine. Information extraction technology plays a critical role in many of these processes, presenting information in a variety of time-saving mechanisms to facilitate browsing, searching, sorting, and scanning of articles. Machine translation provides analysts with access to foreign language information otherwise unavailable. We have created a novel prototype by integrating MiTAP with an expert system to help analysts and public health officials deal with overwhelming amounts of data and information in the biomedical domain, specifically relating to disease outbreaks. By providing the analyst with alerts to indications of disease-related activities, the prototype attempts to detect early signs of disease outbreak in non-traditional data sources, giving the analyst more time to focus on potentially interesting data while reducing the time spent investigating false alarms and insignificant events.