Marti Hearst
Marti Hearst[Word-sense disambiguation|] is a professor in the School of Information at the University of California, Berkeley. She did early work in corpus-based computational linguistics, including some of the first work in automating
sentiment analysis, and
word sense disambiguation. She invented an algorithm that became known as "Hearst patterns" which applies lexico-syntactic patterns to recognize hyponymy relations with high accuracy in large text collections, including an early application of it to WordNet; this algorithm is widely used in commercial text mining applications including ontology learning. Hearst also developed early work in automatic segmentation of text into topical discourse boundaries, inventing a now well-known approach called TextTiling.
Hearst research is on user interfaces for search engine technology and big data analytics. She did early work in user interfaces and information visualization for search user interfaces, inventing the TileBars query term visualization. Her Flamenco research project investigated and developed the now widely used faceted navigation approach for searching and browsing web sites and information collections. She wrote the first academic book on the topic of Search User Interfaces.
Hearst is an Edge Foundation contributing author and a member of the Usage panel of the American Heritage Dictionary of the English Language.
Hearst received her B.A., M.S., and Ph.D. in Computer Science, all from U.C. Berkeley. In 2013 she became a fellow of the Association for Computing Machinery.
