DocumentCode :
2363259
Title :
An integrated approach to text understanding
Author :
Hu, Peng ; Shi, Chunyi ; Wang, Kehong ; Zhang, Xiaomei
Author_Institution :
Dept. of Comput. Sci. & Tech., Tsinghua Univ., Beijing, China
fYear :
1993
fDate :
1-5 Mar 1993
Firstpage :
79
Lastpage :
85
Abstract :
Top-down analysis (prediction-based) and bottom-up analysis (syntax and semantics-based) are two typical approaches to text understanding; however, each has its drawbacks. The authors give a novel approach to Chinese text understanding. With the approach, the top-down analyzer (TA) consisting of a predictor and a substantiator is combined with the bottom-up analyzer (BA) which is made up of a syntactic analyzer and an intersentence relation analyzer through cooperation and result synthesis. In text understanding, TA sends predictions to guide BA´s bottom-up analysis, and the result synthesizer unites TA´s and BA´s results into a more complete and detailed result if they are consistent. This approach has been implemented on a distributed knowledge-based system. The preliminary experiments showed that the integrated approach may give a more complete and detailed result than the top-down or bottom-up approaches
Keywords :
computational linguistics; distributed processing; knowledge based systems; natural languages; BA; Chinese text understanding; TA; bottom-up analysis; distributed knowledge-based system; integrated approach; intersentence relation analyzer; result synthesis; syntactic analyzer; top-down analyzer; Artificial intelligence; Computer science; Knowledge based systems; Moon; Natural language processing; Synthesizers;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Artificial Intelligence for Applications, 1993. Proceedings., Ninth Conference on
Conference_Location :
Orlando, FL
Print_ISBN :
0-8186-3840-0
Type :
conf
DOI :
10.1109/CAIA.1993.366658
Filename :
366658
Link To Document :
بازگشت