Parallel Text Mining and Cross-language Information Retrieval Using Statistical Translation Model

Jiang Chen

RALI, DIRO

Le 13 octobre 1999 à 11 h 30

Salle 3195, Pavillon André-Aisenstadt


In this research we established a Chinese-English statistical translation model trained by the parallel corpus collected from the Web. The objective is to test the performance of such a model in cross-language information1. Parallel text mining -- We will firstly present the adopted algorithm mining system for parallel texts.2. Translation model training -- The important procedures in the training process will then be discussed. We evaluated the translation quality of the model by examining a set of randomly chosentranslations.3. CLIR -- Experimental results on both English-Chinese and Chinese-English CLIR will be given. Some preliminary analysis will indicate the problems in this approach and suggest possible solutions.


Pour recevoir les annonces hebdomadaires par courriel, visitez http://rali.iro.umontreal.ca/rali/?q=fr/node/1631

Liste des autres séminaires