The problem of language barriers emerges in the internationalization of Taobao, the largest online shopping marketplace in China. In order to translate the information of the Taobao listings for overseas purchasers, machine translation is superior to human translation in terms of speed and quantity. However, full automatic translation might fail or require too many resources to tackle products reviews due to complicated syntactico-semantic factors. In this talk, we extend the Multi-Connectionist Selective Modification Model to represent the aspect-based opinion extraction from Taobao products reviews, followed by its bilingual alignment based on the comparable corpora from Amazon.
Professor Dr Xuening LI teaches English linguistics and computer-aided translation at the School of Foreign Studies, Jiangnan University. His current research interests include computational linguistics and CAT.