报告题目：Transfer Learning for Cross-Lingual and Cross-Domain SentimentAnalysis with Deep Neural Networks

报告人： 周光有 副教授

报告时间：2016年11月8日（周二）16:00

报告地点：逸夫楼C座314会议室

摘要：

Sentiment analysis (also known as opinion mining)refers to the use of natural language processing, text analysis and computational linguistics to identify and extract subjective information in source materials.In this talk, I will focus on two important tasks in sentiment analysis:Cross-lingual and Cross-domain sentiment classification. The fundamentalchallenge of the former task stems from a lack of overlap between the featurespaces of source language data and that of target language data. While the key challengefor the latter task is how to reduce the domain discrepancy and manual labelingcosts. To address the above challenges, I will present two novel deep neuralnetwork recently published in my group: one is called weakly shared deep neuralnetworks (WSDNNs), and the other is called bi-transferring deep neural networks(BTDNNs).

报告人简介：

Guangyou Zhou received his Ph.D.degree from National Laboratory of Pattern Recognition (NLPR), Institute ofAutomation, Chinese Academy of Sciences (IACAS) in 2013. Currently, he workedas an Associate Professor at the School of Computer, Central China NormalUniversity. His research interests include natural language processing andinformation retrieval. He has won the best paper award in COLING 2014 and NLPCC2014. Now he has served several program committees of the major internationalconferences in the field of natural language processing and knowledgeengineering, and also served as editorial board for several journals. In the past five years, he has publishedmore than 40 papers in the leading journals and top conferences, such as ACMTWEB, ACM TIST, IEEE TKDE, IEEE TASLP, ACL, SIGIR, IJCAI, CIKM, COLING etc.