Neural POS Taggers for Vietnamese Social Media Text

Thời gian: 15:00 đến 16:30 Ngày 13/06/2019

Địa điểm: B4-705, VIASM

Báo cáo viên: Dr. Ngo Xuan Bach

Tóm tắt:

This talk focuses on the task of part-of-speech (POS) tagging for Vietnamese social media text, which poses several challenges compared with tagging for conventional text. We discuss how to  take advantages of deep learning and manually engineered features to overcome the challenges of the task. Different neural network architectures and graphical models including RNNs, CNNs, CRFs, and their integration will be investigated. We present various types of manually designed features in addition to automatically learned features to capture the characteristics of Vietnamese social media data. Experiment results and open questions will be also discussed. 

Affiliation: Department of Computer Science, Faculty of Information Technology, Posts and Telecommunications Institute of Technology, Hanoi.

Short bio: Ngo Xuan Bach received his B.Sc. degree in computer science from UET, VNU Hanoi (2006), M.Sc. and Ph.D. degrees in information science from JAIST Japan (2011 and 2014). He won several awards, including Honda Y-E-S Award (2006), Japan Institute of Electronics, Information, and Communication Engineers (IEICE) Award (2011), Outstanding Paper Award for Young Researchers in Computer & Communications, Japan (2013). He is now Head of Department of Computer Science, Faculty of Information Technology, PTIT. His research interests include statistical natural language processing, machine learning, and recommender systems.