The program overview is available. Please click Program Overview


NLP for Microblog Summarization

Kam Fai Wong

Massive volume of textual information, eg news, over the Internet results in the problem of information explosion. Direct reading of massive text is impractical if not incomprehensible. For this reason, there is a growing demand for automatic summarization technology in the digital era. Traditionally, the goal of automatic summarization is to identify relevant excerpts from the target document(s), digest them and represent them in a succinct form for easy reading. While research in summarization has been in good progress in past decades, its effectiveness has been roved extremely low for social media applications, such as microblogging in twitter, WeChat, etc. This is mainly due to word limitation of microblog messages, which in turn leads to lack of proper grammatical structures and discourse information. We investigate how conventional NLP echniques can be best applied to summarizing microblogs. Resembling a document in a traditional natural language text, we first cluster related microblogs to form a microblog document (m-doc). Within a m-doc, microblog paragraphs (m-para) are formed by the leading topics. Practically, a m-para is a microblog re-post tree. Each m-para is comprised of inter-related short microblog messages ('sentences'), which is referred to as m-sent. Also, links between m-sent and m-para are viewed as contextual information. In our research, we propose a new discourse integration technique over m-doc for microblog summarization purpose.

K.F. Wong obtained his PhD from Edinburgh University, Scotland, in 1987. After his PhD, he performed research in Heriot-Watt University (Scotland), UniSys (Scotland) and ECRC (Germany). At present he is Associate Dean (External Affairs) of the Faculty of Engineering, Professor in the Department of Systems Engineering and Engineering Management, and Director, Centre for Innovation and Technology (CINTEC), of the Chinese University of Hong Kong (CUHK) as well as Adjunct Professor of MoE Lab on High Confidence Software Technologies (PKU), Harbin Institute of Technology Shenzhen Graduate School, Northeastern University, China. His research interest centers on Chinese information processing, parallel database and information retrieval. He has published over 250 technical papers in these areas in various international journals and conferences and books. He is Fellow of BCS, HKIE, and IET. He is the founding Editor-In-Chief of ACM Transactions on Asian Language Processing (TALIP), Member of the Editorial Board, International Journal on Computational Linguistics and Chinese Language Processing (Taiwan) and Journal of Chinese Processing (China). He is the conference co-chair of BigComp2016, NLPCC2015, IJCNLP2011, APWeb'08, Shenyang and AIRS'2008, Harbin, China; Finance Co-chair of IJCNLP'2008, Hyderabad, India; and Conference co-chair of APWEB'2008; panel chair of VLDB2002, PC co-chair of IRAL03, ICCPOL01, ICCPOL99 and IJCNLP05; General Chair of AIRS04 and IRAL00; and also PC members of many international conferences.


Technologies behind Digital Personal Assistants

Mei-Yuh Hwang

From Apple Siri, Google Now, Microsoft Cortana, and Amazon Echo, the industry is proving that machine intelligence is becoming a reality. In this talk we will analyze Cortana architecture and go deep to look at some of the advanced technologies using neural network. The advancement of speech recognition enables users to interact with machines in a more natural way, especially in hands-free mobile scenarios. Statistical semantic understanding extracts structured semantics from unstructured text and thus enables machines to act on behalf of their masters. Is your digital personal assistant truly learning about you every day? Can she know your need before your ask? By examining the cost of building such an intelligent system, we learn about the constraints of current technologies and continue to innovate and seek for more advanced unsupervised and/or semi-supervised learning and self-learning intelligence. With neural network intelligence, maybe we are truly in a groundbreaking era.

Dr. Mei-Yuh Hwang is a principal science manager of the language understanding and speech recognition groups at Microsoft, China. Her teams have been deeply involved in building speech recognition and language understanding technologies that power Windows Phones Cortana, especially the Mandarin version. Expanding from Windows Phones, Dr. Hwang continues to march Cortana into all Microsoft devices, including Windows 10 desktop. Dr. Hwang has been working at Microsoft since 1994-2004 and 2008 through present. From 2004-2008, she led the Mandarin speech recognition component of the GALE project at University of Washington in Seattle and her system ranked the best in 2007. She obtained her PhD in Computer Science from Carnegie Mellon University in December 1993 and was one of the major contributors of SPHINX-II speech recognition system. Her passion in products brought her to Cortana and brought intelligence into personal assistants. She has published many conference and journal papers and currently serves as an associated editor for IEEE transactions on Audio, Speech and Language Processing. She is a senior member of IEEE.


Speech Recognition and Keyword Search for Low Resource Languages

Bin Ma

The performance of speech recognition systems for rich resource languages has been improved dramatically in recent years, benefited from a large amount of speech and text training data, large vocabulary pronunciation dictionary, advanced acoustic modeling technologies and large-scale n-gram language modeling. However, for the most of the spoken languages in the world, due to the cost of speech and text data collection and the lack of linguistic knowledge and language expertise, the development of a good performance speech recognition system is still a big challenge for these languages and thus the performance of spoken keyword search using speech recognition as the front-end for producing indexed audio representation is significantly affected. In this talk, the state-of-the-art speech recognition and keyword search technologies for low resource languages will be reviewed and our work for low resource spoken keyword search in the NIST (National Institute of Standards and Technology, US) Open Keyword Search Evaluations will be introduced.

Bin Ma received the B.Sc. degree in Computer Science from Shandong University, China, in 1990, the M.Sc. degree in Pattern Recognition & Artificial Intelligence from the Institute of Automation, Chinese Academy of Sciences (IACAS), China, in 1993, and the Ph.D. degree in Computer Engineering from The University of Hong Kong, in 2000.
He was a Research Assistant from 1993 to 1996 at the National Laboratory of Pattern Recognition in IACAS. In 2000, he joined Lernout & Hauspie Asia Pacific in Singapore as a Researcher working on speech recognition. From 2001 to 2004, he worked for InfoTalk Corp., Ltd in Singapore, as a Senior Researcher for speech recognition. He joined the Institute for Infocomm Research, Singapore in 2004 and is now working as a Senior Scientist and the Lab Head of Automatic Speech Recognition.
Dr Ma has served as a Subject Editor for Speech Communication (Elsevier) in 2009-2012, an Area Chair for INTERSPEECH 2013, the Technical Program Co-Chair for INTERSPEECH 2014, and is now serving as an Area Chair for INTERSPEECH 2016 and an Associate Editor for IEEE/ACM Transactions on Audio, Speech, and Language Processing. He has served as the Secretary of the Special Interest Group on Chinese Spoken Language Processing (CSLP), International Speech Communication Association (ISCA) in 2007-2010, and is now serving as the Vice Chairperson of ISCA-CSLP. He is also serving as a member of APSIPA Speech, Language, and Audio (SLA) Technical Committee.
Copyright © 2015 Chinese and Oriental Languages Information Processing Society