Prof. Haizhou Li, IEEE Fellow, National University of Singapore, Singapore
Speech Title: Brain-inspired Computational Model for Selective Listening
Abstract: Humans have a remarkable ability to pay their auditory attention only to the speech of interest, that we call selective listening, in a multi-talker environment or a Cocktail Party. However, signal processing approach to speech separation and/or speaker extraction from multi-talker speech remains a challenge for machines. In this talk, we study the a brain-inspired computational model for monaural speech separation and speaker extraction that enables selective listening, speech recognition, speaker recognition at Cocktail Party. We discuss the computational auditory models, technical challenges and the recent advances in the field.
Biography: Haizhou Li received the B.Sc, M.Sc, and Ph.D degrees in electrical and electronic engineering from South China University of Technology, Guangzhou, China in 1984, 1987, and 1990 respectively. He is now a Professor at the Department of Electrical and Computer Engineering, and the Department of Mechanical Engineering of the National University of Singapore.
Professor Li has worked on human language technology in academia and industry since 1988. He has taught in The University of Hong Kong (1988-1989), South China University of Technology in Guangzhou, China (1990-1994), Nanyang Technological University, Singapore (2006-2016), University of Eastern Finland (2009), and the University of New South Wales (2011-). He was a Visiting Professor at CRIN/INRIA in France (1994-1995). As a technologist, he was a Research Manager in Apple-ISS Research Centre (1996-1998), Research Director of Lernout & Hauspie Asia Pacific (1999-2001), Vice President of InfoTalk Corp. Ltd (2001-2003), Principal Scientist and Department Head of Human Language Technology at the Institute for Infocomm Research, Singapore (2003-2016), and the Research Director of the Institute for Infocomm Research (2014-2016). He co-founded Baidu-I2R Research Centre in Singapore (2012). Professor Li was known for his technical contributions to several award-winning speech products, such as Apple's Chinese Dictation Kits for Macintosh (1996) and Lernout & Hauspie's Speech-Pen-Keyboard Text Entry Solution for Asian languages (1999). He was the architect of a series of major technology deployments that include TELEFIQS voice-automated call centre service in Singapore Changi International Airport (2001), voiceprint engine for Lenovo A586 Smartphone (2012), and Baidu Music Search (2013).
Professor Li's research interests include speech information processing, natural language processing, and human-robot interaction. He has published over 300 technical papers. Professor Li has served as Associate Editor (2008-2012), Senior Area Editor (2014-2016), and Editor-in-Chief (2015-2017) of IEEE/ACM TRANSACTIONS ON AUDIO, SPEECH AND LANGUAGE PROCESSING. He has also served as Associate Editor of Computer Speech and Language (2012-), and Springer International Journal of Social Robotics (2008-), and a Member of IEEE Speech and Language Processing Technical Committee (2013-2015). Professor Li is the President of the International Speech Communication Association (ISCA, 2015-2017), the President of Asia Pacific Signal and Information Processing Association (APSIPA, 2015-2016), the President of the Chinese and Oriental Language Information Processing Society (COLIPS, 2011-2013), the Vice President of the Asian Federation of Natural Language Processing (AFNLP, 2015-2016). Professor Li served as the General Chair of ACL 2012 and INTERSPEECH 2014, the Local Arrangement Chair of SIGIR 2008 and ACL-IJCNLP 2009, and the Technical Program Chair of ISCSLP 1998, APSIPA Annual Summit and Conference 2010, IEEE Spoken Language Technology Workshop 2014, and IEEE ChinaSIP 2015. Professor Li was the recipient of National Infocomm Awards 2002, Institution of Engineers Singapore (IES) Prestigious Engineering Achievement Award 2013 and 2015, President's Technology Award 2013, and MTI Innovation Activist Gold Award 2015 in Singapore. He was named one of the two Nokia Visiting Professors in 2009 by Nokia Foundation. He is a Fellow of IEEE, and Member of ACL, ACM, IEICE and ISCA.
Prof. Wong Kam Fai, ACL Fellow, The Chinese University of Hong Kong, China
Speech Title: XC - an Explainable AI Method to make your Chatbot Trustable
Abstract: Chatbots are popular AI applications in today’s commercial sector. They make extensive use of Natural Language Processing (NLP) and Dialogue System techniques to understand what a human user wants and guide him/her to the desired outcomes. Existing research in chatbot mainly focuses on performance improvement (e.g. to minimize the number of turns to answer a query) and deep learning approaches are commonly used for this purpose. Due to the opaque “black box” nature of deep learning functions, layman users cannot understand the reason and logic behind the chatbot’s decision of his/her request. In fact, lack of explanability renders chatbots confusing and user-unfriendly. By the same token, developers are not sure about the features and factors which contribute most to the trained model. Recently, eXplainable AI (XAI) technology are introduced to overcome this predicament. However, existing XAI methods are mainly used to explain NLP applications based on deep learning on static data. Yet, in practice, each turn (ie text presented by the chatbot system) is generated dynamically based on the conversation history and system interactions with the users. For this reason, existing XAI methods are ineffective for chatbots which involves dynamic data generated from chatting through multiple turns of man-machine interaction. This lays down the objective of or project: to research and develop a novel XAI method, referred to as XC (Explainable Chatbot), to explain features and temporal factors that determine a chatbot’s response or decision. XC can identify (1) the importance of the key phrases leading to the answer and (2) the change of the phrases between conversation turns, which can be viewed as the derivative of the phrase over turns. The former is the “feature” which enables users to uncover the relationships between aspects, sentences and keywords embedded in a dialogue; and the latter is the “time factor” which reveals the chatting behavior of the users reflected by the change of the feature over turns. Respectively, these two pieces of information make the decision and the decision process of a chatbot transparent and traceable. Furthermore, they identify potential flaws in a chatbot and analyze the behavioural changes of a user, which in turn helps developers adjust the underlying chatbot algorithms.
Biography: Kam-Fai Wong obtained his PhD from Edinburgh University, Scotland, in 1987. After his PhD, he was researchers 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, Director, Centre for Innovation and Technology (CINTEC), and Associate Director, Centre for Entrepreneurship, The Chinese University of Hong Kong (CUHK) as well as Associate Director, Key-Laboratory of High Confidence Software Technologies (PKU), Ministry of Education, China. Also, he is a co-founder of the first Chinese news search portal, namely Wisers Information Ltd., in 1998. His research interest focuses on Natural Language Processing especially in Chinese and Social Media processing. He has published over 250 peer reviewed papers in these areas. He has published the book "Introduction in Chinese Natural Language Processing" (2009), which is the first of its king written in English. His research work is summarized by the book: "Social Media Content Analysis: NLP and Beyond" which was published in 2018. He is the founding Editor-In-Chief of ACM Transactions on Asian Language Processing (TALIP) and the General Chair of AACL2020, BigComp2016, NLPCC2015, IJCNLP2011, AIRS2008 and ICCPOL2006. Also, he was the President of Asia Federation of NLP (AFNLP), 2014-2016.
Prof. Tianrui Li, Southwest Jiaotong University, China
Speech Title: Big Data Intelligence: Challenges and our Solutions
Abstract: Big Data Intelligence has become a hot research topic in the area of artificial intelligence. This talk aims to outline the challenges on Big Data Intelligence. Then our solutions for Big Data Intelligence are provided, which cover the following aspects. 1) To improve the data quality, a hierarchical entropy-based approach is demonstrated to evaluate the effectiveness of data collection. 2) To preprocess the collected data, a multi-view-based method is illustrated for filling missing data. 3) To fuse multiple different sources of data, several deep-learning-based models are developed. Finally, some recent progress on applications of big data in natural language processing are provided.
Biography: Dr Tianrui Li is a Professor and the Director of the Key Lab of Cloud Computing and Intelligent Technique of Sichuan Province, Southwest Jiaotong University, China. Since 2000, he has co-edited 6 books, 11 special issues of international journals, received 16 Chinese invention patents and published over 360 research papers (e.g., AI, IEEE TKDE, IEEE TEC, IEEE TFS, IEEE TIFS, IEEE ASLP, IEEE TIE, IEEE TC, IEEE TVT) in refereed journals and conferences (e.g., ACL, IJCAI, KDD, UbiComp, WWW, ICDM, CIKM, EMNLP). 5 papers were ESI Hot Papers and 18 papers was ESI Highly Cited Papers. He serves as the area editor of International Journal of Computational Intelligence Systems (SCI), editor of Knowledge-based Systems (SCI) and Information Fusion (SCI), associate editor of ACM Transactions on Intelligent Systems and Technology, etc. He is an IRSS Fellow and Steering Committee Chair (2019-2020), IEEE CIS ETTC member (2019-2020), IEEE CIS SMC member (2018-2020), a senior member of ACM and IEEE, ACM SIGKDD member, Chair of IEEE CIS Chengdu Chapter (2013-2018) and Treasurer of ACM SIGKDD China Chapter. Over 80 graduate students (including 9 Post-Docs, 25 Doctors) have been trained. Their employment units include Microsoft Research Asia, Sichuan University, Huawei, JD, Baidu, Alibaba, and Tencent. They have received Best Papers/Dissertation Awards 20 times, Champion of Sina Weibo Interaction-prediction at Tianchi Big Data Competition (Bonus 200,000 RMB), Second Place of Social Influence Analysis Contest of IJCAI-2016 Competitions and Second Place of Weather forecast Contest of AI Challenger 2018.
Prof. Dr. Uwe Quasthoff, Universität Leipzig, Germany
Speech Title: Corpora as A Resource for IR
Abstract: Knowledge about words is helpful for IR. Knowledge about single words like word frequencies and knowledge about replacement candidates like inflected forms or synonyms are used heavily. Syntactic and semantic relations between consecutive words are of interest for text understanding. POS tagging and syntactic parsing is the bases for a deeper semantic analysis with statistical methods.
The talk will give an overview of the complete pipeline of corpus building and exploration: Crawling and preprocessing (language identification, sentence segmentation, tokenization, de-duplication, POS-Tagging etc.), word co-occurrences and semantic similarities using word embeddings, word and text classification problems.
As an approach to relation extraction and sentence understanding, so-called typical sentences are used: Sentences of simple syntactic structure repeatedly found with rich lexical variability. The syntactic structures are selected by high frequency, and with large corpora they allow the usage of word similarities to cluster such sentences to basic statements.
Biography: Prof. Uwe Quasthoff works at the Natural Language Processing Group at the Department of Computer Science at Leipzig University in Germany. His main research topics are language independent methods in Natural Language Processing, building very large corpora, and the structure of natural language. The research method is the analysis of large text corpora with statistical and pattern based as well as machine learning methods. The Leipzig Corpora collection (http://corpora.uni-leipzig.de/) started in 1995 and now contains pre-processed text collections and monolingual dictionaries in more than 250 languages. The approach is language independent, hence the algorithms for further processing apply usually to a large group of languages. The analysis of word co-occurrence patterns was the starting point for machine learning used for semantic similarities in different granularities.
Prof. Wookey Lee, VOICE AI Research Institute & Biomedical Science and Engineering, Inha University, Korea
Speech Title: How to Communicate BEYOND MARS AND VENUS
Abstract: Many people say that they cannot understand some people. It may come from just personality difference, or characteristics, MBTI diversity, or superiority complex issue, or racism, just mother tongue or multi-language semantics. What about ethnicity or nationalism, religious difference, or historical events, family history or vengeance, or generation gap, or of which the husband from Mars and the wife from Venus. Moreover when AI society is coming soon, which will probably say that my AI partner is problematic. As long as they have to stay in the same community, some kind of solution scheme will be needed. Thing is that they never originated different planets, and yet have to face and solve problems together as a team. Extremely different opinions and abilities will sometimes be a blessing which is like a "Synthese" from "These" and "Antithese". We can think about team chemistry. Can we sure that the team is greater than the sum of its members’ capabilities? Forging a team depends upon strong collaborations among the team members amalgamated with each member’s abilities. A team will be a combination if it is composed of talented experts that can effectively collaborate and communicate with each other. Moreover, when forming a team, one might not always try to find ”similar” members nor an ”isomorphic” ones, but different complementary skills are of importance. For example, in most military operations, sports teams, research communities, hospital surgeries, business enterprises, highly skilled experts, and their interactions are required. Therefore, forming a team with high team chemistry results in team members that play complementary roles, communicate, and collaborate effectively to reinforce teamwork.
Biography: Wookey Lee received the B.S., M.S., and Ph.D. degrees in Industrial Engineering from Seoul National University, Korea, in 1987 and 1993, and 1996, respectively. He finished his FSE in Carnegie Mellon University, USA in 2000, and got a diploma in TEFL, ISS, Canada, 2003. He is now a professor in the Department of Industrial Engineering as well as the Program in Medical Science and Engineering in Inha University, Incheon, South Korea.
Professor Lee is a director of VOICE AI Research Institute, Inha University, and has been the director of Business Industry center, and Information Communication center, and is now an adjunct professor of Research Center for Social Informatics, Kwansei Gakuin University, Japan, and research professor of iHub-Anubhuti Execution committee, India, was a visiting scholar in the Department of Computer Science at the UBC, Canada co-hosted by Prof. Raymond Ng and Prof. Laks Lakshumanan. He has served several international academic societies and conferences such as ACM CIKM, BigComp, Dasfaa, Dexa, Enterprise Architecture, EDB, ICDE, SCM, SocialCom, VLDB, etc. He has been the editor in chief of Information Tech. and Arch. and Journal of Bigdata Services and Applications, and an associate editor for World Wide Web Journal, Supercomputing, Sensors, Distributed Sensor Networks, Cluster Computing, Journal of Big Data, Computers in Industrial Engineering, etc. He was the president of Korea Database Society, and is serving as a Steering Committee member of BigComp, and the IEEE TCDE executive committee members. He got the best paper awards from IEEE TCSC, ACM Bigdas, KORMS, KIISE, KOSIPER, etc. His research areas are Deep Learning, VOICE with Artificial Intelligence, Patent Information Systems, Trajectory Security, and database and graph theory. He has published more than 300 technical papers in internationally refereed journals and conferences including “Effective privacy preserving data publishing by vectorization." Information Sciences (2020), and pertained more than 100 patents including Device for Supplementing Voice and Method for Controlling the Same(US9232297B2).
Assoc. Prof. Phayung Meesad, King Mongkut's University of Technology North Bangkok, Thailand
Speech Title: Deep Learning and Applications
Abstract: Deep Learning has been extremely successful in many fields such as image processing and natural language processing. Convolutional Neural Network (CNN) and Long Short Term Memory Recurrent Neural Network (LSTM-RNN) are probably most search hit in Deep Learning fields. CNNs are popular in image processing while LSTMs seem to play big roles in Time series data and natural language processing. This talk gives brief reviews about Deep Learning focusing on CNNs and LSTMs as well as their applications.
Biography: Phayung Meesad was graduated in Master of Science and Doctoral of Philosophy in Electrical Engineering from Oklahoma State University, Stillwater, USA in 1998 and 2002, respectively. He is now an Associate Professor and Director of Central Library, KMUTNB Smart Digital Library, King Mongkut's University of Technology North Bangkok, Thailand. His research of interests are Computational Intelligence, Artificial Intelligence, Machine Learning, Deep Learning, Data Analytics, Data Science, Data Mining, Digital Signal Processing, Image Processing, Business Intelligence, Time Serires Analysis, Natural Language Processing.