Hua Wu, Doctor, Baidu

Hua Wu is the Chair of Baidu Technical Committee and Chief Scientist of Baidu NLP. Her research fields include machine translation, natural language processing(NLP), and knowledge graph. She was the Program Co-Chair of ACL (the Association for Computational Linguistics) in 2014, and had served as area chairs for international conferences including ACL and IJCAI. She was named as one of the “20+ leading women in AI research” by Forbes, rewarded as Distinguished Engineer and one of the 10 Best Practioners in Science and Technology by Chinese Institute of Electronics(CIE), and selected as a member of Beijing's Hundred-Thousand-Ten-Thousand Talent Program. She has won one second prize of the National Science and Technology Progress Award, four first prizes of the CIE Science and Technology Award and one silver prize of China Patent Award.

Qi Zhang, Professor, Fudan University

Zhang Qi is a professor in the School of Computer Science and Technology of Fudan University. His main research direction is information retrieval and natural language processing. He has published more than 70 papers in important international conferences and journals such as ACL, SIGIR, TKDE, and EMNLP. He also served as area chair, senior program committee member or program committee nember for a few international conferences, including ACL, SIGIR, EMNLP, IJCAI and AAAI. He won the second prize of Shanghai Science and Technology Progress Award, the second prize of Science and Technology Progress Award of the Ministry of Education, the ACM Shanghai New Star Nomination Award, the IBM Faculty Award, and the Chinese Information Society Qian Weichang Chinese Information Processing Science and Technology Award--Hanwang Youth Innovation First Prize, etc.

Xiaojun Wan, Professor, Peking University

Xiaojun Wan is a professor in Institute of Computer Science and Technology of Peking University and the director of Lab of Language Computing and Web Mining. He obtained his bachelor, master and doctor’s degrees from Peking University. His research interests include document summarization, text generation, sentiment analysis and semantic computing. He has been editorial board member or associate editor for TACL, Computational Linguistics and Natural Language Engineering. He is PC co-chair of EMNLP-IJCNLP 2019 and also served as area chair or SPC for a few international conferences, including ACL, NAACL, EMNLP, IJCAI and AAAI. He has received ACL 2017 Outstanding Paper Award, IJCAI 2018 Distinguished Paper Award, CAAI WUWENJUN Technical Innovation Award (2017) and CCF NLPCC Distinguished Young Scientist (2017).

Title: Deep Generative Models for Text Generation

Speakers: Lei Li,Hao Zhou ,ByteDance AI Lab

Abstract : Natural language generation has been a fundamental technology in many applications such as machine writing, machine translation, chatbots, web search, and computational advertising. In this tutorial, we will give an overview about deep neural models for natural language generation. We will first introduce the basic sequence to sequence learning model and its variants (i.e. Transformer). Then we will describe two deep generative models for text generations from theory to practice, including Variational Auto-encoder (VAE) and Generative Adversarial Networks (GAN). Both of VAE and GAN have advantages than vanilla sequence to sequence models. VAE has a smoothed latent space which enables interpolation and sampling for text generation, and GAN based generation models can be equipped with a task-oriented designed discriminator, always achieving better text generation qualities. The third paradigm is to generate using Bayesian sampling methods, through which we can control the generation in diversity and interpretability. Different deep generation models arouse different applications of text generation. In the last part, we will present various practical applications like story-telling, data-to-text, Question generation and etc.

Author Introduction: Dr. Lei Li is Director and a research scientist of ByteDance AI Lab. His research interest is on machine learning, language understanding and generation. Lei received his B.S. in Computer Science and Engineering from Shanghai Jiao Tong University (ACM class) and Ph.D. in Computer Science from Carnegie Mellon University, respectively. His dissertation work on fast algorithms for mining co-evolving time series was awarded ACM KDD best dissertation (runner up). His recent work on AI writer Xiaomingbot received 2nd-class award of WU Wenjun AI prize in 2017. Before ByteDance, he worked at EECS department of UC Berkeley and Baidu's Institute of Deep Learning in Silicon Valley. He has served in the Program Committee for ICML 2014, ECML/PKDD 2014/2015, SDM 2013/2014, IJCAI 2011/2013/2016/2019, KDD 2015/2016/2019, 2017 KDD Cup co-Chair, KDD 2018 hands-on tutorial co-chair, EMNLP 2018/2019, AAAI 2019 senior PC, and as a lecturer in 2014 summer school on Probabilistic Programming for Advancing Machine Learning. He has published over 40 technical papers and holds 3 US patents. (Homepage: https://lileicc.github.io/)

Author Introduction: Dr. Hao Zhou is a researcher at ByteDance AI Lab. His research interests are machine learning and its applications for natural language processing, including syntax parsing, machine translation and text generation. Currently he focuses on deep generative models for NLP. Previously he received his Ph.D. degrees in 2017, from Nanjing University. He has served in the Program Committee for ACL, EMNLP, IJCAI, AAAI and NIPS. He has more than 20 publications in prestigious conferences and journals, including ACL, EMNLP, NAACL, TACL, AAAI, IJCAI, NIPS and JAIR. He will give a tutorial on discreteness of neural NLP at EMNLP’2019 (Homepage: https://zhouh.github.io/).

Title: Natural Language and Multimodal Intelligence, and their Industrial Applications

Speakers: Xiaodong He, JD AI Research

Abstract : I will first briefly review the impact of deep learning techniques on speech, language, and computer vision, and then focus on the research frontier in natural language processing (NLP) from two aspects. Language intelligence is cognitional intelligence that is unique to human beings. NLP technology can be roughly divided into two aspects: First, how to let AI understand humans, such as understanding intentions, parsing semantics, recognizing emotions, and information matching (e.g., IR & recommendation); Second, how to make AI understandable by humans, such as text summarization, content generation, topic expansion, Emotional expression and dialogue, etc. On the other hand, build upon these fundamental AI technologies, the research on human-machine dialogue and interactive systems has also made breakthroughs. I will also explore the frontier directions of semantic analysis, knowledge mapping, machine reading, emotion and style expression. Meanwhile, I will introduce some recent advances in the field of multi-modal intelligence, including language and image understanding, reasoning and generation. Specifically, I will introduce language-vision semantic representation modeling, and then introduce the progress in image description, that is, understanding visual content and generating natural language description; visual question answering, that is, reasoning across natural language and vision to infer answers; and text-to-image synthesis, that is, generating images according to natural language description. On the other hand, the development of AI technology will also subvert the pattern of many fields, such as democratizing content creation for everyone. I will review some recent examples of AI technology in the field of artistic creation, too. At the same time, we are more concerned than ever about how these technological advances will change our lives. Taking retail industry as an example, I will introduce how we can use AI technology to drive retail from simple transactions to more experienced conversational services in JD.COM. E.g., the application of advanced dialogical artificial intelligence technology will enable us to build an intelligent interactive dialog system capable of large-scale complex scenes, which can not only quickly answer customer's questions, but also provide sympathetic comfort and guidance for emotional appeals, thus providing timely and thoughtful service for customers. Finally, I will have an open conversation with the audience on the progress of AI technology and its impact of industry.

Author Introduction: Dr. He Xiaodong is the deputy Managing Director of JD AI Research, and Head of the Deep Learning, Speech and Language Lab, at JD.COM. He also holds adjunct and honorary faculty positions at the University of Washington (Seattle), CUHK-SZ, Tongji University, and Central Academy of Fine Arts. Dr. Xiaodong He is IEEE Fellow. His research mainly focuses on artificial intelligence (AI) areas, including deep learning, natural language processing, speech recognition, computer vision, and multimodal intelligence. He and his colleagues have invented a series of algorithms including the Deep Structured Semantic Model (DSSM), the Hierarchical Attention Network (HAN), the Image Caption Bot, Bottom-up Attention, that have been widely applied to important AI scenarios, received more than 10000 citations, and won multiple paper awards and key AI Challenges. He also holds editorial positions of IEEE and ACL journals and serves as area chairs of major conferences in these fields. Prior to joining JD.COM, Dr. He Xiaodong is a Principal Researcher and Research Manager of the Deep Learning Technology Center (DLTC) of Microsoft Research, Redmond, USA. Dr. He Xiaodong holds a bachelor’s degree from Tsinghua University and a PhD degree from University of Missouri—Columbia, USA.

Title: Lifelong and Continual Learning

Speakers: Bing Liu, University of Illinois at Chicago (UIC)

Abstract : Classic machine learning is characterized by isolated single task learning in closed environments. It typically requires a large amount of labeled training data in order to learn effectively, and is only suitable for well-defined and narrow tasks. Going forward, this isolated learning paradigm is no longer sufficient. For example, it is almost impossible to pre-train a chatbot, a self-driving car, or any other AI agent so that it can function seamlessly in a real-world open environment, because it is very hard, if not impossible, for humans to provide knowledge or labeled data to cover all possible scenarios that the agent may encounter. The agent thus must learn on the job in its interaction with the environment continually, retain the learned knowledge, and use it to learn more and better in the future. When faced with an unfamiliar situation, it must adapt its knowledge to deal with it and learn from it. This general learning capability is one of the hallmarks of the human intelligence. Without it, an AI agent is probably never going to be truly intelligent. In recent years, we have seen an emerging and growing research trend under the name of lifelong learning, continual learning, open-world learning, and/or never-ending learning that try to endow AI agents with the capability. In this tutorial, I introduce these topics and discuss some recent progresses.

Author Introduction: Bing Liu is a distinguished professor of Computer Science at the University of Illinois at Chicago (UIC). He received his Ph.D. in Artificial Intelligence from University of Edinburgh. Before joining UIC, he was a faculty member at the School of Computing, National University of Singapore. His research interests include sentiment analysis, lifelong learning, chatbot, natural language processing (NLP), data mining, and machine learning. He has published extensively in top conferences and journals. Three of his papers have received Test-of-Time awards: two from SIGKDD (ACM Special Interest Group on Knowledge Discovery and Data Mining), and one from WSDM (ACM International Conference on Web Search and Data Mining). He is also a recipient of ACM SIGKDD Innovation Award, the most prestigious technical award from SIGKDD. He has authored four books: two on sentiment analysis, one on lifelong learning, and one on Web mining. Some of his work has been widely reported in the international press, including a front-page article in the New York Times. On professional services, he has served as program chair of many leading data mining conferences, including KDD, ICDM, CIKM, WSDM, SDM, and PAKDD, as associate editor of many leading journals such as TKDE, TWEB, DMKD and TKDD, and as area chair or senior PC member of numerous NLP, AI, Web, and data mining conferences. Additionally, he served as the Chair of ACM SIGKDD from 2013-2017. He is a Fellow of the ACM, AAAI, and IEEE.

Title: Low Resource Neural Machine Translation

Speakers: Shujie Liu, Microsoft Research Asia

Abstract : With the help of deep learning, neural machine translation (NMT) has made rapid advances in recent years. For rich language pairs with large bilingual corpus, NMT model even reaches human quality especially for news domain. But for low resource translation tasks, such as  minority language, or specific domains, large bilingual data is not easy to acquire, and leads to poor translation performance. In this talk, based on the brief introduction of NMT technique,  I will introduce several techniques to deal with the low resource NMT problem. In this talk, two  semi-supervised methods are introduced, of which, bidirectional agreement encourages agreement between left-to-right and right-to-left decoding results, and joint training enhances the effect of monolingual source and target data by iteratively boosting the  source-to-target and target-to-source translation models. By leveraging the large bilingual corpus from rich languages, a triangular architecture is introduced to jointly train four translation models between these two rich languages and one poor language. At the end, I will also introduce the recent progress on unsupervised NMT in the extreme case that no bilingual corpus is available. 

Author Introduction: Shujie Liu is a lead researcher in Microsoft Research Asia. Shujie received B.S. and Ph.D. degrees from Harbin Institute of Technology.  His research interests include natural language processing, speech processing and deep learning. He is now working on fundamental NLP and speech problems, models, algorithms and innovations. His research work has been applied in many Microsoft products, including Microsoft Translator, Skype Translator, Microsoft IME, XiaoIce, Microsoft speech service. He has published more than 40 papers in the top conferences and journals, including 11 ACL, 6 AAAI and 5 EMNLP. He also wrote one book , published by Higher Education Press in 2018.

Title: Knowledge-Guided Natural Language Processing

Speakers: Zhiyuan Liu, Tsinghua University

Abstract : Recent years have witnessed the advances of deep learning for NLP. As a typical data-driven approach, deep learning has issue of model interpretability. How can we incorporate rich human knowledge on languages, world and cognition into deep learning? The knowledge-guided deep learning is expected to be more interpretable and robust for NLP tasks. In this tutorial, we will introduce the recent advances of knowledge-guided NLP and outlook several future research directions.

Author Introduction: Zhiyuan Liu is an associate professor at the Department of Computer Science and Technology, Tsinghua University. He received his Ph.D. degree in Computer Science from Tsinghua in 2011. His research interests include representation learning, knowledge graphs and social computation, and has published more than 60 papers in top-tier conferences and journals of AI and NLP including ACL, IJCAI and AAAI, cited by more than 3,600 according to Google Scholar. He is the recipient of the Excellent Doctoral Dissertation of Tsinghua University, the Excellent Doctoral Dissertation of CAAI (Chinese Association for Artificial Intelligence), and Outstanding Post-Doctoral Fellow in Tsinghua University. He serves as Youth Associate Editor of Frontiers of Computer Science, Area Chairs of ACL, EMNLP, COLING, IJCNLP, etc.

Title: Foundations and Trends for Personalized Recommendation

Speakers: Min Zhang, Tsinghua University

Abstract : Personalized recommendation has played vital roles in current information consuming environment. This tutorial mainly includes two parts: the foundations and the trends. In the first section, we will introduce the fundamental problems for personalized recommender systems, including user intent and requirements, challenging problems and the state-of-art technologies. In the second section, we will focus on the new trending topics in the related area, including (and not limited to): user satisfaction and evaluation, explainable recommendation, recommendation based on knowledge graphs and inferences, cross-domain heterogeneous recommendation, and fairness issue in recommender system, etc. Finally， we will make some discussions on the future direction with the participants.

Author Introduction: Dr. Min Zhang is a tenured associate professor in the Dept. of Computer Science & Technology, Tsinghua University, specializes in Web search and recommendation, and user modeling. She is the vice director of Intelligent Technology & Systems lab at CS Dept., the executive director of Tsinghua-MSRA Lab on Media and Search. She also serves as associate editor for the ACM Transaction of Information Systems (TOIS), Tutorial Chair of SIGIR 2019, Short Paper Chair of SIGIR 2018, Program Chair of WSDM 2017, etc. She has published more than 100 papers with 3500+ citations and H-index scores of 32. She was awarded Beijing Science and Technology Award (First Prize) in 2016, and the Excellent Teacher Award for Computer Sciences in Universities of China in 2018. etc. She also owns 12 patents. And she has made a lot of cooperation with international and domestic enterprises. Homepage: http://www.thuir.cn/group/~mzhang/