Feiyu Xu

Organization: AI Lab, Lenovo Research, Lenovo Group
Homepage: http://research.lenovo.com/webapp/view_English/researchField.html

Feiyu Xu is Vice President of Lenovo Group and Head of AI Lab at Lenovo Research since March 2017. She was Principal Researcher and Head of Research Group Text Analytics in the Language Technology Lab of German Research Center of Artificial Intelligence (DFKI) before she joined Lenovo. Since 2004, Dr. Xu is vice-director of the Joint Research Laboratory for Language Technology of Shanghai Jiao Tong University and Saarland University. Feiyu Xu was also co-founder and managing director of Yocoy Technologies GmbH, a 2007 spin-off from DFKI that is developing next generation mobile language and travelguides.

Dr. Xu has extensive experience in multilingual information systems, information extraction, text mining, big data analytics, business intelligence, question answering and mobile applications of NLP technologies. She has broad and in-depth experience of the total cycle of innovation in her expert areas, from basic research, to application and development and finally to products and their commercialization.

In 2013, Feiyu Xu has won a Google Focused Research Award for Natural Language Understanding as co-PI with Hans Uszkoreit and Roberto Navigli. In 2014, Feiyu Xu was honored as DFKI Research Fellow. Her research is documented in almost 100 publications including conference papers for ACL, COLING, EMNLP, CONLL, NAACL, LREC etc. She was area chair of EACL 2017 for text mining, information extraction and question answering and is area chair of ACL 2018 for Information Extraction and Text Mining.

Keynote talk - Multilingual and Multimodal Smart Conversational Agent for Real World Customer Relationship Management

Artificial Intelligence can contribute to the improvement of different processes in companies, for example, supply chain management or customer relationship management. Call centers play important roles for after sale customer care. Automated dialog systems, or chatbots, have the potential to greatly facilitate customer support and help desk services. Customer support and help desk services for consumer products such as mobile phones or computers are often very labor- and time-intensive, due to the technical nature of many customer questions and the complexity of the devices and their software. In many cases, resolving such issues is a multi-step process of determining the user's problem and intent, and then guiding the user through potentially several issue resolution steps. This talk will describe a multilingual and multimodal conversational agent for real world call centers. This conversational agent can, on the one hand, provide solutions to the trouble shooting and how-to questions and, on the other hand, deal with small talks and fact-based requests. The conversational agent was automatically learned from a large collection of the real world call center chats. This conversational agent is embedded in a comprehensive solution where human call center agents can interact with the automatic results.

John D. Kelleher

Organization: Dublin Institute of Technology, Kevin Street, Dublin 8
Phone: +353 1 402 4789
Homepage: http://www.linkedin.com/in/johndkelleher/, http://www.comp.dit.ie/jkelleher

Professor John D. Kelleher is Academic Leader (Senior Lecturer III) at the Information, Communications and Entertainment (ICE) research institute at Dublin Institute of Technology that integrates research from a number of different research centres, spanning domains from photonics and antennae research, through communications and networking infrastructure, and onto data science, digital media, and arts and humanities research. Prof. Kelleher is motivated by developing solutions to technical challenges (be they research or business focused) and by mentoring people to develop their careers. He has expertise in Machine Learning, Data Science, Natural Language Processing and Artificial Intelligence. He is a co-author of two books on Machine Learning and Data Science published by MIT Press and a senior author of 75 scientific publications. In 2015 Prof. Kelleher was granted an IBM Shared University Research Award.

Keynote talk - Machine Learning and Deep Learning in AI

The talk will introduce the field of Deep Learning and position it relative to Artificial Intelligence and Machine Learning. It will overview the state of the art in Deep Learning across a number of fields, with a particular focus on applications to language and vision processing. The talk will conclude by discussing some of the current challenges faced by Deep Learning researchers and some of the potential future directions for Deep Learning research.

Milica Gasic

Organization: University of Cambridge, UK
Homepage: http://mi.eng.cam.ac.uk/~mg436/

Milica Gasic is the Head of Dialogue Systems Group, a Lecturer in Spoken Dialogue Systems at the Cambridge University Engineering Department and a Fellow of Murray Edwards College. She holds a BS in Computer Science and Mathematics from the University of Belgrade (2006), a MPhil in Computer Speech, Text and Internet Technology (2007) and a PhD in Engineering from the University of Cambridge (2011). The topic of her PhD was statistical dialogue modelling and she was awarded an EPSRC PhD plus award for her dissertation. She has published around 50 journal articles and peer reviewed conference papers and received a number of best paper awards. She is an elected committee member of IEEE SLTC and an appointed board member of Sigdial.

Keynote talk - Deep reinforcement learning for dialogue policy optimization

In spoken dialogue systems, we aim to deploy artificial intelligence to build automated dialogue agents that can converse with humans. As part of this effort, we need to find ways to optimize the dialogue policy, i.e. we need to optimize a function that takes the current state of the dialogue as input and returns the response of the system. This is normally done via reinforcement learning. Deep reinforcement learning approaches have produced state-of-the-art results on games. In this talk I will discuss the necessary steps needed to deploy deep reinforcement learning for dialogue policy optimization. I will also discuss the necessity for common benchmarks and the efforts in the Dialogue Systems Group to provide these.