題　目：Enhancing Response Generation Using Chat Flow Identification
講   者：Dr. Wen-Ling Hsu (Data Science and AI Research, AT&T Labs)
時   間：2019.1.9(三)13:20 - 15:20
地   點：本校工程三館345室
摘   要：Conversational artificial intelligence (AI) has been widely applied to different products and services in the past few years. As a common engagement channel between customers and businesses, live chat can benefit substantially from conversational AI. Currently, businesses spend a tremendous amount of time and money to train and educate chat agents. The training and educating processes include providing agents with chat flows, FAQs, and answer templates to guide chat agents answering customer questions efficiently and properly. However, creating such training materials for chat agents is time-consuming. In addition, the materials need to be constantly updated as new services and products are deployed to customers. In this work, we study the problem of extracting common chat flow from online customer care chats to enhance question response generation, with the goal of using them to facilitate chat agent training.

講者簡介：Dr. Wen-Ling Hsu is a Lead Inventive Scientist in Data Science and AI Research, AT&T Labs. Her current interests include machine learning, text mining, analytical computing and intelligent systems, focusing on analytics and innovation in customer care. She holds a PhD in Information Systems from Purdue University, and a BSE in Computer Science from National Chiao Tung University in Taiwan. Prior to joining AT&T, Wen-Ling was a Principal Research Scientist in the Department of Distributed AI and Machine Learning at Siemens Corporate Research, an Assistant Professor at the Heinz School at Carnegie Mellon University, and an Adjunct Professor in the school of IEOR at Columbia University.
Research Interests: Data mining, text mining, analytical computing and intelligent systems. Applications include solving real problems in various areas of telecommunication systems such as customer care, network planning, call traffic analysis, risk management, and strategic planning.