CS Lectures

Prof. Verena Rieser

Prof. Verena Rieser / A Short History of Data-driven Dialogue Systems in 5 Acts: Where Do We Go from Here?

Heriot-Watt University, Edinburgh


In this talk I will provide a short review of the last 20 years of data-driven development of conversational AI (aka dialogue systems) through the lens of 5 major systems and initiatives I was involved in: I will start by reviewing the beginnings of customer-driven rule-based systems, corpus-based studies, optimising systems using Reinforcement Learning and finally neural end-to-end systems -- including open-domain and task-based systems.

Throughout my talk, I will focus on the sub-task of response generation, for which I will highlight lessons learnt and ongoing challenges.

In particular, I will argue that neural models lack control over what is generated, are not customer-driven, nor able to adapt in real-time and don't show system system initiative. I will then provide a short overview of how my team aims to address these challenges, including visual- and symbolic grounding, common sense reasoning, evaluation practises and persona design.

About the Speaker

Verena Rieser is a professor at Heriot-Watt University in Edinburgh, where she leads research on Natural Language Generation and Spoken Dialogue Systems. She is also a co-founder of the Conversational AI company Alana. She received her PhD in 2008 from Saarland University and joined the University of Edinburgh for a postdoctoral position, before taking up a faculty role at Heriot-Watt in 2011.

Verena was recently awarded a Leverhulme Trust Senior Research Fellowship by the Royal Society and she is PI of several publicly funded research projects and industry awards. Verena’s team is a double finalist of the prestigious Amazon Alexa Challenge. Her research has featured in the BBC's documentary The Joy of AI, BBC’s Tomorrow's World and in national and international news. Verena’s current research interests include Ethics for open-domain conversational systems, Data-to-Text and Text-to-Text generation, as well as Multimodal Dialogue.

ITU Host

Barbara Plank