Speech is the most natural mode of communication and yet attempts to build systems which support robust habitable conversations between a human and a machine have so far had only limited success. A key reason is that current systems treat speech input as equivalent to a keyboard or mouse, and behaviour is controlled by predefined scripts that try to anticipate what the user will say and act accordingly. But speech recognisers make many errors and humans are not predictable; the result is systems which are difficult to design and fragile in use.

Statistical methods for spoken dialogue management takes a radically different view. It treats dialogue as the problem of inferring a user's intentions based on what is said. The dialogue is modelled as a probabilistic network and the input speech acts are observations that provide evidence for performing Bayesian inference. The result is a system which is much more robust to speech recognition errors and for which a dialogue strategy can be learned automatically using reinforcement learning. The thesis describes both the architecture, the algorithms needed for fast real-time inference over very large networks, model parameter estimation and policy optimisation.

This ground-breaking work will be of interest both to practitioners in spoken dialogue systems and to cognitive scientists interested in models of human behaviour.

Nominated by University of Cambridge as an outstanding Ph.D. thesis Describes both the architecture and the algorithms needed for fast real-time inference over very large networks, model parameter estimation and policy optimization Aimed at practitioners in spoken dialogue systems and to cognitive scientists interested in models of human behavior Includes supplementary material: sn.pub/extras

Autorentext

Blaise Thomson is a Research Fellow at St John's College in the University of Cambridge. He obtained a Bachelors degree in Pure Mathematics, Computer Science, Statistics and Actuarial Science at the University of Cape Town, South Africa, before completing an MPhil at the University of Cambridge in 2006 and a PhD in Statistical Dialogue Modelling in 2010. He has published around 35 peer-reviewed journal and conference papers, focusing largely on the topics of dialogue management, automatic speech recognition, speech synthesis, natural language understanding and collaborative filtering. In 2008 he was awarded the IEEE Student Spoken Language Processing award for his paper at the International Conference on Acoustics, Speech, and Signal Processing (ICASSP) and in 2010 he co-authored best papers at both the IEEE Spoken Language Technologies workshop and Interspeech. He was co-chair of the 2009 ACL Student Research Workshop and co-presented a tutorial on POMDP dialogue management at Interspeech 2009. In his spare time, he enjoys playing guitar and dancing and represented England at the 2010, 2011 and 2012 world formation Latin championships.

Inhalt
Dialogue system theory.- Maintaining state.- Maintaining state - optimizations.- Policy design.- Evaluation.- Parameter learning.