The design and implementation of dialogue control in voice operated database inquiry systems

This paper describes work performed as part of the U.K. Alvey sponsored Voice Operated Database Inquiry System (VODIS) project in the area of intelligent dialogue control. The principal aims of the work were to develop a habitable interface for the untrained user; to investigate the degree to which dialogue control can be used to compensate for deficiencies in recognition performance; and to examine the requirements on dialogue control for generating natural speech output. A data-driven methodology is described based on the use of frames in which dialogue topics are organized hierarchically. The concept of a dynamically adjustable scope is introduced to permit adaptation to recognizer performance and the use of historical and hierarchical contexts are described to facilitate the construction of contextually relevant output messages. © 1989.

Gaussian mixture modeling with Gaussian process latent variable models

Density modeling is notoriously difficult for high dimensional data. One approach to the problem is to search for a lower dimensional manifold which captures the main characteristics of the data. Recently, the Gaussian Process Latent Variable Model (GPLVM) has successfully been used to find low dimensional manifolds in a variety of complex data. The GPLVM consists of a set of points in a low dimensional latent space, and a stochastic map to the observed space. We show how it can be interpreted as a density model in the observed space. However, the GPLVM is not trained as a density model and therefore yields bad density estimates. We propose a new training strategy and obtain improved generalisation performance and better density estimates in comparative evaluations on several benchmark data sets. © 2010 Springer-Verlag.

Model structure and uncertainty for stochastic non-point source modelling applications

The uncertainty associated with a rainfall-runoff and non-point source loading (NPS) model can be attributed to both the parameterization and model structure. An interesting implication of the areal nature of NPS models is the direct relationship between model structure (i.e. sub-watershed size) and sample size for the parameterization of spatial data. The approach of this research is to find structural limitations in scale for the use of the conceptual NPS model, then examine the scales at which suitable stochastic depictions of key parameter sets can be generated. The overlapping regions are optimal (and possibly the only suitable regions) for conducting meaningful stochastic analysis with a given NPS model. Previous work has sought to find optimal scales for deterministic analysis (where, in fact, calibration can be adjusted to compensate for sub-optimal scale selection); however, analysis of stochastic suitability and uncertainty associated with both the conceptual model and the parameter set, as presented here, is novel; as is the strategy of delineating a watershed based on the uncertainty distribution. The results of this paper demonstrate a narrow range of acceptable model structure for stochastic analysis in the chosen NPS model. In the case examined, the uncertainties associated with parameterization and parameter sensitivity are shown to be outweighed in significance by those resulting from structural and conceptual decisions. © 2011 Copyright IAHS Press.

POMDP-based statistical spoken dialog systems: A review

Statistical dialog systems (SDSs) are motivated by the need for a data-driven framework that reduces the cost of laboriously handcrafting complex dialog managers and that provides robustness against the errors created by speech recognizers operating in noisy environments. By including an explicit Bayesian model of uncertainty and by optimizing the policy via a reward-driven process, partially observable Markov decision processes (POMDPs) provide such a framework. However, exact model representation and optimization is computationally intractable. Hence, the practical application of POMDP-based systems requires efficient algorithms and carefully constructed approximations. This review article provides an overview of the current state of the art in the development of POMDP-based spoken dialog systems. © 1963-2012 IEEE.

Articulating current service development practices: a qualitative analysis of eleven mental health projects.

BACKGROUND: The utilisation of good design practices in the development of complex health services is essential to improving quality. Healthcare organisations, however, are often seriously out of step with modern design thinking and practice. As a starting point to encourage the uptake of good design practices, it is important to understand the context of their intended use. This study aims to do that by articulating current health service development practices. METHODS: Eleven service development projects carried out in a large mental health service were investigated through in-depth interviews with six operation managers. The critical decision method in conjunction with diagrammatic elicitation was used to capture descriptions of these projects. Stage-gate design models were then formed to visually articulate, classify and characterise different service development practices. RESULTS: Projects were grouped into three categories according to design process patterns: new service introduction and service integration; service improvement; service closure. Three common design stages: problem exploration, idea generation and solution evaluation - were then compared across the design process patterns. Consistent across projects were a top-down, policy-driven approach to exploration, underexploited idea generation and implementation-based evaluation. CONCLUSIONS: This study provides insight into where and how good design practices can contribute to the improvement of current service development practices. Specifically, the following suggestions for future service development practices are made: genuine user needs analysis for exploration; divergent thinking and innovative culture for idea generation; and fail-safe evaluation prior to implementation. Better training for managers through partnership working with design experts and researchers could be beneficial.

Improving data quality in DSM modelling: A structural comparison approach

The Dependency Structure Matrix (DSM) has proved to be a useful tool for system structure elicitation and analysis. However, as with any modelling approach, the insights gained from analysis are limited by the quality and correctness of input information. This paper explores how the quality of data in a DSM can be enhanced by elicitation methods which include comparison of information acquired from different perspectives and levels of abstraction. The approach is based on comparison of dependencies according to their structural importance. It is illustrated through two case studies: creation of a DSM showing the spatial connections between elements in a product, and a DSM capturing information flows in an organisation. We conclude that considering structural criteria can lead to improved data quality in DSM models, although further research is required to fully explore the benefits and limitations of our proposed approach.

Optimal data fitting on lie groups: A coset approach

This work considers the problem of fitting data on a Lie group by a coset of a compact subgroup. This problem can be seen as an extension of the problem of fitting affine subspaces in n to data which can be solved using principal component analysis. We show how the fitting problem can be reduced for biinvariant distances to a generalized mean calculation on an homogeneous space. For biinvariant Riemannian distances we provide an algorithm based on the Karcher mean gradient algorithm. We illustrate our approach by some examples on SO(n). © 2010 Springer -Verlag Berlin Heidelberg.