Policy optimisation of POMDP-based dialogue systems without state space compression

The partially observable Markov decision process (POMDP) has been proposed as a dialogue model that enables automatic improvement of the dialogue policy and robustness to speech understanding errors. It requires, however, a large number of dialogues to train the dialogue policy. Gaussian processes (GP) have recently been applied to POMDP dialogue management optimisation showing an ability to substantially increase the speed of learning. Here, we investigate this further using the Bayesian Update of Dialogue State dialogue manager. We show that it is possible to apply Gaussian processes directly to the belief state, removing the need for a parametric policy representation. In addition, the resulting policy learns significantly faster while maintaining operational performance. © 2012 IEEE.

Designing and testing an advanced pneumatic braking system for heavy vehicles

Heavy goods vehicles exhibit poor braking performance in emergency situations when compared to other vehicles. Part of the problem is caused by sluggish pneumatic brake actuators, which limit the control bandwidth of their antilock braking systems. In addition, heuristic control algorithms are used that do not achieve the maximum braking force throughout the stop. In this article, a novel braking system is introduced for pneumatically braked heavy goods vehicles. The conventional brake actuators are improved by placing high-bandwidth, binary-actuated valves directly on the brake chambers. A made-for-purpose valve is described. It achieves a switching delay of 3-4 ms in tests, which is an order of magnitude faster than solenoids in conventional anti-lock braking systems. The heuristic braking control algorithms are replaced with a wheel slip regulator based on sliding mode control. The combined actuator and slip controller are shown to reduce stopping distances on smooth and rough, high friction (μ = 0.9) surfaces by 10% and 27% respectively in hardware-in-the-loop tests compared with conventional ABS. On smooth and rough, low friction (μ = 0.2) surfaces, stopping distances are reduced by 23% and 25%, respectively. Moreover, the overall air reservoir size required on a heavy goods vehicle is governed by its air usage during an anti-lock braking stop on a low friction, smooth surface. The 37% reduction in air usage observed in hardware-in-the-loop tests on this surface therefore represents the potential reduction in reservoir size that could be achieved by the new system. © 2012 IMechE.

Thorium energy futures

The potential for thorium as an alternative or supplement to uranium in fission power generation has long been recognised, and several reactors, of various types, have already operated using thorium-based fuels. Accelerator Driven Subcritical (ADS) systems have benefits and drawbacks when compared to conventional critical thorium reactors, for both solid and molten salt fuels. None of the four options - liquid or solid, with or without an accelerator - can yet be rated as better or worse than the other three, given today's knowledge. We outline the research that will be necessary to lead to an informed choice. Copyright © 2012 by IEEE.

Continuous asr for flexible incremental dialogue

Spoken dialogue systems provide a convenient way for users to interact with a machine using only speech. However, they often rely on a rigid turn taking regime in which a voice activity detection (VAD) module is used to determine when the user is speaking and decide when is an appropriate time for the system to respond. This paper investigates replacing the VAD and discrete utterance recogniser of a conventional turn-taking system with a continuously operating recogniser that is always listening, and using the recogniser 1-best path to guide turn taking. In this way, a flexible framework for incremental dialogue management is possible. Experimental results show that it is possible to remove the VAD component and successfully use the recogniser best path to identify user speech, with more robustness to noise, potentially smaller latency times, and a reduction in overall recognition error rate compared to using the conventional approach. © 2013 IEEE.

On-line policy optimisation of Bayesian spoken dialogue systems via human interaction

A partially observable Markov decision process has been proposed as a dialogue model that enables robustness to speech recognition errors and automatic policy optimisation using reinforcement learning (RL). However, conventional RL algorithms require a very large number of dialogues, necessitating a user simulator. Recently, Gaussian processes have been shown to substantially speed up the optimisation, making it possible to learn directly from interaction with human users. However, early studies have been limited to very low dimensional spaces and the learning has exhibited convergence problems. Here we investigate learning from human interaction using the Bayesian Update of Dialogue State system. This dynamic Bayesian network based system has an optimisation space covering more than one hundred features, allowing a wide range of behaviours to be learned. Using an improved policy model and a more robust reward function, we show that stable learning can be achieved that significantly outperforms a simulator trained policy. © 2013 IEEE.

Free space adaptive optical interconnect, using a ferroelectric liquid crystal SLM for beam steering

A free-space, board-to-board, adaptive optical interconnect demonstrator has been developed. Binary phase gratings displayed on a Ferroelectric Liquid Crystal Spatial Light Modulator are used to maintain data transfer at 1.25Gbps, given varying optical misalignment.© 2005 Optical Society of America.