Effects of the user model on simulation-based learning of dialogue strategies

Over the past decade, a variety of user models have been proposed for user simulation-based reinforcement-learning of dialogue strategies. However, the strategies learned with these models are rarely evaluated in actual user trials and it remains unclear how the choice of user model affects the quality of the learned strategy. In particular, the degree to which strategies learned with a user model generalise to real user populations has not be investigated. This paper presents a series of experiments that qualitatively and quantitatively examine the effect of the user model on the learned strategy. Our results show that the performance and characteristics of the strategy are in fact highly dependent on the user model. Furthermore, a policy trained with a poor user model may appear to perform well when tested with the same model, but fail when tested with a more sophisticated user model. This raises significant doubts about the current practice of learning and evaluating strategies with the same user model. The paper further investigates a new technique for testing and comparing strategies directly on real human-machine dialogues, thereby avoiding any evaluation bias introduced by the user model. © 2005 IEEE.

Model based learning of sigma points in unscented Kalman filtering

The unscented Kalman filter (UKF) is a widely used method in control and time series applications. The UKF suffers from arbitrary parameters necessary for a step known as sigma point placement, causing it to perform poorly in nonlinear problems. We show how to treat sigma point placement in a UKF as a learning problem in a model based view. We demonstrate that learning to place the sigma points correctly from data can make sigma point collapse much less likely. Learning can result in a significant increase in predictive performance over default settings of the parameters in the UKF and other filters designed to avoid the problems of the UKF, such as the GP-ADF. At the same time, we maintain a lower computational complexity than the other methods. We call our method UKF-L. ©2010 IEEE.

Model based learning of sigma points in unscented Kalman filtering

The unscented Kalman filter (UKF) is a widely used method in control and time series applications. The UKF suffers from arbitrary parameters necessary for sigma point placement, potentially causing it to perform poorly in nonlinear problems. We show how to treat sigma point placement in a UKF as a learning problem in a model based view. We demonstrate that learning to place the sigma points correctly from data can make sigma point collapse much less likely. Learning can result in a significant increase in predictive performance over default settings of the parameters in the UKF and other filters designed to avoid the problems of the UKF, such as the GP-ADF. At the same time, we maintain a lower computational complexity than the other methods. We call our method UKF-L. © 2011 Elsevier B.V.

Hydrogen sensor based on silicon carbide (SiC) MOS capacitor

Silicon carbide (SiC) based MOS capacitor devices are used for gas sensing in high temperature and chemically reactive environments. A SiC MOS capacitor structure used as hydrogen sensor is defined and simulated. The effects of hydrogen concentration, temperature and interface traps on C-V characteristics were analysed. A comparison between structures with different oxide layer types (SiO2, TiO2 and ZnO) and thicknesses (50..10nm) was conducted. The TiO2 based structure has better performance than the SiO2 and ZnO structures. Also, the performance of the SiC MOS capacitor increases at thinner oxide layers. © 2012 IEEE.

A thermopile based SOI CMOS MEMS wall shear stress sensor

In this paper we present for the first time, a novel silicon on insulator (SOI) complementary metal oxide semiconductor (CMOS) MEMS thermal wall shear stress sensor based on a tungsten hot-film and three thermopiles. These devices have been fabricated using a commercial 1 μm SOI-CMOS process followed by a deep reactive ion etch (DRIE) back-etch step to create silicon oxide membranes under the hot-film for effective thermal isolation. The sensors show an excellent repeatability of electro-thermal characteristics and can be used to measure wall shear stress in both constant current anemometric as well as calorimetric modes. The sensors have been calibrated for wall shear stress measurement of air in the range of 0-0.48 Pa using a suction type, 2-D flow wind tunnel. The calibration results show that the sensors have a higher sensitivity (up to four times) in calorimetric mode compared to anemometric mode for wall shear stress lower than 0.3 Pa. © 2013 IEEE.

Smart chemical sensor application of ZnO nanowires grown on CMOS compatible SOI microheater platform

Smart chemical sensor based on CMOS(complementary metal-oxide- semiconductor) compatible SOI(silicon on insulator) microheater platform was realized by facilitating ZnO nanowires growth on the small membrane at the relatively low temperature. Our SOI microheater platform can be operated at the very low power consumption with novel metal oxide sensing materials, like ZnO or SnO2 nanostructured materials which demand relatively high sensing temperature. In addition, our sol-gel growth method of ZnO nanowires on the SOI membrane was found to be very effective compared with ink-jetting or CVD growth techniques. These combined techniques give us the possibility of smart chemical sensor technology easily merged into the conventional semiconductor IC application. The physical properties of ZnO nanowire network grown by the solution-based method and its chemical sensing property also were reported in this paper.