Language model combination and adaptation using weighted finite state transducers

In speech recognition systems language model (LMs) are often constructed by training and combining multiple n-gram models. They can be either used to represent different genres or tasks found in diverse text sources, or capture stochastic properties of different linguistic symbol sequences, for example, syllables and words. Unsupervised LM adaptation may also be used to further improve robustness to varying styles or tasks. When using these techniques, extensive software changes are often required. In this paper an alternative and more general approach based on weighted finite state transducers (WFSTs) is investigated for LM combination and adaptation. As it is entirely based on well-defined WFST operations, minimum change to decoding tools is needed. A wide range of LM combination configurations can be flexibly supported. An efficient on-the-fly WFST decoding algorithm is also proposed. Significant error rate gains of 7.3% relative were obtained on a state-of-the-art broadcast audio recognition task using a history dependently adapted multi-level LM modelling both syllable and word sequences. ©2010 IEEE.

Simple and practical construction of high-performance, low-cost rogowski transducers and accompanying circuitry for research applications

Rogowski transducers have become an increasingly popular method of measuring current within prototyping applications and power electronics equipment due to their significant advantages compared to an equivalent current transformer. This paper presents a simple and practical construction technique of high-performance, low-cost Rogowski transducers and accompanying circuitry. Experimental tests were carried out to show the validity of the proposed construction technique. © 2005 IEEE.

Hygroscopic biomimetic transducers made from CNT-hydrogel composites

Plants as well as other biological organisms achieve directed movements by fibres that constraint and direct the isotropic expansion of a matrix material. In order to mimic these actuators, complex arrangements of rigid fibres must be achieved, which is challenging, especially at small scales. In this paper, a new method to organize carbon nanotubes (CNTs) into complex shapes is employed to create a framework for hydrogel infiltration. These CNT frameworks can be realized as iris, needle and bridge architectures, and after hydrogel infiltration, they show directed actuation in response to water uptake. Finally, we show how the latter can be employed as a novel hygroscopic sensor. © 2011 IEEE.