Unsupervised training for mandarin broadcast news and conversation transcription

A significant cost in obtaining acoustic training data is the generation of accurate transcriptions. For some sources close-caption data is available. This allows the use of lightly-supervised training techniques. However, for some sources and languages close-caption is not available. In these cases unsupervised training techniques must be used. This paper examines the use of unsupervised techniques for discriminative training. In unsupervised training automatic transcriptions from a recognition system are used for training. As these transcriptions may be errorful data selection may be useful. Two forms of selection are described, one to remove non-target language shows, the other to remove segments with low confidence. Experiments were carried out on a Mandarin transcriptions task. Two types of test data were considered, Broadcast News (BN) and Broadcast Conversations (BC). Results show that the gains from unsupervised discriminative training are highly dependent on the accuracy of the automatic transcriptions. © 2007 IEEE.

Laser micromachining of thin films for optoelectronic devices and packages

Focused laser micromachining in an optical microscope system is used to prototype packages for optoelectronic devices and to investigate new materials with potential applications in packaging. Micromachined thin films are proposed as mechanical components to locate fibres and other optical and electrical components on opto-assemblies. This paper reports prototype structures which are micromachined in silicon carbide to produce beams 5 μm thick by (i) laser cutting a track in a SiC coated Si wafer, (ii) undercutting by anisotropic silicon etching using KOH in water, and (iii) trimming if necessary with the laser system. This approach has the advantage of fast turn around and proof of concept. Mechanical test data are obtained from the prototype SiC beam package structures by testing with a stylus profilometer. The Youngs modulus obtained for chemical vapour deposited silicon carbide is 360 +/- 50 GPa indicating that it is a promising material for packaging applications.