Morphological decomposition in Arabic ASR systems

In recent years, the use of morphological decomposition strategies for Arabic Automatic Speech Recognition (ASR) has become increasingly popular. Systems trained on morphologically decomposed data are often used in combination with standard word-based approaches, and they have been found to yield consistent performance improvements. The present article contributes to this ongoing research endeavour by exploring the use of the 'Morphological Analysis and Disambiguation for Arabic' (MADA) tools for this purpose. System integration issues concerning language modelling and dictionary construction, as well as the estimation of pronunciation probabilities, are discussed. In particular, a novel solution for morpheme-to-word conversion is presented which makes use of an N-gram Statistical Machine Translation (SMT) approach. System performance is investigated within a multi-pass adaptation/combination framework. All the systems described in this paper are evaluated on an Arabic large vocabulary speech recognition task which includes both Broadcast News and Broadcast Conversation test data. It is shown that the use of MADA-based systems, in combination with word-based systems, can reduce the Word Error Rates by up to 8.1 relative. © 2012 Elsevier Ltd. All rights reserved.

A canonical normal SBLI flow relevant to external compression inlets

The normal shock wave / boundary layer interaction (normal SBLI) is important to the operation and performance of a supersonic inlet, and the normal SBLI is particularly prominent in external compression inlets. To improve our understanding of such interactions, it is helpful to make use of fundamental flows which capture the main elements of inlets, without resorting to the level of complexity and system integration associated with full-geometry inlets. In this paper, several fundamental fiow-fleld configurations have been considered as possible test cases to represent the normal SBLI aspects found in typical external compression inlets, and it was found that the spillage-diffuser more closely retains the basic flow features of an external compression inlet than the other configurations. In particular, this flow-fleld allows the normal shock Mach number as well as the amount and rate of subsonic diffusion to be all held approximately constant mid independent of the application of flow control. In addition, a survey of several external compression inlets was conducted to quantify the flow and geometric parameters of the spillage-diffuser relevant to actual inlets. The results indicated that such a flow may be especially relevant if the terminal Mach number is about 1.3 to 1.4, the confinement parameter is around 10%, the width around twice or three times the height, and with the area expansion just downstream of the shock on the conservative side of the stall limit for incompressible diffusers. © 2013 by the American Institute of Aeronautics and Astronautics, Inc. All rights reserved.

Canonical normal shock wave/boundary-layer interaction flows relevant to external compression inlets

The normal shock wave/boundary-layer interaction is important to the operation and performance of a supersonic inlet, and the normal shock wave/boundary-layer interaction is particularly prominent in external compression inlets. To improve understanding of such interactions, it is helpful to make use of fundamental flows that capture the main elements of inlets, without resorting to the level of complexity and system integration associated with full-geometry inlets. In this paper, several fundamental flowfield configurations have been considered as possible test cases to represent the normal shock wave/boundary-layer interaction aspects found in typical external compression inlets, and it was found that the spillage diffuser more closely retains the basic flow features of an external compression inlet than the other configurations. In particular, this flowfield allows the normal shock Mach number as well as the amount and rate of subsonic diffusion to all be held approximately constant and independent of the application of flow control. In addition, a survey of several external compression inlets was conducted to quantify the flow and geometric parameters of the spillage diffuser relevant to actual inlets. The results indicated that such a flow may be especially relevant if the terminal Mach number is about 1.3 to 1.4, the confinement parameter is around 10%, and the width is around twice or three times the height. In addition, the area expansion downstream of the shock should be limited to the conservative side of incipient stall based on incompressible diffusers. Copyright © 2013 by the authors.

Investigating carbon materials for use as the electron emission source in a parallel electron-beam lithography system

A microelectronic parallel electron-beam lithography system using an array of field emitting microguns is currently being developed. This paper investigates the suitability of various carbon based materials for the electron source in this device, namely tetrahedrally bonded amorphous carbon (ta-C), nanoclustered carbon and carbon nanotubes. Ta-C was most easily integrated into a gated field emitter structure and various methods, such as plasma and heavy ion irradiation, were used to induce emission sites in the ta-C. However, the creation of such emission sites at desired locations appeared to be difficult/random in nature and thus the material was unsuitable for this application. In contrast, nanoclustered carbon material readily field emits with a high site density but the by-products from the deposition process create integration issues when using the material in a microelectronic gated structure. Carbon nanotubes are currently the most promising candidate for use as the emission source. We have developed a high yield and clean (amorphous carbon by-product free) PECVD process to deposit single free standing nanotubes at desired locations with exceptional uniformity in terms of nanotube height and diameter. Field emission from an array of nanotubes was also obtained. © 2001 Elsevier Science B.V.

Materials and devices toward three-dimensional integration

It has been shown that the apparent benefits of a two-layer stacked SOI system, i.e. packing density and speed improvements, are less than could be expected in the context of a VLSI requirement [1]. In this project the stacked SOI system has been identified as having major application in the realization of integrated, mixed technology systems. Zone-melting-recrystallization (ZMR) with lasers and electron beams have been used to produce device quality SOI material and a small test-bed circuit has been designed as a demonstration of the feasibility of this approach. © 1988.