The modified Kanerva model for automatic speech recognition

A parallel processing network derived from Kanerva's associative memory theory Kanerva 1984 is shown to be able to train rapidly on connected speech data and recognize further speech data with a label error rate of 0·68%. This modified Kanerva model can be trained substantially faster than other networks with comparable pattern discrimination properties. Kanerva presented his theory of a self-propagating search in 1984, and showed theoretically that large-scale versions of his model would have powerful pattern matching properties. This paper describes how the design for the modified Kanerva model is derived from Kanerva's original theory. Several designs are tested to discover which form may be implemented fastest while still maintaining versatile recognition performance. A method is developed to deal with the time varying nature of the speech signal by recognizing static patterns together with a fixed quantity of contextual information. In order to recognize speech features in different contexts it is necessary for a network to be able to model disjoint pattern classes. This type of modelling cannot be performed by a single layer of links. Network research was once held back by the inability of single-layer networks to solve this sort of problem, and the lack of a training algorithm for multi-layer networks. Rumelhart, Hinton & Williams 1985 provided one solution by demonstrating the "back propagation" training algorithm for multi-layer networks. A second alternative is used in the modified Kanerva model. A non-linear fixed transformation maps the pattern space into a space of higher dimensionality in which the speech features are linearly separable. A single-layer network may then be used to perform the recognition. The advantage of this solution over the other using multi-layer networks lies in the greater power and speed of the single-layer network training algorithm. © 1989.

Environmental and social impact monitoring of the Bujagali Hydropower Project (BHPP), Uganda Fisheries Component: monthly patterns in physico-chemical conditions and fish catch in the immediate vicinity of the Bujagali Hydropower construction site, 29th September - 28th October 2007

Following the commencement of construction works of a 250 MW hydropower plant at Dumbbell Island in the Upper Victoria Nile in September 2007, BEL requested NaFIRRI to conduct continuous monitoring of fish catches at two transects i.e. the immediate upstream transect of the project site (Kalange-Makwanzi) and the immediate downstream .transect (Buyala-Kikubamutwe). The routine monitoring surveys were designed to be conducted twice a week at each of the tWo transects. It was anticipated that major immediate impacts were to occur during construction, and these needed to be known by BEL as part of a mitigation strategy. For example, the construction of it cofferdam could be accompanied by rapid changes in water quality and quantity downstream of the construction. These changes in turn could affect the fish catch and would probably be missed by the quarterly monitoring already in place. Therefore, a major cbjective of the more regular and rapid monitoring was to discern immediate impacts of construction activities by focusing on selected water quality parameters (total suspended solids, water conductivity, temperature, dissolved oxygen and pH) and fish catch characteristics (total catch, catch rates and value of the catch)

Social and economic implications of the fishery distribution patterns on Lake Victoria, Uganda

The fisheries of Lake Victoria have undergonea dramatic transformation during the last two decades. From being a locally based fishery with little intervention and capital investment from outside,the present fishery is dominated by national and international capital penetrating the industry. It is explosion in the growth of nile perch and the strong demand devloped for this fishin the global markets, which have transformed the fisheries of the Lake victoria. This report presents the results of a survey carried out between October 2001 and February 2002 about the fishery distribution patterns and their impacts on fisher communities of Lake Victoria. The fisheries distribution pattern of the lake is described as well as the flows and benefits from the fisheries resource and the resource constraints and sustainability options. A major part of the paper discusses some of the socio-economic impacts of the rapid changes that are responsible for the present fisheries. It particularly focuses on the effect of the Nile perch boom, its globalization and the development of the fish industry in Uganda, on food security and employment for the local population.

Thin film shape memory alloys for optical sensing applications

Based on shape memory effect of the sputtered thin film shape memory alloys, different types of micromirror structures were designed and fabricated for optical sensing application. Using surface micromachining, TiNi membrane mirror structure has been fabricated, which can be actuated based on intrinsic two-way shape memory effect of the free-standing TiNi film. Using bulk micromachining, TiNi/Si and TiNi/Si 3N 4microcantilever mirror structures were fabricated. © 2007 IOP Publishing Ltd.

Effects of morphine and its withdrawal on Y-maze spatial recognition memory in mice

Effects of morphine on acquisition and retrieval of memory have been proven in the avoidance paradigms. In present study, we used a two-trial recognition Y-maze to test the effects of acute morphine and morphine withdrawal on spatial recognition memory. T

Recent origin of a hominoid-specific splice form of neuropsin, a gene involved in learning and memory

Neuropsin is a secreted-type serine protease involved in learning and memory. The type II splice form of neuropsin is abundantly expressed in the human brain but not in the mouse brain. We sequenced the type II-spliced region of neuropsin gene in humans and representative nonhuman primate species. Our comparative sequence analysis showed that only the hominoid species (humans and apes) have the intact open reading frame of the type II splice form, indicating that the type II neuropsin originated recently in the primate lineage about 18 MYA. Expression analysis using RT-PCR detected abundant expression of the type II form in the frontal lobe of the adult human brain, but no expression was detected in the brains of lesser apes and Old World monkeys, indicating that the type II form of neuropsin only became functional in recent time, and it might contribute to the progressive change of cognitive abilities during primate evolution.

A human-specific mutation leads to the origin of a novel splice form of neuropsin (KLK8), a gene involved in learning and memory

Neuropsin (kallikrein 8, ELKS) is a secreted-type serine protease preferentially expressed in the central nervous system and involved in learning and memory. Its splicing pattern is different in human and mouse, with the longer form (type II) only express

Dentin phosphoprotein frameshift mutations in hereditary dentin disorders and their variation patterns in normal human population

Background: Dentin phosphoprotein ( DPP) is the most abundant non-collagenous protein in dentin, which is highly phosphorylated and plays key roles in dentin biomineralisation. The aetiology of isolated hereditary dentin disorders in most affected familie

Functional characterization of the human-specific (type II) form of kallikrein 8, a gene involved in learning and memory

Kallikrein 8 (KLK8) is a serine protease functioning in the central nervous system, and essential in many aspects of neuronal activities. Sequence comparison and gene expression analysis among diverse primate species identified a human-specific splice for