This work continues in development of the recently proposed Bottle-Neck features for ASR. A five-layers MLP used in bottleneck feature extraction allows to obtai arbitrary feature size without dimensionality reduction by transforms, independently on the MLP training targets. The MLP topology – number and sizes of layers, suitable training targets, the impact of output feature transforms, the ne...

BACKGROUND In chronic heart failure, myocardial expression of the inducible isoform of nitric oxide (NO) synthase (NOS2) is enhanced, leading to a sustained production of NO. We postulated that NO modulates expression of genes in cardiac myocytes that may be functionally important in the context of cardiac hypertrophy and failure. METHODS AND RESULTS As revealed by cDNA expression array analy...

In this paper, we propose a system for recognition handwritten characters Tifinagh, with the use of neural networks (the multi layer perceptron MLP), the hidden Markov model (HMM), the hybrid Model MLP/HMM and a feature extraction method based on mathematical morphology, this method is tested on the database of handwritten isolated characters Tifinagh size consistent (1800 images in learning an...

The adenovirus type 2 major late upstream element factor (UEF) recognizes two similar elements that lie between the major late promoter (MLP) and IVa2 promoter cap sites (the previously characterized MLP-UE from nucleotides -49 to -67 and the IVa2-UE from nucleotides -98 to -122). DNase I footprinting and gel retention assays showed that the UEF has a lower affinity for the IVa2-UE than for the...

MLP (muscle LIM protein)-deficient mice count among the first mouse models for dilated cardiomyopathy (DCM), yet the exact role of MLP in cardiac signalling processes is still enigmatic. Elevated PKCα signalling activity is known to be an important contributor to heart failure. Here we show that MLP directly inhibits the activity of PKCα. In end-stage DCM, PKCα is concentrated at the intercalat...

In this paper we present our latest investigation on initialization schemes for Multilayer Perceptron (MLP) training using multilingual data. We show that the overall performance of a MLP network improves significantly by initializing it with a multilingual MLP. We propose a new strategy called “open target language” MLP to train more flexible models for language adaptation, which is particular...

An adaptive maximum-likelihood procedure (MLP), typically used for threshold measurements in psychoacoustics, was adopted for the measurement of speech intelligibility, with the use of the modified rhyme test (MRT) [1]. The advantage of the MLP over the fixed step up-down adaptive procedures usually employed in adaptive speech intelligibility tests is that the MLP imposes no restrictions on the...