Neural networks used for speech recognition

Gevaert, Wouter; Tsenov, Georgi; Mladenov, Valeri

doi:10.2298/jac1001001g

Cited by 71 publications

(34 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…For arabic automatic speech recognition, the recognition of phonemes constitutes an important step in continuous speech analysis. most research proceeds by extracting isolated phonemes or small phonetic segments (el-obaid, al-Nassiri, and maaly 2006;awais 2003;gevaert, tsenov, and mladenov 2010;al-manie, alkanhal, and al-ghamdi 2009) for analysis of longer speech signals (abushariah et al 2010) and broadcast news (al-manie, alkanhal, and al-ghamdi 2009), using several techniques, such as aNN (essa, tolba, and elmougy 2008), fuzzy hmm (shenouda, Zaki, and goneid 2006), fuzzy logic, concurrent self-organizing maps (sehgal, gondal, andDooley 2004), andhmm (satori, harti, andChenfour 2007;Bourouba et al 2010;Biadsy, moreno, and Jansche 2012). spoken in the middle east and North africa, arabic has different dialects.…”

Section: Tree Representation For Arabic Phonemesmentioning

confidence: 99%

Cortical Algorithms

Awad¹,

Khanna²

2015

Efficient Learning Machines

View full text Add to dashboard Cite

Computational models inspired by the structural and functional properties of the human brain have seen impressive gains since the mid-1980s, owing to significant discoveries in neuroscience and advancements in computing technology. Among these models, cortical algorithms (CAs) have emerged as a biologically inspired approach, modeled after the human visual cortex, which stores sequences of patterns in an invariant form and recalls those patterns autoassociatively. This chapter details the structure and mathematical formulation of CA then presents a case study of CA generalization accuracy in identifying isolated Arabic speech using an entropy-based weight update. Cortical Algorithm PrimerInitially developed by Edelman and Mountcastle (1978), and inspired by the visual human cortex, CAs are positioned to be superior to the early generations of artificial neural networks (ANNs), which do not use temporal and spatial relationships in data for building machine learning models.The CA model consists of a multilayered network, with the cortical column as the basic structure. The network is trained in a two-stage manner: the first learning stage is unsupervised and trains the columns to identify independent features from the patterns occurring; the second stage relies on supervised feedback learning to create invariant representations. Cortical Algorithm StructureThe human brain is a six-layered structure consisting of a very large number of neurons strongly connected via feedforward and feedback connections. An important property of the neocortex is its structural and functional uniformity: all units in the network seem similar, and they perform the same basic operation. Like this brain architecture, CA architecture has minicolumns of varying thickness (Edelman and Mountcastle 1978). A minicolumn is a group of neurons that share the same receptive field: neurons belonging to a minicolumn are associated with the same sensory input region. The minicolumn is the basic structure in a cortical network, in contrast to neurons in a classical ANN. An association of minicolumns is called a hypercolumn or layer

show abstract

Section: Tree Representation For Arabic Phonemesmentioning

confidence: 99%

Cortical Algorithms

Awad¹,

Khanna²

2015

Efficient Learning Machines

View full text Add to dashboard Cite

show abstract

“…Feed-forward multilayer networks with non-linear node functions can overcome these limitations, and can be used for many applications. Hence a more powerful supervised learning mechanism called back-propagation is used for multi-class, multi-level discrimination [3], [5].…”

Section: (F) Adalinesmentioning

confidence: 99%

“…The main benefit of this work would be its contribution towards employing the neural network-based techniques for solving common but difficult problem of pattern recognition, particularly in ASR. There are three major types of pattern recognition techniques namely dynamic time warping (DTW), Hidden Markov model (HMM) and artificial neural networks (ANN) [1], [5]. This paper is organized as follows.…”

Section: Iintroductionmentioning

confidence: 99%

MATLAB Based Back-Propagation Neural Network for Automatic Speech Recognition

JOSHI¹,

CHEERAN²

2014

IJAREEIE

View full text Add to dashboard Cite

Speech interface to computer is the next big step that the technology needs to take for general users. Automatic speech recognition (ASR) will play an important role in taking technology to the people. There are numerous applications of speech recognition such as direct voice input in aircraft, data entry, speech-to-text processing, voice user interfaces such as voice dialling. ASR system can be divided into two different parts, namely feature extraction and feature recognition. In this paper we present MATLAB based feature recognition using backpropagation neural network for ASR. The objective of this research is to explore how neural networks can be employed to recognize isolated-word speech as an alternative to the traditional methodologies. The general techniques developed here can be further extended to other applications such as sonar target recognition, missile tracking and classification of underwater acoustic signals. Back-propagation neural network algorithm uses input training samples and their respective desired output values to learn to recognize specific patterns, by modifying the activation values of its nodes and weights of the links connecting its nodes. Such a trained network is later used for feature recognition in ASR systems.

show abstract

“…In the approach of Artificial Intelligence to speech recognition various sources of knowledge [2] are required to be set up. Thus, artificial intelligence is classified in two processes broadly: a) Automatic knowledge acquisitions learning and b) Adaptation.…”

Section: Fig 8: Simplified View Of An Artificial Neural Networkmentioning

confidence: 99%

Reviewing Human-Machine Interaction through Speech Recognition approaches and Analyzing an approach for Designing an Efficient System

Kantlavania¹,

Sharma²,

Sharma³

2012

IJCA

View full text Add to dashboard Cite

Speech is most natural way of interaction for human. It has broad applications in the human-machine and humancomputer interaction. This paper reviews the literature and the technological aspects of human-machine interaction through various speech recognition approaches. It also discusses the various techniques used in each step of a speech recognition process and attempts to analyze an approach for designing an efficient system for speech recognition. It also discusses that how this system works and its application in various areas.

show abstract

Neural networks used for speech recognition

Cited by 71 publications

References 7 publications

Cortical Algorithms

Cortical Algorithms

MATLAB Based Back-Propagation Neural Network for Automatic Speech Recognition

Reviewing Human-Machine Interaction through Speech Recognition approaches and Analyzing an approach for Designing an Efficient System

Contact Info

Product

Resources

About