Fast Likelihood Computation in Speech Recognition using Matrices

Acoustic modeling using mixtures of multivariate Gaussians is the prevalent approach for many speech processing problems. Computing likelihoods against a large set of Gaussians is required as a part of many speech processing systems and it is the computationally dominant phase for Large Vocabulary Continuous Speech Recognition (LVCSR) systems. We express the likelihood computation as a multiplication of matrices representing augmented feature vectors and Gaussian parameters. The computational gain of this approach over traditional methods is by exploiting the structure of these matrices and efficient implementation of their multiplication. In particular, we explore direct low-rank approximation of the Gaussian parameter matrix and indirect derivation of low-rank factors of the Gaussian parameter matrix by optimum approximation of the likelihood matrix. We show that both the methods lead to similar speedups but the latter leads to far lesser impact on the recognition accuracy. Experiments on 1,138 work vocabulary RM1 task and 6,224 word vocabulary TIMIT task using Sphinx 3.7 system show that, for a typical case the matrix multiplication based approach leads to overall speedup of 46 % on RM1 task and 115 % for TIMIT task. Our low-rank approximation methods provide a way for trading off recognition accuracy for a further increase in computational performance extending overall speedups up to 61 % for RM1 and 119 % for TIMIT for an increase of word error rate (WER) from 3.2 to 3.5 % for RM1 and for no increase in WER for TIMIT. We also express pairwise Euclidean distance computation phase in Dynamic Time Warping (DTW) in terms of matrix multiplication leading to saving of approximately of computational operations. In our experiments using efficient implementation of matrix multiplication, this leads to a speedup of 5.6 in computing the pairwise Euclidean distances and overall speedup up to 3.25 for DTW.

Computational framework to integrate the effect of antigen recognition on disease epidemiology outcome: multi-scale approach

Human Leukocyte Antigen (HLA) plays an important role, in presenting foreign pathogens to our immune system, there by eliciting early immune responses. HLA genes are highly polymorphic, giving rise to diverse antigen presentation capability. An important factor contributing to enormous variations in individual responses to diseases is differences in their HLA profiles. The heterogeneity in allele specific disease responses decides the overall disease epidemiological outcome. Here we propose an agent based computational framework, capable of incorporating allele specific information, to analyze disease epidemiology. This framework assumes a SIR model to estimate average disease transmission and recovery rate. Using epitope prediction tool, it performs sequence based epitope detection for a given the pathogenic genome and derives an allele specific disease susceptibility index depending on the epitope detection efficiency. The allele specific disease transmission rate, that follows, is then fed to the agent based epidemiology model, to analyze the disease outcome. The methodology presented here has a potential use in understanding how a disease spreads and effective measures to control the disease.

A mode-shape classification technique for robust speech rate estimation and syllable nuclei detection

Acoustic feature based speech (syllable) rate estimation and syllable nuclei detection are important problems in automatic speech recognition (ASR), computer assisted language learning (CALL) and fluency analysis. A typical solution for both the problems consists of two stages. The first stage involves computing a short-time feature contour such that most of the peaks of the contour correspond to the syllabic nuclei. In the second stage, the peaks corresponding to the syllable nuclei are detected. In this work, instead of the peak detection, we perform a mode-shape classification, which is formulated as a supervised binary classification problem - mode-shapes representing the syllabic nuclei as one class and remaining as the other. We use the temporal correlation and selected sub-band correlation (TCSSBC) feature contour and the mode-shapes in the TCSSBC feature contour are converted into a set of feature vectors using an interpolation technique. A support vector machine classifier is used for the classification. Experiments are performed separately using Switchboard, TIMIT and CTIMIT corpora in a five-fold cross validation setup. The average correlation coefficients for the syllable rate estimation turn out to be 0.6761, 0.6928 and 0.3604 for three corpora respectively, which outperform those obtained by the best of the existing peak detection techniques. Similarly, the average F-scores (syllable level) for the syllable nuclei detection are 0.8917, 0.8200 and 0.7637 for three corpora respectively. (C) 2016 Elsevier B.V. All rights reserved.

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.

Syllable language models for Mandarin speech recognition: exploiting character language models.

Mandarin Chinese is based on characters which are syllabic in nature and morphological in meaning. All spoken languages have syllabiotactic rules which govern the construction of syllables and their allowed sequences. These constraints are not as restrictive as those learned from word sequences, but they can provide additional useful linguistic information. Hence, it is possible to improve speech recognition performance by appropriately combining these two types of constraints. For the Chinese language considered in this paper, character level language models (LMs) can be used as a first level approximation to allowed syllable sequences. To test this idea, word and character level n-gram LMs were trained on 2.8 billion words (equivalent to 4.3 billion characters) of texts from a wide collection of text sources. Both hypothesis and model based combination techniques were investigated to combine word and character level LMs. Significant character error rate reductions up to 7.3% relative were obtained on a state-of-the-art Mandarin Chinese broadcast audio recognition task using an adapted history dependent multi-level LM that performs a log-linearly combination of character and word level LMs. This supports the hypothesis that character or syllable sequence models are useful for improving Mandarin speech recognition performance.

A Cancer Recognition Method Based on DNA Microarray

The accurate cancer classification is of great importance in clinical treatment. Recently, the DNA microarray technology provides a promising approach to the diagnosis and prognosis of cancer types. However, it has no perfect method for the multiclass classification problem. The difficulty lies in the fact that the data are of high dimensionality with small sample size. This paper proposed an automatic classification method of multiclass cancers based on Biomimetic pattern recognition (BPR). To the public GCM data set, the average correct classification rate reaches 80% under the condition that the correct rejection rate is 81%.

An Automatic Classification Method of Metabolic Cycle Based on Biomimetic Pattern Recognition

Correct classification of different metabolic cycle stages to identification cell cycle is significant in both human development and clinical diagnostics. However, it has no perfect method has been reached in classification of metabolic cycle yet. This paper exploringly puts forward an automatic classification method of metabolic cycle based on Biomimetic pattern recognition (BPR). As to the three phases of yeast metabolic cycle, the correct classification rate reaches 90%, 100% and 100% respectively.

Iris recognition algorithm based on point covering of high-dimensional space and neural network

In this paper, we constructed a Iris recognition algorithm based on point covering of high-dimensional space and Multi-weighted neuron of point covering of high-dimensional space, and proposed a new method for iris recognition based on point covering theory of high-dimensional space. In this method, irises are trained as "cognition" one class by one class, and it doesn't influence the original recognition knowledge for samples of the new added class. The results of experiments show the rejection rate is 98.9%, the correct cognition rate and the error rate are 95.71% and 3.5% respectively. The experimental results demonstrate that the rejection rate of test samples excluded in the training samples class is very high. It proves the proposed method for iris recognition is effective.

Manifold covering theory in biomimetic pattern recognition

We describe a new model which is based on the concept of cognizing theory. The method identifies subsets of the data which are embedded in arbitrary oriented lower dimensional space. We definite manifold covering in biomimetic pattern recognition, and study its property. Furthermore, we propose this manifold covering algorithm based on Biomimetic Pattern Recognition. At last, the experimental results for face recognition demonstrates that the correct rejection rate of the test samples excluded in the classes of training samples is very high and effective.

Research on continuous speech recognition with small vocabulary based on biomimetic pattern recognition

Based on biomimetic pattern recognition theory, we proposed a novel speaker-independent continuous speech keyword-spotting algorithm. Without endpoint detection and division, we can get the minimum distance curve between continuous speech samples and every keyword-training net through the dynamic searching to the feature-extracted continuous speech. Then we can count the number of the keywords by investigating the vale-value and the numbers of the vales in the curve. Experiments of small vocabulary continuous speech with various speaking rate have got good recognition results and proved the validity of the algorithm.

Cancer subtypes recognition and its gene expression profiles analysis based on geometrical learning

The accurate recognition of cancer subtypes is very significant in clinic. Especially, the DNA microarray gene expression technology is applied to diagnosing and recognizing cancer types. This paper proposed a method of that recognized cancer subtypes based on geometrical learning. Firstly, the cancer genes expression profiles data was pretreated and selected feature genes by conventional method; then the expression data of feature genes in the training samples was construed each convex hull in the high-dimensional space using training algorithm of geometrical learning, while the independent test set was tested by the recognition algorithm of geometrical learning. The method was applied to the human acute leukemia gene expression data. The accuracy rate reached to 100%. The experiments have proved its efficiency and feasibility.

A method of Biomimetic Pattern Recognition for face recognition

A new method of face recognition, based on Biomimetic Pattern Recognition and Multi-Weights Neuron Network, had been proposed. A model for face recognition that is based on Biomimetic Pattern Recognition had been discussed, and a new method of facial feature extraction also had been introduced. The results of experiments with BPR and K-Nearest Neighbor Rules showed that the method based on BPR can eliminate the error recognition of the samples of the types that not be trained, the correct rate is also enhanced.

An improved BP algorithm for pattern recognition

An improved BP algorithm for pattern recognition is proposed in this paper. By a function substitution for error measure, it resolves the inconsistency of BP algorithm for pattern recognition problems, i.e. the quadratic error is not sensitive to whether the training pattern is recognized correctly or not. Trained by this new method, the computer simulation result shows that the convergence speed is increased to treble and performance of the network is better than conventional BP algorithm with momentum and adaptive step size.

Single-walled carbon nanotubes binding to human telomeric i-motif DNA: significant acceleration of S1 nuclease cleavage rate

Single-walled carbon nanotubes (SWNTs) binding to human telomeric i-motif DNA can significantly accelerate S1 nuclease cleavage rate by increasing the enzyme turnover number.

Quantifying the kinetic paths of flexible biomolecular recognition

Biomolecular recognition often involves large conformational changes, sometimes even local unfolding. The identification of kinetic pathways has become a central issue in understanding the nature of binding. A new approach is proposed here to study the dynamics of this binding-folding process through the establishment of a path-integral framework on the underlying energy landscape. The dominant kinetic paths of binding and folding can be determined and quantified. The significant coupling between the binding and folding of biomolecules often exists in many important cellular processes. In this case, the corresponding kinetic paths of binding are shown to be intimately correlated with those of folding and the dynamics becomes quite cooperative. This implies that binding and folding happen concurrently. When the coupling between binding and folding is weak (strong), the kinetic process usually starts with significant folding (binding) first, with the binding (folding) later proceeding to the end. The kinetic rate can be obtained through the contributions from the dominant paths. The rate is shown to have a bell-shaped dependence on temperature in the concentration-saturated regime consistent with experiment. The changes of the kinetics that occur upon changing the parameters of the underlying binding-folding energy landscape are studied.