The Experts below are selected from a list of 198384 Experts worldwide ranked by ideXlab platform
Li Deng - One of the best experts on this subject based on the ideXlab platform.
-
tracking vocal tract resonances using a quantized nonlinear Function embedded in a temporal constraint
IEEE Transactions on Audio Speech and Language Processing, 2006Co-Authors: Li Deng, Alex Acero, I BazziAbstract:This paper presents a new technique for high-accuracy tracking of vocal-tract resonances (which coincide with formants for nonnasalized vowels) in natural speech. The technique is based on a discretized nonlinear Prediction Function, which is embedded in a temporal constraint on the quantized input values over adjacent time frames as the prior knowledge for their temporal behavior. The nonlinear Prediction is constructed, based on its analytical form derived in detail in this paper, as a parameter-free, discrete mapping Function that approximates the “forward” relationship from the resonance frequencies and bandwidths to the Linear Predictive Coding (LPC) cepstra of real speech. Discretization of the Function permits the “inversion” of the Function via a search operation. We further introduce the nonlinear-Prediction residual, characterized by a multivariate Gaussian vector with trainable mean vectors and covariance matrices, to account for the errors due to the Functional approximation. We develop and describe an expectation–maximization (EM)-based algorithm for training the parameters of the residual, and a dynamic programming-based algorithm for resonance tracking. Details of the algorithm implementation for computation speedup are provided. Experimental results are presented which demonstrate the effectiveness of our new paradigm for tracking vocal-tract resonances. In particular, we show the effectiveness of training the Prediction-residual parameters in obtaining high-accuracy resonance estimates, especially during consonantal closure.
-
a structured speech model with continuous hidden dynamics and Prediction residual training for tracking vocal tract resonances
International Conference on Acoustics Speech and Signal Processing, 2004Co-Authors: Li Deng, Hagai Attias, Alejandro AceroAbstract:A novel approach is developed for efficient and accurate tracking of vocal tract resonances, which are natural frequencies of the resonator from larynx to lips, in fluent speech. The tracking algorithm is based on a version of the structured speech model consisting of continuous-valued hidden dynamics and a piecewise-linearized Prediction Function from resonance frequencies and bandwidths to LPC cepstra. We present details of the piecewise linearization design process and an adaptive training technique for the parameters that characterize the Prediction residuals. An iterative tracking algorithm is described and evaluated that embeds both the Prediction-residual training and the piecewise linearization design in an adaptive Kalman filtering framework. Experiments on tracking vocal tract resonances in Switchboard speech data demonstrate high accuracy in the results, as well as the effectiveness of residual training embedded in the algorithm. Our approach differs from traditional formant trackers in that it provides meaningful results even during consonantal closures when the supra-laryngeal source may cause no spectral prominences in speech acoustics.
I Bazzi - One of the best experts on this subject based on the ideXlab platform.
-
tracking vocal tract resonances using a quantized nonlinear Function embedded in a temporal constraint
IEEE Transactions on Audio Speech and Language Processing, 2006Co-Authors: Li Deng, Alex Acero, I BazziAbstract:This paper presents a new technique for high-accuracy tracking of vocal-tract resonances (which coincide with formants for nonnasalized vowels) in natural speech. The technique is based on a discretized nonlinear Prediction Function, which is embedded in a temporal constraint on the quantized input values over adjacent time frames as the prior knowledge for their temporal behavior. The nonlinear Prediction is constructed, based on its analytical form derived in detail in this paper, as a parameter-free, discrete mapping Function that approximates the “forward” relationship from the resonance frequencies and bandwidths to the Linear Predictive Coding (LPC) cepstra of real speech. Discretization of the Function permits the “inversion” of the Function via a search operation. We further introduce the nonlinear-Prediction residual, characterized by a multivariate Gaussian vector with trainable mean vectors and covariance matrices, to account for the errors due to the Functional approximation. We develop and describe an expectation–maximization (EM)-based algorithm for training the parameters of the residual, and a dynamic programming-based algorithm for resonance tracking. Details of the algorithm implementation for computation speedup are provided. Experimental results are presented which demonstrate the effectiveness of our new paradigm for tracking vocal-tract resonances. In particular, we show the effectiveness of training the Prediction-residual parameters in obtaining high-accuracy resonance estimates, especially during consonantal closure.
Budiman Minasny - One of the best experts on this subject based on the ideXlab platform.
-
Quantitatively Predicting Soil Carbon Across Landscapes
2014Co-Authors: Budiman Minasny, Alex B Mcbratney, Brendan P Malone, Marine Lacoste, Christian WalterAbstract:Quantitative Prediction of soil carbon (C) in the landscape can be achieved by empirical or mechanistic models, or a combination of both. The empirical approach called digital soil mapping, usually involves: collection of a database of soil carbon observations over an area of interest; compilation of relevant covariates for the area; calibration or training of a spatial Prediction Function based on the observed dataset; interpolation and/or- extrapolation of the Prediction Function over the whole area; and finally validation using existing or independent datasets. The resulting digital maps of C can be used in landscape mechanistic models simulating soil organic C evolution laterally and vertically (within the profile). Here we demonstrate the two approaches in predicting C stock evolution in a landscape in Northwest of France. We introduce the pedogeomorphometry approach which can combine the two approaches to map soil carbon dynamics at the landscape scale.
-
digital mapping of soil carbon
Advances in Agronomy, 2013Co-Authors: Budiman Minasny, Alex B Mcbratney, Brendan P Malone, Ichsani WheelerAbstract:There is a global demand for soil data and information for food security and global environmental management. There is also great interest in recognizing the soil system as a significant terrestrial sink of carbon. The reliable assessment of soil carbon (C) stocks is of key importance for soil conservation and in mitigation strategies for increased atmospheric carbon. In this article, we review and discuss the recent advances in digital mapping of soil C. The challenge to map carbon is demonstrated with the large variation of soil C concentration at a field, continental, and global scale. This article reviews recent studies in mapping soil C using digital soil mapping approaches. The general activities in digital soil mapping involve collection of a database of soil carbon observations over the area of interest; compilation of relevant covariates (scorpan factors) for the area; calibration or training of a spatial Prediction Function based on the observed dataset; interpolation and/or extrapolation of the Prediction Function over the whole area; and finally validation using existing or independent datasets. We discuss several relevant aspects in digital mapping: carbon concentration and carbon density, source of data, sampling density and resolution, depth of investigation, map validation, map uncertainty, and environmental covariates. We demonstrate harmonization of soil depths using the equal-area spline and the use of a material coordinate system to take into consideration the varying bulk density due to management practices. Soil C mapping has evolved from 2-D mapping of soil C stock at particular depth ranges to a semi-3-D soil map allowing the estimation of continuous soil C concentration or density with depth. This review then discusses the dynamics of soil C and the consequences for Prediction and mapping of soil C change. Finally, we illustrate the Prediction of soil carbon change using a semidynamic scorpan approach.
-
on digital soil mapping
Geoderma, 2003Co-Authors: Alex B Mcbratney, M Mendonca L Santos, Budiman MinasnyAbstract:We review various recent approaches to making digital soil maps based on geographic information systems (GIS) data layers, note some commonalities and propose a generic framework for the future. We discuss the various methods that have been, or could be, used for fitting quantitative relationships between soil properties or classes and their ‘environment’. These include generalised linear models, classification and regression trees, neural networks, fuzzy systems and geostatistics. We also review the data layers that have been, or could be, used to describe the ‘environment’. Terrain attributes derived from digital elevation models, and spectral reflectance bands from satellite imagery, have been the most commonly used, but there is a large potential for new data layers. The generic framework, which we call the scorpanSSPFe (soil spatial Prediction Function with spatially autocorrelated errors) method, is particularly relevant for those places where soil resource information is limited. It is based on the seven predictive scorpan factors, a generalisation of Jenny’s five factors, namely: (1) s: soil, other or previously measured attributes of the soil at a point; (2) c: climate, climatic properties of the environment at a point; (3) o: organisms, including land cover and natural vegetation; (4) r: topography, including terrain attributes and classes; (5) p: parent material, including lithology; (6) a: age, the time factor; (7) n: space, spatial or geographic position. Interactions (*) between these factors are also considered. The scorpan-SSPFe method essentially involves the following steps:
Bertrand Thirion - One of the best experts on this subject based on the ideXlab platform.
-
Multi-scale Mining of fMRI Data with Hierarchical Structured Sparsity
2011Co-Authors: Rodolphe Jenatton, Francis Bach, Vincent Michel, Alexandre Gramfort, Guillaume Obozinski, Bertrand ThirionAbstract:Inverse inference, or "brain reading", is a recent paradigm for analyzing Functional magnetic resonance imaging (fMRI) data, based on pattern recognition tools. By predicting some cognitive variables related to brain activation maps, this approach aims at decoding brain activity. Inverse inference takes into account the multivariate information between voxels and is currently the only way to assess how precisely some cognitive information is encoded by the activity of neural populations within the whole brain. However, it relies on a Prediction Function that is plagued by the curse of dimensionality, as we have far more features than samples, i.e., more voxels than fMRI volumes. To address this problem, different methods have been proposed. Among them are univariate feature selection, feature agglomeration and regularization techniques. In this paper, we consider a hierarchical structured regularization. Specifically, the penalization we use is constructed from a tree that is obtained by spatially constrained agglomerative clustering. This approach encodes the spatial prior information in the regularization process, which makes the overall Prediction procedure more robust to inter-subject variability. We test our algorithm on a real data acquired for studying the mental representation of objects, and we show that the proposed algorithm yields better Prediction accuracy than reference methods.
-
Multiclass Sparse Bayesian Regression for fMRI-Based Prediction
International Journal of Biomedical Imaging, 2011Co-Authors: Vincent Michel, Evelyn Eger, Christine Keribin, Bertrand ThirionAbstract:Inverse inference has recently become a popular approach for analyzing neuroimaging data, by quantifying the amount of information contained in brain images on perceptual, cognitive, and behavioral parameters. As it outlines brain regions that convey information for an accurate Prediction of the parameter of interest, it allows to understand how the corresponding information is encoded in the brain. However, it relies on a Prediction Function that is plagued by the curse of dimensionality, as there are far more features (voxels) than samples (images), and dimension reduction is thus a mandatory step. We introduce in this paper a new model, called Multiclass Sparse Bayesian Regression (MCBR), that, unlike classical alternatives, automatically adapts the amount of regularization to the available data. MCBR consists in grouping features into several classes and then regularizing each class differently in order to apply an adaptive and efficient regularization. We detail these framework and validate our algorithm on simulated and real neuroimaging data sets, showing that it performs better than reference methods while yielding interpretable clusters of features.
-
Multi-scale mining of fMRI data with hierarchical structured sparsity
Proceedings - International Workshop on Pattern Recognition in NeuroImaging PRNI 2011, 2011Co-Authors: Rodolphe Jenatton, Francis Bach, Vincent Michel, Evelyn Eger, Alexandre Gramfort, Guillaume Obozinski, Bertrand ThirionAbstract:Inverse inference, or "brain reading", is a recent paradigm for analyzing Functional magnetic resonance imaging (fMRI) data, based on pattern recognition and statistical learning. By predicting some cognitive variables related to brain activation maps, this approach aims at decoding brain activity. Inverse inference takes into account the multivariate information between voxels and is currently the only way to assess how precisely some cognitive information is encoded by the activity of neural populations within the whole brain. However, it relies on a Prediction Function that is plagued by the curse of dimensionality, since there are far more features than samples, i.e., more voxels than fMRI volumes. To address this problem, different methods have been proposed, such as, among others, univariate feature selection, feature agglomeration and regularization techniques. In this paper, we consider a sparse hierarchical structured regularization. Specifically, the penalization we use is constructed from a tree that is obtained by spatially-constrained agglomerative clustering. This approach encodes the spatial structure of the data at different scales into the regularization, which makes the overall Prediction procedure more robust to inter-subject variability. The regularization used induces the selection of spatially coherent predictive brain regions simultaneously at different scales. We test our algorithm on real data acquired to study the mental representation of objects, and we show that the proposed algorithm not only delineates meaningful brain regions but yields as well better Prediction accuracy than reference methods.
-
Multi-Class Sparse Bayesian Regression for Neuroimaging data analysis
2010Co-Authors: Vincent Michel, Evelyn Eger, Christine Keribin, Bertrand ThirionAbstract:The use of machine learning tools is gaining popularity in neuroimaging, as it provides a sensitive assessment of the information conveyed by brain images. In particular, finding regions of the brain whose Functional signal reliably predicts some behavioral information makes it possible to better understand how this information is encoded or processed in the brain. However, such a Prediction is performed through regression or classification algorithms that suffer from the curse of dimensionality, because a huge number of features (i.e. voxels) are available to fit some target, with very few samples (i.e. scans) to learn the informative regions. A commonly used solution is to regularize the weights of the parametric Prediction Function. However, model specification needs a careful design to balance adaptiveness and sparsity. In this paper, we introduce a novel method, Multi-Class Sparse Bayesian Regression (MCBR ), that generalizes classical approaches such as Ridge regression and Automatic Relevance Determination. Our approach is based on a grouping of the features into several classes, where each class is regularized with specific parameters. We apply our algorithm to the Prediction of a behavioral variable from brain activation images. The method presented here achieves similar Prediction accuracies than reference methods, and yields more interpretable feature loadings.
-
Adaptive multi-class Bayesian sparse regression - An application to brain activity classification
2009Co-Authors: Vincent Michel, Evelyn Eger, Christine Keribin, Bertrand ThirionAbstract:In this article we describe a novel method for regularized regression and apply it to the Prediction of a behavioural variable from brain activation images. In the context of neuroimaging, regression or classification techniques are often plagued with the curse of dimensionality, due to the extremely high number of voxels and the limited number of activation maps. A commonly-used solution is the regularization of the weights used in the parametric Prediction Function. It entails the difficult issue of introducing an adapted amount of regularization in the model; this question can be addressed in a Bayesian framework, but model specification needs a careful design to balance adaptiveness and sparsity. Thus, we introduce an adaptive multi-class regularization to deal with this cluster-based structure of the data. Based on a hierarchical model and estimated in a Variational Bayes framework, our algorithm is robust to overfit and more adaptive than other regularization methods. Results on simulated data and preliminary results on real data show the accuracy of the method in the context of brain activation images.
Alejandro Acero - One of the best experts on this subject based on the ideXlab platform.
-
a structured speech model with continuous hidden dynamics and Prediction residual training for tracking vocal tract resonances
International Conference on Acoustics Speech and Signal Processing, 2004Co-Authors: Li Deng, Hagai Attias, Alejandro AceroAbstract:A novel approach is developed for efficient and accurate tracking of vocal tract resonances, which are natural frequencies of the resonator from larynx to lips, in fluent speech. The tracking algorithm is based on a version of the structured speech model consisting of continuous-valued hidden dynamics and a piecewise-linearized Prediction Function from resonance frequencies and bandwidths to LPC cepstra. We present details of the piecewise linearization design process and an adaptive training technique for the parameters that characterize the Prediction residuals. An iterative tracking algorithm is described and evaluated that embeds both the Prediction-residual training and the piecewise linearization design in an adaptive Kalman filtering framework. Experiments on tracking vocal tract resonances in Switchboard speech data demonstrate high accuracy in the results, as well as the effectiveness of residual training embedded in the algorithm. Our approach differs from traditional formant trackers in that it provides meaningful results even during consonantal closures when the supra-laryngeal source may cause no spectral prominences in speech acoustics.