检索结果-内蒙古大学图书馆

Modeling sensorimotor learning with linear dynamical systems

NEURAL COMPUTATION 2006年第4期18卷 760-793页

作者： Cheng, S Sabes, PN Univ Calif San Francisco Sloan Swartz Ctr Theoret Neurobiol WM Keck Fdn Ctr Integrat Neurosci San Francisco CA 94143 USA Univ Calif San Francisco Dept Physiol San Francisco CA 94143 USA

Recent studies have employed simple linear dynamical systems to model trial-by-trial dynamics in various sensorimotor learning tasks. Here we explore the theoretical and practical considerations that arise when employing the general class of linear dynamical systems (LDS) as a model for sensorimotor learning. In this framework, the state of the system is a set of parameters that define the current sensorimotor transformation-the function that maps sensory inputs to motor outputs. The class of LDS models provides a first-order approximation for any Markovian (state-dependent) learning rule that specifies the changes in the sensorimotor transformation that result from sensory feedback on each movement. We show that modeling the trial-by-trial dynamics of learning provides a substantially enhanced picture of the process of adaptation compared to measurements of the steady state of adaptation derived from more traditional blocked-exposure experiments. Specifically, these models can be used to quantify sensory and performance biases, the extent to which learned changes in the sensorimotor transformation decay over time, and the portion of motor variability due to either learning or performance variability. We show that previous attempts to fit such models with linear regression have not generally yielded consistent parameter estimates. Instead, we present an expectation-maximization algorithm for fitting LDS models to experimental data and describe the difficulties inherent in estimating the parameters associated with feedback-driven learning. Finally, we demonstrate the application of these methods in a simple sensorimotor learning experiment: adaptation to shifted visual feedback during reaching.

关键词： MOTOR learning SENSORIMOTOR cortex LINEAR systems PERCEPTUAL motor learning MARKOV spectrum expectation-maximization algorithms

来源：评论

学校读者我要写书评

暂无评论

A Region-Based Image Fusion Method Using the expectation-maximization Algorithm

A Region-Based Image Fusion Method Using the Expectation-Max...

引用

Annual Conference on Information Sciences and Systems (CISS)

作者： Jinzhong Yang Rick S. Blum ECE Department Lehigh University Bethlehem PA USA

We present a novel region-based image fusion method using a rigorous application of estimation theory. This method takes advantage of the similar intensity or texture in a region for fusion. A statistical image formation model using Gaussian mixture distortion is built for each region and the EM (expectation-maximization) algorithm is used in conjunction with the model to develop the region-level EM fusion algorithm to produce the fused image. Since in most applications of image fusion, objects carry the information of interest and regions can be used to represent objects, the region-based fusion approaches could be more meaningful than pixel-based methods. Our experiments demonstrate the efficiency of the proposed region-base fusion method and the advantages in dealing with region interface artifacts for concealed weapon detection and night vision applications.

关键词： Image fusion expectation-maximization algorithms Pixel Image segmentation Iterative algorithms Estimation theory Weapons Night vision Image analysis Layout

来源：评论

学校读者我要写书评

暂无评论

On Ryde/spl acute/n's EM algorithm for estimating MMPPs

引用

IEEE Signal Processing Letters 2006年第6期13卷 373-376页

作者： W.J.J. Roberts Y. Ephraim E. Dieguez Atlantic Coast Technologies Inc. Silver Spring MD USA Department of Electrical and Computer Engineering George Mason University Fairfax VA USA

Two aspects of Ryden's expectation-maximization algorithm for estimating the parameter of a Markov modulated Poisson process are addressed. First, a scaling procedure is developed for the forward-backward recursions that circumvents the need for customized floating-point software. Second, evaluation of integrals of matrix exponentials is facilitated by applying a result due to Van Loan. For an MMPP of order four, a speedup of over two orders of magnitude was observed.

关键词： Parameter estimation Signal processing algorithms Maximum likelihood estimation Application software State estimation expectation-maximization algorithms Computer networks Finance Recursive estimation Control theory

来源：评论

学校读者我要写书评

暂无评论

A Probabilistic Ensemble Pruning Algorithm

A Probabilistic Ensemble Pruning Algorithm

引用

IEEE International Conference on Data Mining Workshops (ICDM Workshops)

作者： Huanhuan Chen Peter Tino Xin Yao School of Computer Science University of Binningham Birmingham UK

An ensemble is a group of learners that work together as a committee to solve a problem. However, the existing ensemble training algorithms sometimes generate unnecessary large ensembles, which consume extra computational resource and may degrade the performance. Ensemble pruning algorithm aims to find a good subset of ensemble members to constitute a small ensemble, which saves the computational resource and performs as well as, or better than, the non-pruned ensemble. This paper introduces a probabilistic ensemble pruning algorithm by choosing a set of "sparse" combination weights, most of which are zero, to prune the large ensemble. In order to obtain the set of sparse combination weights and satisfy the non-negative restriction of the combination weights, a left-truncated, non-negative, Gaussian prior is adopted over every combination weight. expectation-maximization algorithm is employed to obtain maximum a posterior (MAP) estimation of weight vector. Four benchmark regression problems and another four benchmark classification problems have been employed to demonstrate the effectiveness of the method

关键词： Machine learning Degradation Bayesian methods Gaussian distribution Computer science expectation-maximization algorithms Bagging Boosting Evolutionary computation Machine learning algorithms

来源：评论

学校读者我要写书评

暂无评论

A Novel Model-based Method for Feature Extraction from Protein Sequences for Classification

A Novel Model-based Method for Feature Extraction from Prote...

引用

IEEE Signal Processing and Communications Applications (SIU)

作者： Sarac Atalay Bilgisayar Mühendisliği Bölümü Orta Doğu Teknik Üniversitesi Ankara Turkey

Representation of amino-acid sequences constitutes the key point in classification of proteins into functional or structural classes. The representation should contain the biologically meaningful information hidden in the primary sequence of the protein. Conserved or similar subsequences are strong indicators of functional and structural similarity. In this study we present a feature mapping that takes into account the models of the subsequences of protein sequences. An expectation-maximization algorithm along with an HMM mixture model is used to cluster and learn the models of subsequences of a given set of proteins

关键词： Feature extraction Proteins Hidden Markov models expectation-maximization algorithms Influenza

来源：评论

学校读者我要写书评

暂无评论

Iterative Synchronization: EM Algorithm Versus Newton-Raphson Method

Iterative Synchronization: EM Algorithm Versus Newton-Raphso...

引用

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： C. Herzet X. Wautelet V. Ramon L. Vandendorpe Communications Laboratory Université catholique de Louvain Louvain-la-Neuve Belgium

This paper deals with iterative maximum-likelihood synchronization of a scalar parameter. An efficient implementation of the Newton-Raphson (NR) maximum-search method is proposed. Considering the latter implementation, the NR approach is shown to be an attractive alternative to synchronization methods based on the expectation-maximization (EM) algorithm. Simulation results for the case of phase-offset synchronization show that NR method usually increases the speed of convergence of the synchronization algorithm

关键词： Iterative methods Iterative algorithms Newton method Maximum likelihood estimation Convergence Parameter estimation Laboratories Signal to noise ratio Sum product algorithm expectation-maximization algorithms

来源：评论

学校读者我要写书评

暂无评论

Unsupervised Image Layout Extraction

Unsupervised Image Layout Extraction

引用

IEEE International Conference on Image Processing

作者： David Liu Datong Chen Tsuhan Chen Department of Electrical and Computer Engineering School of Computer Science Carnegie Mellon University Pittsburgh USA

We propose a novel unsupervised learning algorithm to extract the layout of an image by learning latent object-related aspects. Unlike traditional image segmentation algorithms that segment an image using feature similarity, our method is able to learn high-level object characteristics (aspects) from a large number of unlabeled images containing similar objects to facilitate image segmentation. Our method does not require human to annotate the training set and works without supervision. We use a graphical model to address the learning of aspects and layout extraction together. In particular, aspect-feature dependency from multiple images is learned via the expectation-maximization algorithm. We demonstrate that, by associating latent aspects to spatial structure, the proposed method achieves much better layout extraction results than using probabilistic latent semantic analysis.

关键词： Image segmentation Unsupervised learning Data mining Graphical models Layout Image analysis Object detection Computer science Humans expectation-maximization algorithms

来源：评论

学校读者我要写书评

暂无评论

Quantitative Stochastic Analysis of Magnetic Resonance Images of the Brain

Quantitative Stochastic Analysis of Magnetic Resonance Image...

引用

International Conference on Information and Communication Technologies: From Theory to Applications, ICTTA

作者： M. Rouainia N. Doghmane LAS laboratory Skikda university Algeria LAMA Laboratory Badji Mokhtar University Annaba Algeria

The aim of our work is to perform an automated tool for brain MRI tissues quantification. The method we develop is based on MRI intensity stochastic analysis. By the use of the Gaussian mixture model for these intensities, we estimate MRI tissues parameters with a combination of the expectation-maximization algorithm and the Markov random field model witch provide contextual constraints that improve the classification of image pixels into three classes of tissue: white matter, grey matter and cerebro-spinal fluid. The automated model based algorithm is also extended to take in account an important MRI artefact: the bias field caused by electromagnetic field inhomogeneities. The resulting automated MRI analysis method simultaneously corrects from MR field inhomogeneities, estimates tissue classes distribution parameters, classifies the image and detects multiple sclerosis lesions when treated images present this pathology. We validate our method on simulated data then on real MRI scans

关键词： Stochastic resonance Image analysis Magnetic analysis Magnetic resonance Magnetic resonance imaging Context modeling Stochastic processes Parameter estimation expectation-maximization algorithms Markov random fields

来源：评论

学校读者我要写书评

暂无评论

Probabilistic Image-Based Rendering with Gaussian Mixture Model

Probabilistic Image-Based Rendering with Gaussian Mixture Mo...

引用

International Conference on Pattern Recognition

作者： Wenfeng Li Baoxin Li Department of Computer Science & Engineering Arizona State University Tempe AZ USA

ISBN: (纸本)0769525210

One major challenge in traditional image-based rendering is 3D scene reconstruction by estimating accurate dense depth map, which suffers from the ambiguities in textureless or periodically textured regions. Alternatively, statistical methods may be used to estimate a most likely color for each pixel for photorealistic rendering from multiple views of the same scene. Such statistical methods normally require a relatively large number of input images to achieve reasonable quality for the synthesized image, if the estimation is purely nonparametric. In this paper, based on some reasonable assumptions on the configuration of the multiple views, we propose to use a two-component Gaussian mixture model for the appearance of a given pixel in all the views so that both the problem of occlusion and the problem of noise can be considered simultaneously. Then we use the expectation-maximization algorithm to estimate the model parameters. The virtual pixel is given as a maximum likelihood estimate for one of the mixture components. Experiments shows that reasonable performance can be obtained even with only a few input images

关键词： Rendering (computer graphics) Layout Pixel Image reconstruction Statistical analysis expectation-maximization algorithms Parameter estimation Maximum likelihood estimation Information geometry Cameras

来源：评论

学校读者我要写书评

暂无评论

Acoustic Model Adaptation Based on Coarse/Fine Training of Transfer Vectors Using Directional Statistics

Acoustic Model Adaptation Based on Coarse/Fine Training of T...

引用

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： S. Watanabe A. Nakamura NTT Communication Science Laboratories NTT Corporation Japan

In this paper, we reformulate an adaptation scheme of coarse/fine training (CFT) of transfer vectors in acoustic modeling by using directional statistics. In CFT, the transfer vector is decomposed into a unit direction vector and a scaling factor. By using coarse tied Gaussian class (coarse class) estimation for the unit direction vector, and by using fine tied Gaussian class (fine class) estimation for the scaling factor, we can obtain accurate transfer vectors with a small number of free parameters. Directional statistics is a method for analyzing geometric parameters (e.g. angle and unit vector) using directional data, and is suited for the analysis of the CFT representation. Using directional statistics as a basis, we construct expectation-maximization algorithms for CFT parameters analytically using the maximum likelihood and Bayesian (maximum a posteriori) approaches. In particular, with the Bayesian approach, prior and posterior distributions for unit direction vectors are represented with a von Mises distribution, a representative distribution in directional statistics. Speaker adaptation experiments show that our proposal improves the performance of large vocabulary continuous speech recognition due to the efficient coarse/fine representation of transfer vectors, compared with the conventional transfer vector adaptation

关键词： Adaptation model Statistics Statistical analysis Statistical distributions Bayesian methods expectation-maximization algorithms Algorithm design and analysis Maximum likelihood estimation Proposals Vocabulary

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：