Study of phase synchronization in multichannel seizure EEG using nonlinear recurrence measure

Complex biological systems such as the human brain can be expected to be inherently nonlinear and hence difficult to model. Most of the previous studies on investigations of brain function have either used linear models or parametric nonlinear models. In this paper, we propose a novel application of a nonlinear measure of phase synchronization based on recurrences, correlation between probabilities of recurrence (CPR), to study seizures in the brain. The advantage of this nonparametric method is that it makes very few assumptions thus making it possible to investigate brain functioning in a data-driven way. We have demonstrated the utility of CPR measure for the study of phase synchronization in multichannel seizure EEG recorded from patients with global as well as focal epilepsy. For the case of global epilepsy, brain synchronization using thresholded CPR matrix of multichannel EEG signals showed clear differences in results obtained for epileptic seizure and pre-seizure. Brain headmaps obtained for seizure and preseizure cases provide meaningful insights about synchronization in the brain in those states. The headmap in the case of focal epilepsy clearly enables us to identify the focus of the epilepsy which provides certain diagnostic value. Comparative studies with linear correlation have shown that the nonlinear measure CPR outperforms the linear correlation measure. (C) 2014 Elsevier Ltd. All rights reserved.

Compiler/Runtime Framework for Dynamic Dataflow Parallelization of Tiled Programs

Task-parallel languages are increasingly popular. Many of them provide expressive mechanisms for intertask synchronization. For example, OpenMP 4.0 will integrate data-driven execution semantics derived from the StarSs research language. Compared to the more restrictive data-parallel and fork-join concurrency models, the advanced features being introduced into task-parallelmodels in turn enable improved scalability through load balancing, memory latency hiding, mitigation of the pressure on memory bandwidth, and, as a side effect, reduced power consumption. In this article, we develop a systematic approach to compile loop nests into concurrent, dynamically constructed graphs of dependent tasks. We propose a simple and effective heuristic that selects the most profitable parallelization idiom for every dependence type and communication pattern. This heuristic enables the extraction of interband parallelism (cross-barrier parallelism) in a number of numerical computations that range from linear algebra to structured grids and image processing. The proposed static analysis and code generation alleviates the burden of a full-blown dependence resolver to track the readiness of tasks at runtime. We evaluate our approach and algorithms in the PPCG compiler, targeting OpenStream, a representative dataflow task-parallel language with explicit intertask dependences and a lightweight runtime. Experimental results demonstrate the effectiveness of the approach.

Multiscale model reduction methods for deterministic and stochastic partial differential equations

Partial differential equations (PDEs) with multiscale coefficients are very difficult to solve due to the wide range of scales in the solutions. In the thesis, we propose some efficient numerical methods for both deterministic and stochastic PDEs based on the model reduction technique.

For the deterministic PDEs, the main purpose of our method is to derive an effective equation for the multiscale problem. An essential ingredient is to decompose the harmonic coordinate into a smooth part and a highly oscillatory part of which the magnitude is small. Such a decomposition plays a key role in our construction of the effective equation. We show that the solution to the effective equation is smooth, and could be resolved on a regular coarse mesh grid. Furthermore, we provide error analysis and show that the solution to the effective equation plus a correction term is close to the original multiscale solution.

For the stochastic PDEs, we propose the model reduction based data-driven stochastic method and multilevel Monte Carlo method. In the multiquery, setting and on the assumption that the ratio of the smallest scale and largest scale is not too small, we propose the multiscale data-driven stochastic method. We construct a data-driven stochastic basis and solve the coupled deterministic PDEs to obtain the solutions. For the tougher problems, we propose the multiscale multilevel Monte Carlo method. We apply the multilevel scheme to the effective equations and assemble the stiffness matrices efficiently on each coarse mesh grid. In both methods, the $\KL$ expansion plays an important role in extracting the main parts of some stochastic quantities.

For both the deterministic and stochastic PDEs, numerical results are presented to demonstrate the accuracy and robustness of the methods. We also show the computational time cost reduction in the numerical examples.

The design and implementation of dialogue control in voice operated database inquiry systems

This paper describes work performed as part of the U.K. Alvey sponsored Voice Operated Database Inquiry System (VODIS) project in the area of intelligent dialogue control. The principal aims of the work were to develop a habitable interface for the untrained user; to investigate the degree to which dialogue control can be used to compensate for deficiencies in recognition performance; and to examine the requirements on dialogue control for generating natural speech output. A data-driven methodology is described based on the use of frames in which dialogue topics are organized hierarchically. The concept of a dynamically adjustable scope is introduced to permit adaptation to recognizer performance and the use of historical and hierarchical contexts are described to facilitate the construction of contextually relevant output messages. © 1989.

POMDP-based statistical spoken dialog systems: A review

Statistical dialog systems (SDSs) are motivated by the need for a data-driven framework that reduces the cost of laboriously handcrafting complex dialog managers and that provides robustness against the errors created by speech recognizers operating in noisy environments. By including an explicit Bayesian model of uncertainty and by optimizing the policy via a reward-driven process, partially observable Markov decision processes (POMDPs) provide such a framework. However, exact model representation and optimization is computationally intractable. Hence, the practical application of POMDP-based systems requires efficient algorithms and carefully constructed approximations. This review article provides an overview of the current state of the art in the development of POMDP-based spoken dialog systems. © 1963-2012 IEEE.

Smoothing LDA Model for Text Categorization

Abstract. Latent Dirichlet Allocation (LDA) is a document level language model. In general, LDA employ the symmetry Dirichlet distribution as prior of the topic-words’ distributions to implement model smoothing. In this paper, we propose a data-driven smoothing strategy in which probability mass is allocated from smoothing-data to latent variables by the intrinsic inference procedure of LDA. In such a way, the arbitrariness of choosing latent variables'priors for the multi-level graphical model is overcome. Following this data-driven strategy,two concrete methods, Laplacian smoothing and Jelinek-Mercer smoothing, are employed to LDA model. Evaluations on different text categorization collections show data-driven smoothing can significantly improve the performance in balanced and unbalanced corpora.

Heavy-ion conformal irradiation in the shallow-seated tumor therapy terminal at HIRFL

Basic research related to heavy-ion cancer therapy has been done at the Institute of Modern Physics (IMP), Chinese Academy of Sciences since 1995. Now a plan of clinical trial with heavy ions has been launched at IMP. First, superficially placed tumor treatment with heavy ions is expected in the therapy terminal at the Heavy Ion Research Facility in Lanzhou (HIRFL), where carbon ion beams with energy up to 100 MeV/u can be supplied. The shallow-seated tumor therapy terminal at HIRFL is equipped with a passive beam delivery system including two orthogonal dipole magnets, which continuously scan pencil beams laterally and generate a broad and uniform irradiation field, a motor-driven energy degrader and a multi-leaf collimator. Two different types of range modulator, ripple filter and ridge filter with which Guassian-shaped physical dose and uniform biological effective dose Bragg peaks can be shaped for therapeutic ion beams respectively, have been designed and manufactured. Therefore, two-dimensional and three-dimensional conformal irradiations to tumors can be performed with the passive beam delivery system at the earlier therapy terminal. Both the conformal irradiation methods have been verified experimentally and carbon-ion conformal irradiations to patients with superficially placed tumors have been carried out at HIRFL since November 2006.

Capability verification of the beam delivery system in the superficially-placed tumor therapy terminal at HIRFL

The passive beam delivery system in the superficially-placed tumor therapy terminal at Heavy Ion Researc h Facility in Lanzhou (HIRFL), which includes two orthogonal dipole magnets as scanning system, a motor-driven energy degrader as range-shifter, series of ridge filters as range modulator and a multileaf collimator, is introduced in detail. The capacities of its important components and the whole system have been verified experimentally. The tests of the ridge filter for extending Bragg peak and the range shifter for energy adjustment show both work well. To examine the passive beam delivery system, a beam shaping experiment were carried out, simulating a three-dimensional (3D) conformal irradiation to a tumor. The encouraging experimental result confirms that 3D layer-stacking conformal irradiation can be performed by means of the passive system. The validation of the beam delivery system establishes a substantial basis for upcoming clinical trial for superficially-placed tumors with heavy ions in the therapy terminal at HIRFL.

数据驱动型多次波衰减方法的研究

In the practical seismic profile multiple reflections tend to impede the task of even the experienced interpreter in deducing information from the reflection data. Surface multiples are usually much stronger, more broadband, and more of a problem than internal multiples because the reflection coefficient at the water surface is much larger than the reflection coefficients found in the subsurface. For this reason most attempts to remove multiples from marine data focus on surface multiples, as will I. A surface-related multiple attenuation method can be formulated as an iterative procedure. In this essay a fully data-driven approach which is called MPI —multiple prediction through inversion (Wang, 2003) is applied to a real marine seismic data example. This is a pretty promising scheme for predicting a relative accurate multiple model by updating the multiple model iteratively, as we usually do in a linearized inverse problem. The prominent characteristic of MPI method lie in that it eliminate the need for an explicit surface operator which means it can model the multiple wavefield without any knowledge of surface and subsurface structures even a source signature. Another key feature of this scheme is that it can predict multiples not only in time but also in phase and in amplitude domain. According to the real data experiments it is shown that this scheme for multiple prediction can be made very efficient if a good initial estimate of the multiple-free data set can be provided in the first iteration. In the other core step which is multiple subtraction we use an expanded multi-channel matching filter to fulfil this aim. Compared to a normal multichannel matching filter where an original seismic trace is matched by a group of multiple-model traces, in EMCM filter a seismic trace is matched by not only a group of the ordinary multiple-model traces but also their adjoints generated mathematically. The adjoints of a multiple-model trace include its first derivative, its Hilbert transform and the derivative of the Hilbert transform. The third chapter of the thesis is the application for the real data using the previous methods we put forward from which we can obviously find the effectivity and prospect of the value in use. For this specific case I have done three group experiments to test the effectiveness of MPI method, compare different subtraction results with fixed filter length but different window length, invest the influence of the initial subtraction result for MPI method. In terms of the real data application, we do fine that the initial demultiple estimate take on a great deal of influence for the MPI method. Then two approaches are introduced to refine the intial demultiple estimate which are first arrival and masking filter respectively. In the last part some conclusions are drawn in terms of the previous results I have got.

Learning Object-Independent Modes of Variation with Feature Flow Fields

We present a unifying framework in which "object-independent" modes of variation are learned from continuous-time data such as video sequences. These modes of variation can be used as "generators" to produce a manifold of images of a new object from a single example of that object. We develop the framework in the context of a well-known example: analyzing the modes of spatial deformations of a scene under camera movement. Our method learns a close approximation to the standard affine deformations that are expected from the geometry of the situation, and does so in a completely unsupervised (i.e. ignorant of the geometry of the situation) fashion. We stress that it is learning a "parameterization", not just the parameter values, of the data. We then demonstrate how we have used the same framework to derive a novel data-driven model of joint color change in images due to common lighting variations. The model is superior to previous models of color change in describing non-linear color changes due to lighting.

A Radial Basis Function Approach to Financial Time Series Analysis

Nonlinear multivariate statistical techniques on fast computers offer the potential to capture more of the dynamics of the high dimensional, noisy systems underlying financial markets than traditional models, while making fewer restrictive assumptions. This thesis presents a collection of practical techniques to address important estimation and confidence issues for Radial Basis Function networks arising from such a data driven approach, including efficient methods for parameter estimation and pruning, a pointwise prediction error estimator, and a methodology for controlling the "data mining'' problem. Novel applications in the finance area are described, including customized, adaptive option pricing and stock price prediction.

Approximation-based feature selection and application for algae population estimation

Q. Shen and R. Jensen, 'Approximation-based feature selection and application for algae population estimation,' Applied Intelligence, vol. 28, no. 2, pp. 167-181, 2008. Sponsorship: EPSRC RONO: EP/E058388/1

Power system parameters forecasting using Hilbert-Huang transform and machine learning

A novel hybrid data-driven approach is developed for forecasting power system parameters with the goal of increasing the efficiency of short-term forecasting studies for non-stationary time-series. The proposed approach is based on mode decomposition and a feature analysis of initial retrospective data using the Hilbert-Huang transform and machine learning algorithms. The random forests and gradient boosting trees learning techniques were examined. The decision tree techniques were used to rank the importance of variables employed in the forecasting models. The Mean Decrease Gini index is employed as an impurity function. The resulting hybrid forecasting models employ the radial basis function neural network and support vector regression. A part from introduction and references the paper is organized as follows. The second section presents the background and the review of several approaches for short-term forecasting of power system parameters. In the third section a hybrid machine learningbased algorithm using Hilbert-Huang transform is developed for short-term forecasting of power system parameters. Fourth section describes the decision tree learning algorithms used for the issue of variables importance. Finally in section six the experimental results in the following electric power problems are presented: active power flow forecasting, electricity price forecasting and for the wind speed and direction forecasting.

Comparison of pattern detection methods in microarray time series of the segmentation clock.

While genome-wide gene expression data are generated at an increasing rate, the repertoire of approaches for pattern discovery in these data is still limited. Identifying subtle patterns of interest in large amounts of data (tens of thousands of profiles) associated with a certain level of noise remains a challenge. A microarray time series was recently generated to study the transcriptional program of the mouse segmentation clock, a biological oscillator associated with the periodic formation of the segments of the body axis. A method related to Fourier analysis, the Lomb-Scargle periodogram, was used to detect periodic profiles in the dataset, leading to the identification of a novel set of cyclic genes associated with the segmentation clock. Here, we applied to the same microarray time series dataset four distinct mathematical methods to identify significant patterns in gene expression profiles. These methods are called: Phase consistency, Address reduction, Cyclohedron test and Stable persistence, and are based on different conceptual frameworks that are either hypothesis- or data-driven. Some of the methods, unlike Fourier transforms, are not dependent on the assumption of periodicity of the pattern of interest. Remarkably, these methods identified blindly the expression profiles of known cyclic genes as the most significant patterns in the dataset. Many candidate genes predicted by more than one approach appeared to be true positive cyclic genes and will be of particular interest for future research. In addition, these methods predicted novel candidate cyclic genes that were consistent with previous biological knowledge and experimental validation in mouse embryos. Our results demonstrate the utility of these novel pattern detection strategies, notably for detection of periodic profiles, and suggest that combining several distinct mathematical approaches to analyze microarray datasets is a valuable strategy for identifying genes that exhibit novel, interesting transcriptional patterns.

Component Neural Systems for the Creation of Emotional Memories during Free Viewing of a Complex, Real-World Event.

To investigate the neural systems that contribute to the formation of complex, self-relevant emotional memories, dedicated fans of rival college basketball teams watched a competitive game while undergoing functional magnetic resonance imaging (fMRI). During a subsequent recognition memory task, participants were shown video clips depicting plays of the game, stemming either from previously-viewed game segments (targets) or from non-viewed portions of the same game (foils). After an old-new judgment, participants provided emotional valence and intensity ratings of the clips. A data driven approach was first used to decompose the fMRI signal acquired during free viewing of the game into spatially independent components. Correlations were then calculated between the identified components and post-scanning emotion ratings for successfully encoded targets. Two components were correlated with intensity ratings, including temporal lobe regions implicated in memory and emotional functions, such as the hippocampus and amygdala, as well as a midline fronto-cingulo-parietal network implicated in social cognition and self-relevant processing. These data were supported by a general linear model analysis, which revealed additional valence effects in fronto-striatal-insular regions when plays were divided into positive and negative events according to the fan's perspective. Overall, these findings contribute to our understanding of how emotional factors impact distributed neural systems to successfully encode dynamic, personally-relevant event sequences.