Enmei Tu

Mapping Temporal Variables Into the NeuCube for Improved Pattern Recognition, Predictive Modeling, and Understanding of Stream Data

This paper proposes a new method for an optimized mapping of temporal variables, describing a temporal stream data, into the recently proposed NeuCube spiking neural network (SNN) architecture. This optimized mapping extends the use of the NeuCube, which was initially designed for spatiotemporal brain data, to work on arbitrary stream data and to achieve a better accuracy of temporal pattern recognition, a better and earlier event prediction, and a better understanding of complex temporal stream data through visualization of the NeuCube connectivity. The effect of the new mapping is demonstrated on three benchmark problems. The first one is the early prediction of patient sleep stage event from temporal physiological data. The second one is the pattern recognition of dynamic temporal patterns of traffic in the Bay Area of California and the last one is the Challenge 2012 contest data set. In all the cases, the use of the proposed mapping leads to an improved accuracy of pattern recognition and event prediction and a better understanding of the data when compared with traditional machine learning techniques or SNN reservoirs with an arbitrary mapping of the variables.

NeuCube (ST) for spatio-temporal data predictive modelling with a case study on ecological data

Early event prediction challenges most of existing modeling methods especially when dealing with complex spatio-temporal data. In this paper we propose a new method for predictive data modelling based on a new development of the recently proposed NeuCube spiking neural network architecture, called here NeuCube(ST). The NeuCube uses a Spiking Neural Network reservoir (SNNr) and dynamic evolving Spiking Neuron Network (deSNN) classifier. NeuCube(ST) is an integrated environment including data conversion into spike trains, input variable mapping, unsupervised learning in the SNNr, supervised classification learning, activity visualization and network structure analysis. A case study on a real world ecological data set is presented to demonstrate the validity of the proposed method.

Feasibility of NeuCube SNN architecture for detecting motor execution and motor intention for use in BCIapplications

The paper is a feasibility analysis of using the recently introduced by one of the authors spiking neural networks architecture NeuCube for modelling and recognition of complex EEG spatio-temporal data related to both physical and intentional (imagined) movements. The preliminary experiments reported in the paper suggest that NeuCube is much more efficient for the task than standard machine learning techniques, resulting in high recognition accuracy, a better adaptability to new data, a better interpretation of the models, leading to a better understanding of the brain data and the processes that generated it.

Improved predictive personalized modelling with the use of Spiking Neural Network system and a case study on stroke occurrences data

This paper is a continuation of previous published work by the same authors on Personalized Modelling and Evolving Spiking Neural Network Reservoir architecture (PMeSNNr). The focus is on improvement of predictive modeling methods for the stroke occurrences case study utilizing an enhanced NeuCube architecture. The adaptability of the new architecture leads towards understanding feature correlations that affect the outcome of the study and extracts new knowledge from hidden patterns that reside within the associations. Through this new method, estimation of the earliest time point for stroke prediction is possible. This study also highlighted the improvement from designing a new experimental dataset compared to previous experiments. Comparative experiments were also carried out using conventional machine learning algorithms such as kNN, wkNN, SVM and MLP to prove that our approach can result in much better accuracy level.

An Experimental Comparison of Semi-supervised Learning Algorithms for Multispectral Image Classification

Semi-Supervised Learning (SSL) method has recently caught much attention in the fields of machine learning and computer vision owing to its superiority in classifying abundant unlabelled samples using a few labeled samples. The goal of this paper is to provide an experimental efficiency comparison between graph based SSL algorithms and traditional supervised learning algorithms (e.g., support vector machines) for multispectral image classification. This research shows that SSL algorithms generally outperform supervised learning algorithms in both classification accuracy and anti-noise ability. In the experiments carried out on two data sets (hyperspectral image and Landsat image), the mean overall accuracies (OAs) of supervised learning algorithms are 15 percent and 86 percent, while the mean OAS of SSL algorithms are 26 percent and 99 percent. To overcome the polynomial complexity of SSL algorithms, we also developed a linear-complexity algorithm by employing multivariate Taylor Series Expansion (TSE) and Woodbury Formula.

Efficient multiresolution level set image segmentation with multiple regions

This study is to investigate a new representation of a partition of an image domain into a fixed but arbitrary number of regions via active contours and level sets. The proposed algorithm is composed of simple closed evolving planar curves by an explicit correspondence to minimize the energy functional containing three terms: multiregion fitting energy, regularization related to the length of the curve, and the distance regular- izing term to penalize the deviation of the level set function from a signed distance function. This formulation leads to a system of coupled curve evolution equations, which is easily amenable to a level set implemen- tation, and an unambiguous segmentation because the evolving regions form a partition of the image domain at all times during curve evolution. In order to increase the robustness of the method to noise and to reduce the computational cost, a multiresolution level set schema is proposed, which can perform the evolution curves of the partitioned image at a different resolution. Given these advantages, the proposed method can get good performance and experiments show promising segmentation results on both synthetic and real images. C 2011 Society of Photo-Optical Instrumentation

Statistical approaches to automatic level set image segmentation with multiple regions

This study is to investigate a new representation of a partition of an image domain into a number of regions using a level set method derived from a statistical framework. The proposed model is composed of evolving simple closed planar curves by a region-based force determined by maximizing the posterior image densities over all possible partitions of the image plane containing three terms: a Bayesian term based on the prior probability, a regularity term adopted to avoid the generation of excessively irregular and small segmented regions, and a term based on a region merging prior related to region area, which is applied to allow the number of regions to vary automatically during curve evolution and therefore can optimize the objective functional implicitly with respect to the number of regions. This formulation leads to a system of coupled curve evolution equations, which is easily amenable to a level set implementation, and an unambiguous segmentation because the evolving regions form a partition of the image domain at all times during curve evolution. Given these advantages, the proposed method can get good performance and experiments show promising segmentation results on both synthetic and real images.

Multilayer level set method for multiregion image segmentation

The purpose of this study is to propose a novel method of a partition of an image domain into an adaptive number of regions using a multilayer foreground-filled method. First, two coupled curves based on a three-region Chan-Vese model, which is built based on the techniques of evolving simple closed planar curves by an explicit correspondence to minimize energy functional containing a fitting term and a regularization term, evolve simultaneously to segment images containing two objects and one background region in each image layer. Second, a foreground-filled technique is used to generate a new image and the three-region Chan-Vese model is repeated to segment the new image for the next image layer. To avoid the long iteration process for level set evolution, an efficient termination criterion is presented on the basis on the length change of an evolving curve. This iterative process is repeated until the background image layer is detected. Numerical experiments on some synthetic and real images have demonstrated the efficiency and robustness of our method.

From von Neumann Architecture and Atanasoffs ABC to Neuro-Morphic Computation and Kasabov’s NeuCube: Principles and Implementations

During the 1940s John Atanasoff with the help of one of his students Clifford E. Berry, at Iowa State College, created the ABC (Atanasoff-Berry Computer) that was the first electronic digital computer. The ABC computer was not a general-purpose one, but still, it was the first to implement three of the most important ideas used in computers nowadays: binary data representation; using electronics instead of mechanical switches and wheels; using a von Neumann architecture, where the memory and the computations are separated. A new computational paradigm, named as Neuromorphic, utilises the above two principles, but instead of the von Neumann principle, it integrates the memory and the computation in a single module a spiking neural network structure. This chapter first reviews the principles of the earlier published work by the team on neuromorphic computational architecture NeuCube. NeuCube is not a general purpose machine but is still the first neuromorphic spatio/spectro-temporal data machine for learning, pattern recognition and understanding of spatio/spectro-temporal data. The chapter further presents the software/hardware implementation of the NeuCube as a development system for efficient applications on temporal or spatio/spectro-temporal across domain areas, including: brain data (EEG, fMRI), brain computer interfaces, robot control, multi-sensory data modelling, seismic stream data modelling and earthquake prediction, financial time series forecasting, climate data modelling and personalised, on-line risk of stroke prediction, and others. A limited version of the NeuCube software implementation is available from http://www.kedri.aut.ac.nz/neucube/.

Posterior Distribution Learning (PDL): A Novel Supervised Learning Framework

In order to obtain a robust supervised model with good generalization ability, traditional supervised learning method has to be trained with sufficient well labeled and uniformly distributed samples. However, in many real applications, the cost of labeled samples is generally very expensive. How to make use of ample easily available unlabeled samples to remedy the insufficiency of labeled samples to train a supervised model is of great interest and practical significance. In this paper we propose a new supervised learning framework, Posterior Distribution Learning (PDL), which could train a robust supervised model with very a few labeled samples by including those unlabeled samples into training stage. Experimental results on both synthetic and real world data sets are presented to demonstrate the effectiveness of the proposed framework.