Cross-modality earth mover’s distance-driven convolutional neural network for different-modality data

Neural Computing and Applications - Cross-modality matching refers to the problem of comparing similarity/dissimilarity of a pair of data points of different modalities, such as an image and a...     

Machine learning technique for early detection of Alzheimer’s disease

Microsystem Technologies - Alzheimer’s disease (AD) is non-repairable brain disorder which impacts a person’s thinking along with shrinking the size of the brain, ultimately resulting...     

Segmentation of abnormal liver region based on earth mover’s distance between histograms with mapping of the distances by multidimensional scaling

In this paper, we propose using earth mover’s distance (EMD) to obtain the appropriate similarity between each histogram for segmentation of abnormal liver regions with mapping of the distances by multidimensional scaling. Conventionally, the similarity between each histogram is calculated by integrating the difference between each bin of the histograms. However, this similarity is unsuitable for appropriate comparison of the histograms because the number of bins for calculating the local histograms of computed tomography images varies. We used EMD to resolve this problem regarding the difference in bin numbers, and the obtained distances are used for mapping the local histograms by multidimensional scaling to low-dimensional space. In the low-dimensional space, the abnormal liver region was well segmented by support vector machine in the test datasets.     

Numerical Studies of Three-dimensional Stochastic Darcy’s Equation and Stochastic Advection-Diffusion-Dispersion Equation

Solute transport in randomly heterogeneous porous media is commonly described by stochastic flow and advection-dispersion equations with a random hydraulic conductivity field. The statistical distribution of conductivity of engineered and naturally occurring porous material can vary, depending on its origin. We describe solutions of a three-dimensional stochastic advection-dispersion equation using a probabilistic collocation method (PCM) on sparse grids for several distributions of hydraulic conductivity. Three random distributions of log hydraulic conductivity are considered: uniform, Gaussian, and truncated Gaussian (beta). Log hydraulic conductivity is represented by a Karhunen-Loève (K-L) decomposition as a second-order random process with an exponential covariance function. The convergence of PCM has been demonstrated. It appears that the accuracy in both the mean and the standard deviation of PCM solutions can be improved by using the Jacobi-chaos representing the truncated Gaussian distribution rather than the Hermite-chaos for the Gaussian distribution. The effect of type of distribution and parameters such as the variance and correlation length of log hydraulic conductivity and dispersion coefficient on leading moments of the advection velocity and solute concentration was investigated.     

Re-transfer learning and multi-modal learning assisted early diagnosis of Alzheimer’s disease

Nowadays more and more elderly people are suffering from Alzheimer’s disease (AD). Finely recognizing mild cognitive impairment (MCI) in early stage of the symptom is vital for AD therapy. However, brain image samples are relatively scarce, meanwhile have multiple modalities, which makes finely classifying brain images by computers extremely difficult. This paper proposes a fine-grained brain image classification approach for diagnosing Alzheimer’s disease, with re-transfer learning and multi-modal learning. First of all, an end-to-end deep neural network classifier CNN4AD is designed to finely classify diffusion tensor image (DTI) into four categories. And according to the characteristics of multi-modal brain image dataset, the re-transfer learning method is proposed based on transfer learning and multi-modal learning theories. Experimental results show that the proposed approach obtain higher accuracy with less labeled training samples. This could help doctors diagnose Alzheimer’s disease more timely and accurately.

     

Group sparse based locality – sensitive dictionary learning for video semantic analysis

Multimedia Tools and Applications - Sparse Representation-based Classifier (SRC) and Dictionary Learning (DL), have significantly impacted greatly on the classification performance of image...     

Machine learning Ensemble for the Parkinson’s disease using protein sequences

Multimedia Tools and Applications - The most challenging issue in diagnosing and treating neurological disorders is gene identification that causes the disease. Classification of the genes that...     

A matrix-based methodology for determining a part family’s learning rate

Learning is a decrease in the time to perform an operation due to repetition and is an important consideration when forecasting process times or product costs. This paper presents a new method for calculating the learning rate for a family of parts using a matrix-based approach to organize historical data on production times. By calculating a single learning rate for the entire family, the data on individual parts is pooled, creating a larger sample size and reducing the variance of the estimate. Applying this method to forecasting costs of a family of jet engine parts shows that it provides much more accurate estimates than the previously available method of taking a weighted average of individual parts’ learning rates. The matrix-based method also allows for calculation of first-unit costs more reliably (since the estimates are less affected by outliers in a larger sample) and for calculation of confidence limits on the estimates, to provide users with information on the reliability of the estimates.     

Augmenting machine learning for Amharic speech recognition: a paradigm of patient’s lips motion detection

Multimedia Tools and Applications - The method of automatic lip motion recognition is an essential input for visual speech detection. It is a technological approach to demystify people who are hard...     

Insight into an unsupervised two-step sparse transfer learning algorithm for speech diagnosis of Parkinson’s disease

Neural Computing and Applications - Speech diagnosis of Parkinson’s disease (PD) as a non-invasive and simple diagnosis method is particularly worth exploring. However, the number of samples...     

Interpretable learning based Dynamic Graph Convolutional Networks for Alzheimer’s Disease analysis

Graph Convolutional Networks (GCNs) are widely applied in classification tasks by aggregating the neighborhood information of each sample to output robust node embedding. However, conventional GCN methods do not update the graph during the training process so that their effectiveness is always influenced by the quality of the input graph. Moreover, previous GCN methods lack the interpretability to limit their real applications. In this paper, a novel personalized diagnosis technique is proposed for early Alzheimer’s Disease (AD) diagnosis via coupling interpretable feature learning with dynamic graph learning into the GCN architecture. Specifically, the module of interpretable feature learning selects informative features to provide interpretability for disease diagnosis and abandons redundant features to capture inherent correlation of data points. The module of dynamic graph learning adjusts the neighborhood relationship of every data point to output robust node embedding as well as the correlations of all data points to refine the classifier. The GCN module outputs diagnosis results based on the learned inherent graph structure. All three modules are jointly optimized to perform reliable disease diagnosis at an individual level. Experiments demonstrate that our method outputs competitive diagnosis performance as well as provide interpretability for personalized disease diagnosis.     

Application of Erlang’s formula for non-Poisson flows

An improved method of equivalent random traffic (ERT) has been developed, thus making it possible to compute the parameters of splitting flows of calls, which makes it different from the classical ERT method operating only with combined flows of calls. The numerical analysis indicated that this method has an increased accuracy for independent flows. For dependent flows, the approximation’s accuracy is reduced, and the correlation between them should be taken into account.     

Advancing understanding using Nonaka’s model of knowledge creation and problem-based learning

Nonaka’s model of knowledge creation can provide guidance for designing learning environments and activities. However, Bereiter is critical of the model because it does not address whether understanding is deepened in the process of socialization, externalization, combination and internalization. To address this issue of understanding, this paper proposed a framework that synthesizes the basic phases of problem-based learning with Nonaka’s model. This paper reports on a study investigating if a course designed based on this authentic framework can help to stimulate knowledge creation based on deepening understanding. Several types of data were collected in this design-based research, namely: reflections by the participants and instructor; group discussions; student-created artifacts; and documents, records and artifacts that reflect the overall design of the course. The findings suggest that the participants demonstrated advancing understanding amidst knowledge creating conditions and processes consistent with Nonaka’s model. Other key implications are also discussed.     

Student teachers’ satisfaction for blended learning via Edmodo learning management system

In a digitally driven world, behaviours of future teachers for blended learning (both face-to-face and on-line classes) need to be examined. This study serves three purposes. The first is to examine student teachers’ preferences for Community-of-Inquiry model-driven blended learning via Edmodo. Second, predicting student satisfaction on b-learning from a combination of four variables (gender, having internet access, using the internet for information access, and previous experience in on-line learning) was questioned. And third, b-learning orientations of participants were investigated. One of the mixed methods, the concurrent triangulation design was employed in which both qualitative and quantitative methods were applied. The study group included 135 freshmen and junior students (29 males and 106 females) from a western Turkish educational faculty. The findings for the first question indicated that 70.4% of student teachers prefer b-learning. For the second, 15% of the variance in satisfaction on b-learning was explained by the proposed model with a medium effect. And for the third, the qualitative findings were discussed under Perceived Usefulness (PU) and Perceived Uselessness and Unease-of-Use (PU-UU) themes. Although less than a quarter of participants found b-learning useless, most held positive notions for b-learning practices via Edmodo.     

Enhancing the learning effect of virtual reality 3D modeling: a new model of learner’s design collaboration and a comparison of its field system usability

Universal Access in the Information Society - Knowledge of 3D modeling design is crucial for industrial designers. Therefore, 3D modeling learning is a crucial subject in industrial design...     

LSM Algorithm for Pricing American Option Under Heston–Hull–White’s Stochastic Volatility Model

In this paper, we present American option pricing under Heston–Hull–White’s stochastic volatility and stochastic interest rate model. To do this, we first discretize the stochastic processes with Euler discretization scheme. Then, we price American option by using least-squares Monte Carlo algorithm. We also compare the numerical results of our model with the Heston-CIR model. Finally, numerical results show the efficiency of the proposed algorithm for pricing American option under the Heston–Hull–White model.     

Algorithm for Constructing an Analog of Plan’s Formula

We present an algorithm for constructing an analog of Plan’s formula, which is essential in obtaining a functional relation to the classical Riemann zeta-function. The algorithm is implemented in the Maple computer algebra system.     

Convergence of Craig’s method for linear algebraic systems

Craig??s iterative method is designed to solve linear algebraic systems with an asymmetric (or even rectangular) matrix. A simple representation is constructed for the method. Test examples are used to study iterative convergence and compare the method with the conjugate gradient method. Although round-off errors in Craig??s method proved to slow down iterative convergence significantly, they allowed attaining high accuracy (if the matrix was well-conditioned). An efficient iteration-stopping criterion is found.