Self-Adjustable Neural Network for speech recognition

Neural networks with fixed input length are not able to train and test data with variable lengths in one network size. This issue is very crucial when the neural networks need to deal with signals of variable lengths, such as speech. Though various methods have been proposed in segmentation and feature extraction to deal with variable lengths of the data, the size of the input data to the neural networks still has to be fixed. A novel Self-Adjustable Neural Network (SANN) is presented in this paper, to enable the network to adjust itself according to different data input sizes. The proposed method is applied to the speech recognition of Malay vowels and TIMIT isolated words. SANN is benchmarked against the standard and state-of-the-art recogniser, Hidden Markov Model (HMM). The results showed that SANN was better than HMM in recognizing the Malay vowels. However, HMM outperformed SANN in recognising the TIMIT isolated words.     

A multi-objective facility location model with batch arrivals: two parameter-tuned meta-heuristic algorithms

Many research works in mathematical modeling of the facility location problem have been carried out in discrete and continuous optimization area to obtain the optimum number of required facilities along with the relevant allocation processes. This paper proposes a new multi-objective facility-location problem within the batch arrival queuing framework. Three objective functions are considered: (I) minimizing the weighted sum of the waiting and the traveling times, (II) minimizing the maximum idle time pertinent to each facility, and (III) minimizing the total cost associated with the opened facilities. In this way, the best combination of the facilities is determined in the sense of economical, equilibrium, and enhancing service quality viewpoints. As the model is shown strongly NP-hard, two meta-heuristic algorithms, namely genetic algorithm (GA) and simulated annealing (SA) are proposed to solve the model. Not only new coding is developed in these solution algorithms, but also a random search algorithm is proposed to justify the efficiency of both algorithms. Since the solution-quality of all meta-heuristic algorithms severely depends on their parameters, design of experiments and response surface methodologies have been utilized to calibrate the parameters of both algorithms. Finally, computational results obtained by implementing both algorithms on several problems of different sizes demonstrate the performances of the proposed methodology.     

Modeling the compressive strength of geopolymeric binders by gene expression programming-GEP

GEP has been employed in this work to model the compressive strength of different types of geopolymers through six different schemes. The differences between the models were in their linking functions, number of genes, chromosomes and head sizes. The curing time, Ca(OH)₂ content, the amount of superplasticizer, NaOH concentration, mold type, aluminosilicate source and H₂O/Na₂O molar ratio were the seven input parameters considered in the construction of the models to evaluate the compressive strength of geopolymers. A total number of 399 input-target pairs were collected from the literature, randomly divided into 299 and 100 sets and were trained and tested, respectively. The best performance model had 6 genes, 14 head size, 40 chromosomes and multiplication as linking function. This was shown by the absolute fraction of variance, the absolute percentage error and the root mean square error. These were of 0.9556, 2.4601 and 3.4716 for training phase, respectively and 0.9483, 2.8456 and 3.7959 for testing phase, respectively. However, another model with 7 genes, 12 head size, 30 chromosomes and addition as linking function showed suitable results with the absolute fraction of variance, the absolute percentage error and the root mean square of 0.9547, 2.5665 and 3.4360 for training phase, respectively and 0.9466, 2.8020 and 3.8047 for testing phase, respectively. These models showed that gene expression programming has a strong potential for predicting the compressive strength of different types of geopolymers in the considered range.     

Active fault detection and isolation of discrete-time linear time-varying systems: a set-membership approach

Active fault detection and isolation (AFDI) is used for detection and isolation of faults that are hidden in the normal operation because of a low excitation signal or due to the regulatory actions of the controller. In this paper, a new AFDI method based on set-membership approaches is proposed. In set-membership approaches, instead of a point-wise estimation of the states, a set-valued estimation of them is computed. If this set becomes empty the given model of the system is not consistent with the measurements. Therefore, the model is falsified. When more than one model of the system remains un-falsified, the AFDI method is used to generate an auxiliary signal that is injected into the system for detection and isolation of faults that remain otherwise hidden or non-isolated using passive FDI (PFDI) methods. Having the set-valued estimation of the states for each model, the proposed AFDI method finds an optimal input signal that guarantees FDI in a finite time horizon. The input signal is updated at each iteration in a decreasing receding horizon manner based on the set-valued estimation of the current states and un-falsified models at the current sample time. The problem is solved by a number of linear and quadratic programming problems, which result in a computationally efficient algorithm. The method is tested on a numerical example as well as on the pitch actuator of a benchmark wind turbine.     

Computation of the recoverable region and stabilisation problem in the recoverable region for discrete-time systems

This article considers discrete-time linear time-invariant systems subject to input and state constraints. It is shown that in the recoverable region (which is the largest domain of attraction that is theoretically achievable) system can be semi-globally stabilised by nonlinear feedbacks while satisfying the constraints. Moreover, the recoverable region for the given system is constructed by constructing the same, however, for a reduced-order subsystem of the given system. Such a reduction in the order or dimension of the system generally leads to a considerable simplification in the computational effort.     

Non-fragile synchronisation of mixed delayed neural networks with randomly occurring controller gain fluctuations

This study is concerned with the problem of non fragile synchronisation of mixed delayed neural networks with randomly occurring controller gain fluctuations. By using a novel mathematical approach and considering the neuron activation functions, improved delay-dependent stability results are formulated in terms of linear matrix inequalities (LMIs). An augmented new Lyapunov-Krasovskii functional (LKF) that contains double and triple integral terms is constructed to ensure the asymptotic stability of the error system which guarantees the master system synchronise with the slave system. Finally, numerical examples are provided to show the effectiveness of the proposed theoretical results.     

Line matching of wide baseline images in an affine projection space

ABSTRACT

Line matching plays an important role in vision localization and three-dimensional reconstruction of building structures. The conventional method of line matching is not effective for processing stereo images with wide baselines and large viewing angles. This paper proposes a line matching method in an affine projection space, aiming to solve the problem of change of viewing angles in aerial oblique images. Firstly, monocular image orientation can be performed through geometric structures of buildings. Secondly, according to the pose information of the camera, the affine projection matrix is obtained. The original image can be rectified as a conformal image based on this projection matrix, thereby reducing the difference in the viewing angle between images. Then, line matching is performed on the rectified images to get the matched line pairs. Finally, the inverse affine projection matrix is used to back-project the matched line pairs to the original images. The experimental results of five groups of aerial oblique images show that the matched line segments obtained by the proposed method are basically superior to those of the methods which are directly processed on the original image in terms of quantity, correctness, and efficiency.     

A new approach to satellite-based estimation of precipitation over snow cover

Current satellite-based remote-sensing approaches are largely incapable of estimating precipitation over snow cover. This note reports a proof-of-concept study of a new satellite-based approach to the estimation of precipitation over snow-covered surfaces. The method is based on the principle that precipitation can be inferred from the changes in the snow water equivalent of the snowpack. Using satellite-based snow water equivalent measurements, we derived daily precipitation amounts for the northern hemisphere for three snow-accumulation seasons, and evaluated these against independent reference datasets. The new precipitation estimates captured realistic-looking storm events over largely un-instrumented regions. However, the data are noisy and, on a seasonal scale, the amount of precipitation is believed to be underestimated. Nevertheless, current uncertainty in snow measurements, albeit large (50–100%), is still lower than direct precipitation measurements over snow (100–140%) and therefore this approach is still useful. The method will become more feasible as the quality of remotely sensed snow measurements improves.     

How We Broke the Union Code (148 Years Too Late)

Abstract

We decrypt a hitherto undeciphered Union message from the United States Civil War. Our means of cryptanalysis of the (Stager) cipher combines an appreciation of the limitations of human-generated randomness and several applications of the principle known as Ockham's Razor. We work without any codebooks and, in the process of our analysis, identify and explain the error that was made in the encryption of this letter from Porter to McClellan (about 148 years ago) that likely led to it being undecipherable—until now.     

Classroom furniture and anthropometric characteristics of Iranian high school students: Proposed dimensions based on anthropometric data

The study evaluated the potential mismatch between classroom furniture dimensions and anthropometric characteristics of 978 Iranian high school students (498 girls, 480 boys), aged 15-18 years. Nine anthropometric measurements (stature, sitting height, sitting shoulder height, popliteal height, hip breadth, elbow-seat height, buttock-popliteal length, buttock-knee length and thigh clearance) and five dimensions from the existing classroom furniture were measured and then compared together (using match criterion equations) to identify any potential mismatch between them. The results indicated a considerable mismatch between body dimensions of the students and the existing classroom furniture, with seat height (60.9%), seat width (54.7%) and desktop height (51.7%) being the furniture dimensions with a higher level of mismatch. The levels of mismatch varied between the high-school grade levels and between genders, indicating their special requirements and possible problems. The proposed dimensions of the classroom furniture more appropriate for the students were given. This additional information on students' anthropometry can be used by local furniture industries as a starting point for designing more appropriate furniture for school children, or used by schools to aid in furniture selection.