Search

Computer Vision (CV) has become an essential field in Artificial Intelligence applications. Object detection and recognition (ODR) is one of the fundamental tasks of computer vision implementations. However, developing an efficient ODR model is still a significant problem. The model's execution time and speed are the most critical features during the inference or detection and recognition process, which need to be improved using the latest object detection architectures. In this paper, the handwritten detection and recognition (HDR) model is developed based on previously known algorithms with their efficiency, such as Faster R-CNN and YOLOv4 in the first hand. On the other hand, two new models capable of detecting and recognizing handwritten digits using the latest ODR algorithm are proposed, one based on the latest YOLO family architecture (YOLOv5-HDR) with high speed and accuracy and the other using the transformers architecture (DETR). To the best of our knowledge, this is the first study to achieve a details comparison between YOLOv5 and transformers-based models in handwritten digit detection. Finally, the detailed performance analysis achieved by the paper proves that the YOLOv4-based model achieved the testing inference 13% faster than Faster R-CNN. However, the proposed YOLOv5-based model outperformed the YOLOv4 and the transformers-based one as it increased the testing execution time 25% faster than the YOLOv4, three times faster than the DETR model. A further adversarial attack test has been conducted to ensure the robust performance of the proposed model. Furthermore, numerical experiment results and their analyses demonstrate the robustness and effectiveness of the proposed YOLOv5-based model being the most stable for handwritten digit detection and recognition tasks. [ABSTRACT FROM AUTHOR]

This paper proposes a novel algorithm that combines symbolic execution and data flow testing to generate test cases satisfying multiple coverage criteria of critical software applications. The coverage criteria considered are data flow coverage as the primary criterion, software safety requirements, and equivalence partitioning as sub-criteria. black The characteristics of the subjects used for the study include high-precision floating-point computation and iterative programs. The work proposes an algorithm that aids the tester in automated test data generation, satisfying multiple coverage criteria for critical software. The algorithm adapts itself and selects different heuristics based on program characteristics. The algorithm has an intelligent agent as its decision support system to accomplish this adaptability. Intelligent agent uses the knowledge base to select different low-level heuristics based on the current state of the problem instance during each generation of genetic algorithm execution. The knowledge base mimics the expert's decision in choosing the appropriate heuristics. black The algorithm outperforms by accomplishing 100% data flow coverage for all subjects. In contrast, the simple genetic algorithm, random testing and a hyper-heuristic algorithm could accomplish a maximum of 83%, 67% and 76.7%, respectively, for the subject program with high complexity. black The proposed algorithm covers other criteria, namely equivalence partition coverage and software safety requirements, with fewer iterations. black The results reveal that test cases generated by the proposed algorithm are also effective in fault detection, with 87.2% of mutants killed when compared to a maximum of 76.4% of mutants killed for the complex subject with test cases of other methods. [ABSTRACT FROM AUTHOR]

There has been a lot of research on edge-computing task offloading in deep reinforcement learning (DRL). Deep reinforcement learning is one of the important algorithms in the current AI field, but there is still room for improvement in the time cost and adaptive correction ability of the algorithm. This paper studies the application of DRL algorithms in edge-computing task offloading, and its key innovation is to propose an MADRLCO algorithm, which is based on the design idea of the Actor–Critic framework, uses the DNN model to act as an Actor, and can more accurately locate the initial decision through iterative training, and use the LSTM model to optimize the Critic, which can be more accurate. The optimal decision can be located in a short period of time. The main work of this paper is divided into three parts: (1) The AC algorithm of the Actor–Critic framework in DRL is proposed to be applied to edge-computing task offloading. (2) To address the weak generalization ability of the basic version of the Actor–Critic algorithm in multi-objective optimization, the sequential quantitative correction and adaptive correction parameter K method are used to optimize the Critic frame, thereby improving the generalization ability of the model in multi-objective decision-making and improving the rationality of decision-making results. (3) Aiming at the problem of large time cost in the critical framework of the model, a search algorithm for resource allocation-related parameters based on the time-series prediction method is proposed (time-series forecasting is a research branch of pattern recognition), which reduces the time overhead of the algorithm and improves the adaptive correction capability of the model. The algorithm in this paper can adapt to not only the time-varying network channel state, but also the time-varying number of device connections. Finally, it is proved by experiments that compared with the DRL calculation offloading algorithm based on DNN plus binary search, the MADRLCO algorithm reduces the model training time by 66.27%, and in the environment of the time-varying number of devices in the metasystem, the average model average. The standard calculation rate is 0.0403 higher than that of the current optimal algorithm. [ABSTRACT FROM AUTHOR]

In this paper, the artificial intelligence control algorithm for steering robot of steering wheel is studied. The steering movement of wheeled soccer robot is controlled by artificial intelligence control algorithm, and the steering movement is modeled and simulated. Firstly, the characteristics of artificial neurons are simulated and a similar control model is constructed to complete the simulation of football. The artificial intelligence control algorithm has a dynamic feedback item compared with the traditional intelligent model, which has a better effect on the steering control of the wheeled soccer robot. In this paper, artificial intelligence control algorithm is used to optimize the parameters of artificial intelligence control algorithm, and the output of control signal of each steering part of wheeled soccer robot is simulated in the experiment, and the control of the steering action of wheeled soccer robot by artificial intelligence control algorithm is verified by experiments. Then the artificial intelligence control algorithm forms the connection structure. This method provides a good reference for steering control of wheeled soccer robots. [ABSTRACT FROM AUTHOR]

As short video social apps develop rapidly, feed has become the main approach or algorithm to present recommendation content to users in such apps. There are big differences in the way that video apps make use of feed flow based on artificial intelligence algorithm. Two kinds of short video social apps including DouYin and KuaiShou are studied with a user experiment in this paper. Several indicators are established to quantify the user experience differences of these two apps. The results are analyzed with correlation analysis to find out the relationship between user experience performance and content presentation mode of feed flow. The differences found from the results are explained from the perspectives of user cognition and behavior. [ABSTRACT FROM AUTHOR]

With the advent of the era of big data, people are eager to extract valuable knowledge from the rapidly expanding data, so that they can more effectively use these massive storage data. The traditional data processing technology can only achieve basic functions such as data query and statistics, and cannot achieve the goal of extracting the knowledge existing in the data to predict the future trend. Therefore, along with the rapid development of database technology and the rapid improvement of computer's computing power, data mining (DM) came into existence. Research on DM algorithms includes knowledge of various fields such as database, statistics, pattern recognition and artificial intelligence. Pattern recognition mainly extracts features of known data samples. The DM algorithm using pattern recognition technology is a better method to obtain effective information from massive data, thus providing decision support, and has a good application prospect. Support vector machine (SVM) is a new pattern recognition algorithm proposed in recent years, which avoids dimension disaster by dimensioning and linearization. Based on this, this paper studies the DM algorithm based on pattern recognition, and proposes a DM algorithm based on SVM. The algorithm divides the vector of the SV set into two different types and iterates through multiple iterations to obtain a classifier that converges to the final result. Finally, through the cross-validation simulation experiment, the results show that the DM algorithm based on pattern recognition can effectively reduce the training time and solve the mining problem of massive data. The results show that the algorithm has certain rationality and feasibility. [ABSTRACT FROM AUTHOR]

Recent trends in educational technology focus on designing systems that can support students while learning complex psychomotor skills, such as those required when practicing sports and martial arts, dancing or playing a musical instrument. In this context, artificial intelligence can be key to personalize the development of these psychomotor skills by enabling the provision of effective feedback when the instructor is not present, or scaling up to a larger pool of students the feedback that an instructor would typically provide one-on-one. This paper presents the modeling of human motion gathered with inertial sensors aimed to offer a personalized support to students when learning complex psychomotor skills. In particular, when comparing learner data with those of an expert during the psychomotor learning process, artificial intelligence algorithms can allow to: (i) recognize specific motion learning units and (ii) assess learning performance in a motion unit. However, it seems that this field is still emerging, since when reviewed systematically, search results hardly included the motion modeling with artificial intelligence techniques of complex human activities measured with inertial sensors. [ABSTRACT FROM AUTHOR]

Several propositional satisfiability (SAT) based encodings have been proposed to deal with various data mining problems including itemset and sequence mining problems. This research issue allows to model data mining problems in a declarative way, while exploiting efficient SAT solving techniques. In this paper, we overview our contributions on the application of propositional satisfiability (SAT) to model and solve itemset mining tasks. We first present a SAT based encoding of frequent closed itemset mining task as a propositional formula whose models corresponds to the patterns to be mined. Secondly, we show that some data mining constraints can be avoided by reformulation. We illustrate this issue by reformulating the closeness constraint using the notion of minimal models. Finally, we also addressed the scalability issue, one of the most important challenge of these nice declarative framework. To this end, we proposed a complete partition based approach whose aim is to avoid encoding the whole database as a single formula. Using a partition on the set of items, our new approach leads to several propositional formulas of reasonable size. The experimental evaluation on several known datasets shows huge improvements in comparison to the direct approach without partitioning, while reducing significantly the performance gap with respect to specialized algorithms. [ABSTRACT FROM AUTHOR]

In this paper, we examine the main advances registered in the last ten years in Pattern Recognition methodologies based on graph matching and related techniques, analyzing more than 180 papers; the aim is to provide a systematic framework presenting the recent history and the current developments. This is made by introducing a categorization of graph-based techniques and reporting, for each class, the main contributions and the most outstanding research results. [ABSTRACT FROM AUTHOR]

The representation of the degree of difference between two fuzzy subsets by means of a real number has been proposed in previous papers, and it seems to be useful in some situations. However, the requirement of assigning a precise number may lead us to the loss of essential information about this difference. Thus, (crisp) divergence measures studied in previous papers may not distinguish whether the differences between two fuzzy subsets are in low or high membership degrees. In this paper we propose a way of measuring these differences by means of a fuzzy valued function which we will call fuzzy divergence measure. We formulate a list of natural axioms that these measures should satisfy. We derive additional properties from these axioms, some of them are related to the properties required to crisp divergence measures. We finish the paper by establishing a one-to-one correspondence between families of crisp and fuzzy divergence measures. This result provides us with a method to build a fuzzy divergence measure from a crisp valued one. [ABSTRACT FROM AUTHOR]

Handwritten text recognition is one of the most difficult problems in the field of pattern recognition. In this paper, we describe our efforts towards improving the performance of state-of-the-art handwriting recognition systems through the use of classifier ensembles. There are many examples of classification problems in the literature where multiple classifier systems increase the performance over single classifiers. Normally one of the two following approaches is used to create a multiple classifier system. (1) Several classifiers are developed completely independent of each other and combined in a last step. (2) Several classifiers are created out of one prototype classifier by using so-called classifier ensemble creation methods. In this paper an algorithm which combines both approaches is introduced and it is used to increase the recognition rate of a hidden Markov model (HMM) based handwritten word recognizer. [ABSTRACT FROM AUTHOR]

This paper focuses on image filtering using weighted median (WM) filters, a nonlinear filter class taking advantages of the robust order-statistic theory and capability to adapt a filter behavior for a variety of statistics related to the desired signals and the noise distributions. The main contribution of the paper is the analysis of the four WM optimization schemes, namely genetic WM optimization, non-adaptive WM optimization algorithm and adaptive WM filtering utilizing the linear and the sigmoidal approximation of the sign function. The analysis is done by extensive simulations, in which several features such as noise reduction, edge preservation, error estimation and dependence of error criteria on the degree of the impulse noise corruption, are examined. [ABSTRACT FROM AUTHOR]

This paper proposes using estimation of distribution algorithms for unsupervised learning of Bayesian networks, directly as well as within the framework of the Bayesian structural EM algorithm. Both approaches are empirically evaluated in synthetic and real data. Specifically, the evaluation in real data consists in the application of this paper's proposals to gene expression data clustering, i.e., the identification of clusters of genes with similar expression profiles across samples, for the leukemia database. The validation of the clusters of genes that are identified suggests that these may be biologically meaningful. [ABSTRACT FROM AUTHOR]

In the pickup and delivery problem with time windows (PDPTW), vehicles have to transport loads from origins to destinations respecting capacity and time constraints. In this paper, we present a two-phase method to solve the PDPTW. In the first phase, we apply a novel construction heuristics to generate an initial solution. In the second phase, a tabu search method is proposed to improve the solution. Another contribution of this paper is a strategy to generate good problem instances and benchmarking solutions for PDPTW, based on Solomon' s benchmark test cases for VRPTW. Experimental results show that our approach yields very good solutions when compared with the benchmarking solutions. [ABSTRACT FROM AUTHOR]

In this paper, we present a new approach for determining the reflectance properties of surface and recovering 3D shapes from intensity images. We determine reflectance parameters which minimize the sum squared difference of the intensity distribution between the image of a sample sphere and the calculated image. The estimated reflectance parameters provide the range data with intensity distributions. Therefore, we generate three reference images of a range sphere, which has the same diameter as that of the sample, from the same viewpoint but with different light directions. Direct matching of the object images to the references can precisely reconstruct the shape of the object. This paper uses a plate diffuse illumination to alleviate the effects of specular spike and highlights. The simulation results show that the proposed method can estimate reflectance properties of the hybrid surface, and also recover the object shape. [ABSTRACT FROM AUTHOR]

Multi-core and multi-processor architectures have predominated the domain of embedded systems permitting easy mapping of concurrent applications to such architectures. The programs, in general, are subjected to significant optimizing and parallelizing transformations, automated and also human guided, before being mapped to an architecture. Modelling parallel behaviour and formally verifying that their functionality is preserved during synthesis are challenging tasks. Untimed PRES+ models are found to be suitable for the specification of parallel behaviour. Path cover oriented equivalence checking methods have been found to be quite effective for sequential behaviour. Path construction for parallel behaviour, however, is significantly more complex than that for sequential behaviour due to all possible interleavings of the parallel operations. Identification of the path covers depends upon choosing appropriate cut-points. In this paper, the need for introducing cut-points dynamically has been underlined and a mechanism to achieve this task is proposed. Details on how to construct a path cover using dynamic cut-points is presented. [ABSTRACT FROM AUTHOR]

In this paper, we consider optimal scheduling algorithms for scientific workows with two typical structures, fork&join and tree, on a set of provisioned (virtual) machines under budget and deadline constraints in cloud computing. First, given a total budget B, by leveraging a bi-step dynamic programming technique, we propose optimal algorithms in pseudo-polynomial time for both workows with minimum scheduling length as a goal. Our algorithms are efficient if the total budget B is polynomially bounded by the number of jobs in respective workows, which is usually the common case in practice. Second, we consider the dual of this optimization problem to minimize the cost when the deadline of the computation D is fixed. We change this problem into the standard multiple-choice knapsack problem via a parallel transformation. [ABSTRACT FROM AUTHOR]

Differing from the conventional back-propagation (BP) neural networks, a novel multi-input Euler polynomial neural network, in short, MIEPNN (specifically, 4-input Euler polynomial neural network, 4IEPNN) is established and investigated in this paper. In order to achieve satisfactory performance of the established MIEPNN, a weights and structure determination (WASD) algorithm with pruning-while-growing (PWG) and twice-pruning (TP) techniques is built up for the established MIEPNN. By employing the weights direct determination (WDD) method, the WASD algorithm not only determines the optimal connecting weights between hidden layer and output layer directly, but also obtains the optimal number of hidden-layer neurons. Specifically, a sub-optimal structure is obtained via the PWG technique, then the redundant hidden-layer neurons are further pruned via the TP technique. Consequently, the optimal structure of the MIEPNN is obtained. To provide a reasonable choice in practice, several different MATLAB computing routines related to the WDD method are studied. Comparative numerical-experiment results of the 4IEPNN using these different MATLAB computing routines and the standard multi-layer perceptron (MLP) neural network further verify the superior performance and efficacy of the proposed MIEPNN equipped with the WASD algorithm including PWG and TP techniques in terms of training, testing and predicting. [ABSTRACT FROM AUTHOR]

In this paper, we proposed an edge weight method for performing a community detection on mixed scale-free networks.We use the phrase 'mixed scale-free networks' for networks where some communities have node degree that follows a power law similar to scale-free networks, while some have node degree that follows normal distribution. In this type of network, community detection algorithms that are designed for scale-free networks will have reduced accuracy because some communities do not have scale-free properties. On the other hand, algorithms that are not designed for scale-free networks will also have reduced accuracy because some communities have scale-free properties. To solve this problem, our algorithm consists of two community detection steps; one is aimed at extracting communities whose node degree follows power law distribution (scale-free), while the other one is aimed at extracting communities whose node degree follows normal distribution (non scale-free). To evaluate our method, we use NMI - Normalized Mutual Information - to measure our results on both synthetic and real-world datasets comparing with both scale-free and non scale-free community detection methods. The results show that our method outperforms all other based line methods on mixed scale-free networks. [ABSTRACT FROM AUTHOR]

Linking named entities to structured knowledge sources paves the way for state-of-the-art Web intelligence applications which assign sentiment to the correct entities, identify trends, and reveal relations between organizations, persons and products. For this purpose this paper introduces Recognyze, a named entity linking component that uses background knowledge obtained from linked data repositories, and outlines the process of transforming heterogeneous data silos within an organization into a linked enterprise data repository which draws upon popular linked open data vocabularies to foster interoperability with public data sets. The presented examples use comprehensive real-world data sets from Orell Füssli Business Information, Switzerland's largest business information provider. The linked data repository created from these data sets comprises more than nine million triples on companies, the companies' contact information, key people, products and brands. We identify the major challenges of tapping into such sources for named entity linking, and describe required data pre-processing techniques to use and integrate such data sets, with a special focus on disambiguation and ranking algorithms. Finally, we conduct a comprehensive evaluation based on business news from the New Journal of Zurich and AWP Financial News to illustrate how these techniques improve the performance of the Recognyze named entity linking component. [ABSTRACT FROM AUTHOR]

Community structure is an important topological property common to many social, biological and technological networks. First, by using the concept of the structural weight, we introduced an improved version of the betweenness algorithm of Girvan and Newman to detect communities in networks without ( intrinsic) edge weight and then extended it to networks with ( intrinsic) edge weight. The improved algorithm was tested on both artificial and real-world networks, and the results show that it can more effectively detect communities in networks both with and without ( intrinsic) edge weight. Moreover, the technique for improving the betweenness algorithm in the paper may be directly applied to other community detection algorithms. [ABSTRACT FROM AUTHOR]

In this paper, formation of diverse groups of individuals based on attributes and two diversity measures are discussed. The formation of clusters in a diversity space is described. An algorithm is given for diverse clustering based on separation in the space and not the nearness using application based diversity thresholds and number of clusters. [ABSTRACT FROM AUTHOR]

In protein tertiary structure prediction, assessing the quality of predicted models is an essential task. Over the past years, many methods have been proposed for the protein model quality assessment (QA) and selection problem. Despite significant advances, the discerning power of current methods is still unsatisfactory. In this paper, we propose two new algorithms, CC-Select and MDS-QA, based on multidimensional scaling and k-means clustering. For the model selection problem, CC-Select combines consensus with clustering techniques to select the best models from a given pool. Given a set of predicted models, CC-Select first calculates a consensus score for each structure based on its average pairwise structural similarity to other models. Then, similar structures are grouped into clusters using multidimensional scaling and clustering algorithms. In each cluster, the one with the highest consensus score is selected as a candidate model. For the QA problem, MDS-QA combines single-model scoring functions with consensus to determine more accurate assessment score for every model in a given pool. Using extensive benchmark sets of a large collection of predicted models, we compare the two algorithms with existing state-of-the-art quality assessment methods and show significant improvement. [ABSTRACT FROM AUTHOR]

The quantron is a hybrid neuron model related to perceptrons and spiking neurons. The activation of the quantron is determined by the maximum of a sum of input signals, which is difficult to use in classical learning algorithms. Thus, training the quantron to solve classification problems requires heuristic methods such as direct search. In this paper, we present an approximation of the quantron trainable by gradient search. We show this approximation improves the classification performance of direct search solutions. We also compare the quantron and the perceptron's performance in solving the IRIS classification problem. [ABSTRACT FROM AUTHOR]

In this paper we introduce a new cardinality constraint: Ordered Distribute. Given a set of variables, this constraint limits for each value v the number of times v or any value greater than v is taken. It extends the global cardinality constraint, that constrains only the number of times a value v is taken by a set of variables and does not consider at the same time the occurrences of all the values greater than v. We design an algorithm for achieving generalized arc-consistency on Ordered Distribute, with a time complexity linear in the sum of the number of variables and the number of values in the union of their domains. In addition, we give some experiments showing the advantage of this new constraint for problems where values represent levels whose overrunning has to be under control. Finally, we present three extensions of our constraint that can be particularly useful in practice. [ABSTRACT FROM AUTHOR]

The development of satellite image acquisition tools helped improving the extraction of information about natural scenes. In the proposed approach, we try to minimize imperfections accompanying the image interpretation process and to maximize useful information extracted from these images through the use of blind source separation (BSS) and fusion methods. In order to extract maximum information from multi-sensor images, we propose to use three algorithms of BSS that are FAST- ICA2D, JADE2D, and SOBI2D. Then by employing various fusion methods such as the probability, possibility, and evidence methods we can minimize both imprecision and uncertainty. In this paper, we propose a hybrid approach based on five main steps. The first step is to apply the three BSS algorithms to the satellites images; it results in obtaining a set of image sources representing each a facet of the land cover. A second step is to choose the image having the maximum of kurtosis and negentropy. After the BSS evaluation, we proceed to the training step using neural networks. The goal of this step is to provide learning regions which are useful for the fusion step. The next step consists in choosing the best adapted fusion method for the selected source images through a case-based reasoning (CBR) module. If the CBR module does not contain a case similar to the one we are seeking, we proceed to apply the three fusion methods. The evaluation of fusion methods is a necessary step for the learning process of our CBR. [ABSTRACT FROM AUTHOR]

This paper explores the phase transitions of the EXPSPACE-complete problems, which mainly focus on the conformant planning problems. The research presents two conformant planning algorithms-CONFORMANT PLAN-NONEXISTENCE algorithm and CONFORMANT PLAN-EXISTENCE algorithm. By analyzing the features of the two algorithms, the phase transition area of the conformant planning problems is obtained. If the number of the operators isn't greater than θub, the CONFORMANT PLAN-NONEXISTENCE algorithm can prove that nearly all the conformant planning instances have no solution. If the number of the operators isn't lower than θlb, the CONFORMANT PLAN-EXISTENCE algorithm can prove that nearly all the conformant planning instances have solutions. The results of the experiments show that there exist phase transitions from a region where almost all the conformant planning instances have no solution to a region where almost all the conformant planning instances have solutions. [ABSTRACT FROM AUTHOR]

Game playing is commonly cited in debates concerning human versus machine intelligence, and Chess is often at the center of such debates. However, the role of Chess in delineating the difference between natural and artificial intelligence has been significantly diminished since a World Chess Champion lost in a tournament against a computer. Computer brute force is regularly blamed for the human defeat. This paper proposes a Quantum Chess Board in an attempt to bring back some equilibrium, putting humans and computers on an ostensibly equal footing when faced with the uncertainties of quantum physics. [ABSTRACT FROM AUTHOR]

Abstract-Cardiac arrhythmia is one of the major causes of human death, and most of the time it cannot be predicted well in advance at the right time. Computational intelligence algorithms can help in extracting the hidden patterns of biological datasets. This paper explores the use of advanced and intelligent computational algorithms for automated detection, classification and clustering of cardiac arrhythmia (CA). Application of Fuzzy C-Mean and Extended Fuzzy C-Mean method to the arrhythmia dataset (165 normal healthy and 138 with CA) demonstrated their good CA classification capabilities. Fuzzy C Mean algorithm was able to classify the two group of data set with an overall accuracy of 97.2% [sensitivity 96.4%, specificity 98.12% and area under the receiver operating curve (AUC-ROC = 0.963)]. The classification accuracy improved significantly when GK-based extended Fuzzy was employed, and an overall accuracy of 99.14% was achieved (sensitivity 97.11%, specificity 99.18% and AUC-ROC = 0.995). These accuracy results were respectively, 19.02%, 7%, 9.14% and 11.06% higher when compared to multi-input single layer perceptron (SLP), feed forward back propagation (FFBP), self organizing maps (SOM) and support vector machine (SVM). The performance measures of fuzzy techniques were found to be better if a Principal Component Analysis (PCA) technique was used to preprocess the arrhythmia datasets. [ABSTRACT FROM AUTHOR]

This paper presents a repair algorithm for the design of a Radial Basis Function (RBF) neural network. The proposed repair RBF (RRBF) algorithm starts from a single prototype randomly initialized in the feature space. The algorithm has two main phases: an architecture learning phase and a parameter adjustment phase. The architecture learning phase uses a repair strategy based on a sensitivity analysis (SA) of the network's output to judge when and where hidden nodes should be added to the network. New nodes are added to repair the architecture when the prototype does not meet the requirements. The parameter adjustment phase uses an adjustment strategy where the capabilities of the network are improved by modifying all the weights. The algorithm is applied to two application areas: approximating a non-linear function, and modeling the key parameter, chemical oxygen demand (COD) used in the waste water treatment process. The results of simulation show that the algorithm provides an efficient solution to both problems. [ABSTRACT FROM AUTHOR]

Unlike feedforward neural networks (FFNN) which can act as universal function approximators, recursive, or recurrent, neural networks can act as universal approximators for multi-valued functions. In this paper, a real time recursive backpropagation (RTRBP) algorithm in a vector matrix form is developed for a two-layer globally recursive neural network that has multiple delays in its feedback path. This algorithm has been evaluated on two GRNNs that approximate both an analytic and nonanalytic periodic multi-valued function that a feedforward neural network is not capable of approximating. [ABSTRACT FROM AUTHOR]

This paper presents the graffiti interpretation of one-stroke handwritten digits (0 to 9) and commands (backspace, carriage return and space) using support vector machines (SVMs). A number of SVM-based graffiti interpreters are proposed for the recognition of graffiti. The performance of the proposed SVM-based graffiti interpreters subject to various kernel functions and parameters are investigated. Simulation and experimental results are presented to show the applicability and the merits of various graffiti interpreters. [ABSTRACT FROM AUTHOR]

This paper investigates robust exponential stability for discrete-time recurrent neural networks with both time-varying delay (0 ≤ τm ≤ τ(k) ≤ τM) and distributed one. Through partitioning delay intervals [0,τm] and [τm,τM], respectively, and choosing an augmented Lyapunov-Krasovskii functional, the delay-dependent sufficient conditions are obtained by using free-weighting matrix and convex combination methods. These criteria are presented in terms of linear matrix inequalities (LMIs) and their feasibility can be easily checked by resorting to LMI in Matlab Toolbox in Ref. 1. The activation functions are not required to be differentiable or strictly monotonic, which generalizes those earlier forms. As an extension, we further consider the robust stability of discrete-time delayed Cohen-Grossberg neural networks. Finally, the effectiveness of the proposed results is further illustrated by three numerical examples in comparison with the reported ones. [ABSTRACT FROM AUTHOR]

The K-means algorithm is very popular in the machine learning community due to its inherent simplicity. However, in its basic form, it is not suitable for use in problems which contain periodic attributes, such as oscillator phase, hour of day or directional heading. A commonly used technique of trigonometrically encoding periodic input attributes to artificially generate the required topology introduces a systematic error. In this paper, a metric which induces a conceptually correct topology for periodic attributes is embedded into the K-means algorithm. This requires solving a non-convex minimization problem in the maximization step. Results of numerical experiments comparing the proposed algorithm to K-means with trigonometric encoding on synthetically generated data are reported. The advantage of using the proposed K-means algorithm is also shown on a real example using gas load data to build simple predictive models. [ABSTRACT FROM AUTHOR]

This paper presents a new approach, known as symmetry axis based feature extraction and recognition (SAFER), for recognizing objects under translation, rotation and scaling. Unlike most previous invariant object recognition (IOR) systems, SAFER puts emphasis on both simplicity and accuracy of the recognition system. To achieve simplicity, it uses simple formulae for extracting invariant features from an object. The scheme used in feature extraction is based on the axis of symmetry and angles of concentric circles drawn around the object. SAFER divides the extracted features into a number of groups based on their similarity. To improve the recognition performance, SAFER uses a number of neural networks (NNs) instead of single NN are used for training and recognition of extracted features. The new approach, SAFER, has been tested on two of real world problems i.e., English characters with two different fonts and images of different shapes. The experimental results show that SAFER can produce good recognition performance in comparison with other algorithms. [ABSTRACT FROM AUTHOR]

This paper presents a semi-automatic approach for confounding-aware subgroup discovery: Confounding essentially disturbs the measured effect of an association between variables due to the influence of other parameters that were not considered. The proposed method is embedded into a general subgroup discovery approach, and provides the means for detecting potentially confounded subgroup patterns, other unconfounded relations, and/or patterns that are affected by effect-modification. Since there is no purely automatic test for confounding, the discovered relations are presented to the user in a semi-automatic approach. Furthermore, we utilize (causal) domain knowledge for improving the results of the algorithm, since confounding is itself a causal concept. The applicability and benefit of the presented technique is illustrated by real-world examples from a case-study in the medical domain. [ABSTRACT FROM AUTHOR]

The ability of an algorithm to accurately estimate the parameters of the geometric transformation which aligns two image profiles even in the presence of photometric distortions can be considered as a basic requirement in many computer vision applications. Projective transformations constitute a general class which includes as special cases the affine, as well as the metric subclasses of transformations. In this paper the applicability of a recently proposed iterative algorithm, which uses the Enhanced Correlation Coefficient as a performance criterion, in the projective image registration problem is investigated. The main theoretical results concerning the proposed iterative algorithm are presented. Furthermore, the performance of the iterative algorithm in the presence of nonlinear photometric distortions is compared against the leading Lucas-Kanade algorithm and its simultaneous inverse compositional variant with the help of a series of experiments involving strong or weak geometric deformations, ideal and noisy conditions and even over-modelling of the warping process. Although under ideal conditions the proposed algorithm and simultaneous inverse compositional algorithm exhibit a similar performance and both outperform the Lucas-Kanade algorithm, under noisy conditions the proposed algorithm outperforms the other algorithms in convergence speed and accuracy, and exhibits robustness against photometric distortions. [ABSTRACT FROM AUTHOR]

Having more hidden units than necessary can produce a neural network that has a poor generalization. This paper proposes a new algorithm for pruning unnecessary hidden units away from the single-hidden layer feedforward neural networks, resulting in a Spartan network. Our approach is simple and easy to implement, yet produces a very good result. The idea is to train the network until it begins to lose its generalization. Then the algorithm measures the sensitivity and automatically prunes away the most irrelevant unit. We define this sensitivity as the absolute difference between the desirable output and the output of the pruned network. Unlike other pruning methods, our algorithm is distinct in calculating the sensitivity from the validation set, instead of the training set, without increasing the asymptotic time complexity of the back-propagation algorithm. In addition, for a classification problem, we raise a point that the sensitivities of some well-known pruning algorithms may still underestimate the irrelevance of hidden unit even though the validation set is used in measuring the sensitivity. We resolve this problem by considering the number of misclassified patterns as the main concern. The Spartan simplicity algorithm is applied to three artificial and seven standard benchmarks. In most problems, the algorithm can produce a compact-sized network with high generalization ability in comparison with other pruning algorithms. [ABSTRACT FROM AUTHOR]

HAM (Hopfield Associative Memory) and BAM (Bidirectinal Associative Memory) are representative associative memories by neural networks. The storage capacity by the Hebb rule, which is often used, is extremely low. In order to improve it, some learning methods, for example, pseudo-inverse matrix learning and gradient descent learning, have been introduced. Oh introduced pseudo-relaxation learning algorithm to HAM and BAM. In order to accelerate it, Hattori proposed quick learning. Noest proposed CAM (Complex-valued Associative Memory), which is complex-valued HAM. The storage capacity of CAM by the Hebb rule is also extremely low. Pseudo-inverse matrix learning and gradient descent learning have already been generalized to CAM. In this paper, we apply pseudo-relaxation learning algorithm to CAM in order to improve the capacity. [ABSTRACT FROM AUTHOR]

The present manuscript aims at illustrating fundamental challenges and solutions arising in the design of learning theories by optimization on manifolds in the context of complex-valued neural systems. The special case of a unitary unimodular group of matrices is dealt with. The unitary unimodular group under analysis is a low dimensional and easy-to-handle matrix group. Notwithstanding, it exhibits a rich geometrical structure and gives rise to interesting speculations about methods to solve optimization problems on manifolds. Also, its low dimension allows us to treat most of the quantities involved in computation in closed form as well as to render them in graphical format. Some numerical experiments are presented and discussed within the paper, which deal with complex-valued independent component analysis. [ABSTRACT FROM AUTHOR]

This paper introduces an Independent Component Analysis (ICA) approach to the separation of nonlinear mixtures in the complex domain. Source separation is performed by a complex INFOMAX approach. The neural network which realizes the separation employs the so called "Mirror Model" and is based on adaptive activation functions, whose shape is properly modified during learning. Nonlinear functions involved in the processing of complex signals are realized by pairs of spline neurons called "splitting functions", working on the real and the imaginary part of the signal respectively. Theoretical proof of existence and uniqueness of the solution under proper assumptions is also provided. In particular a simple adaptation algorithm is derived and some experimental results that demonstrate the effectiveness of the proposed solution are shown. [ABSTRACT FROM AUTHOR]

Formula extraction is the first stage of optical formula recognition which converts printed scientific documents into their corresponding electronic format. So far, little research has been done in this area. In this paper, an improved method using fuzzy classification and irregularity rate feature is proposed to separate formulas from texts in the printed documents. Firstly, according to a statistical threshold of distance, connected components are extracted and merged to form the areas of characters and lines. Secondly, the isolated formulas are extracted based on the line features. Finally, the formula symbols in the rest lines are labeled using irregularity degree, and the embedded formulas are located by extending kernel symbols using the propagation of context. In these steps, fuzzy classification algorithm and irregularity degree feature are introduced to solve the problems existing in traditional methods and improve the extracting accuracy. The experimental results show that the method is of great significance in both theory and practice. [ABSTRACT FROM AUTHOR]

This paper introduces a stochastic graph-based algorithm, called OutRank, for detecting outliers in data. We consider two approaches for constructing a graph representation of the data, based on the object similarity and number of shared neighbors between objects. The heart of this approach is the Markov chain model that is built upon this graph, which assigns an outlier score to each object. Using this framework, we show that our algorithm is more robust than the existing outlier detection schemes and can effectively address the inherent problems of such schemes. Empirical studies conducted on both real and synthetic data sets show that significant improvements in detection rate and false alarm rate are achieved using the proposed framework. [ABSTRACT FROM AUTHOR]

The goal of spatial co-location pattern mining is to find subsets of spatial features frequently located together in spatial proximity. Example co-location patterns include services requested frequently and located together from mobile devices (e.g., PDAs and cellular phones) and symbiotic species in ecology (e.g., Nile crocodile and Egyptian plover). Spatial clustering groups similar spatial objects together. Reusing research results in clustering, e.g. algorithms and visualization techniques, by mapping co-location mining problem into a clustering problem would be very useful. However, directly clustering spatial objects from various spatial features may not yield well-defined co-location patterns. Clustering spatial objects in each layer followed by overlaying the layers of clusters may not applicable to many application domains where the spatial objects in some layers are not clustered. In this paper, we propose a new approach to the problem of mining co-location patterns using clustering techniques. First, we propose a novel framework for co-location mining using clustering techniques. We show that the proximity of two spatial features can be captured by summarizing their spatial objects embedded in a continuous space via various techniques. We define the desired properties of proximity functions compared to similarity functions in clustering. Furthermore, we summarize the properties of a list of popular spatial statistical measures as the proximity functions. Finally, we show that clustering techniques can be applied to reveal the rich structure formed by co-located spatial features. A case study on real datasets shows that our method is effective for mining co-locations from large spatial datasets. [ABSTRACT FROM AUTHOR]

Standard tabu search methods are based on the complete exploration of current solution neighborhood. However, for some problems due to the neighborhood size or to the fitness evaluation complexity, the total exploration of the neighborhood is impractical. The main purpose of this paper is to propose a local search method with no enumeration procedure. In other words, a single solution is examined at each iteration and retained for the future iterations. The idea is to randomly select one solution among a sub-set of the neighborhood of the current one. An adaptive exploration of neighborhood, using extension and restriction mechanisms represented by loop detection and tabu list structure, determines this sub-set. This approach is applied to the K-coloring problem and evaluated on standard benchmarks like DIMACS. The objective is to show how a generic method, without full neighborhood exploration, degradation control and problem-oriented operators, provides a very competitive results comparing to the best dedicated algorithms for graph coloring problems published in the literature. [ABSTRACT FROM AUTHOR]

Precedence constraints specify that an activity must finish before another activity starts and hence such constraints play a crucial role in planning and scheduling problems. Many real-life problems also include dependency constraints expressing logical relations between the activities – for example, an activity requires presence of another activity in the plan. For such problems a typical objective is a maximization of the number of activities satisfying the precedence and dependency constraints. In the paper we propose new incremental filtering rules integrating propagation through both precedence and dependency constraints. We also propose a new filtering rule using the information about the requested number of activities in the plan. We demonstrate efficiency of the proposed rules on log-based reconciliation problems and min-cutset problems. [ABSTRACT FROM AUTHOR]

A method for arbitrary surface reconstruction from 3D large scattered points is proposed in this paper. According to the properties of 3D points, e.g. the non-uniform distribution and unknown topology, an implicit surface model is adopted based on radial basis functions network. And because of the property of locality of radial basis function, the method is fast and robust in surface reconstruction. Furthermore, an adapted K-Means algorithm is used to reduce reconstruction centers. For features completeness, two effective methods are introduced to extract the feature points before the adapted K-Means algorithm. Experiment results show that the presented approach is a good solution for reconstruction from 3D large scattered points. [ABSTRACT FROM AUTHOR]

In this paper we propose a new word sense disambiguation method called Multi-engine Collaborative Bootstrapping (MCB) that combines different types of corpora and also uses two languages for bootstrapping. MCB uses the bilingual bootstrapping as its core algorithm that leading to incremental knowledge acquisition. The EM model is applied to train parameters in a base learner. The feature translation model is improved by semantic correlation estimation. In addition we use multi-engine selection to produce qualified starting seeds from parallel corpora and monolingual corpora. Those seeds that are generated through unsupervised machine learning approaches can also ensure bootstrapping effectiveness in contrast with manually selected seeds in spite of their different selection mechanisms. Experimental results prove the effectiveness of MCB. Some factors including feature space and starting seed number are concerned involved in our experiments because the EM algorithm is sensitive to starting values. Limitation of resources is also a concern. [ABSTRACT FROM AUTHOR]

A theory, called the Kinematic Theory of Rapid Human Movement, was proposed a few years ago to analyze rapid human movements, called the Kinematic Theory of Rapid Human Movements, based on a delta-lognormal equation that globally describes the basic properties of the velocity profiles of an end-effector using seven parameters. This realistic model has been very useful for proposing original solutions to various pattern recognition problems (signature segmentation and verification, handwriting analysis and synthesis, etc.). Most of these applications rely on the use of an efficient algorithm to extract the delta-lognormal parameters from real data with the best possible fit. In this paper, we compare two such algorithms: a deterministic one, based on nonlinear regression, and a Breeder Genetic algorithm. The performance of these two algorithms and of their combinations are compared using the same artificial database, composed of analytical delta-lognormal profiles and their noisy versions (20 dB SNR). In the free-noise case, the analysis of the experimental results shows that the deterministic approach leads to better results than the evolutionary one, while under the extremely noisy conditions selected, the evolutionary approach seems to be less sensitive to noise, but is nevertheless less successful than the deterministic search. [ABSTRACT FROM AUTHOR]

In this paper we present an alternative strategy to define zoning for handwriting recognition, which is based on nonsymmetrical perceptual zoning. The idea is to extract some knowledge from the confusion matrices in order to make the zoning process less empirical. The feature set considered in this work is based on concavities/convexities deficiencies, which are obtained by labeling the background pixels of the input image. To better assess the nonsymmetrical zoning we carried out experiments using four different zonings strategies. Experiments show that the nonsymmetrical zoning could be considered as a tool to build more reliable handwriting recognition systems. [ABSTRACT FROM AUTHOR]