Search

In various approaches for combinatorial optimization, a problem instance is represented by a numerical vector that summarizes to some extent the actual solution state of such instance. Such representation intends to include the most relevant features related to the instance and the problem domain. The proper selection of these features has a direct impact on the performance of a hyper-heuristic. Previous approaches for hyper-heuristics have been relying on intuitive ways to determine the feature set, usually based on the domain knowledge of a particular problem. In this paper, a more general methodology for establishing an adequate problem-state representation is proposed. We chose the irregular case of the two-dimensional Bin Packing Problem (2D irregular BPP) and a GA-based hyper-heuristic model to test the methodology. As far as we know, this is the only hyper-heuristic model applied to the 2D irregular BPP and it has been successful when solving a wide range of instances. Our developed representation shows a significant improvement in performance with respect to a more conventional representation for the 2D irregular BPP. [ABSTRACT FROM AUTHOR]

This article presents a method based on the multi-objective evolutionary algorithm NSGA-II to approximate hyper-heuristics for solving irregular 2D cutting stock problems under multiple objectives. In this case, additionally to the traditional objective of minimizing the number of sheets used to fit a finite number of irregular pieces, the time required to perform the placement task is also minimized, leading to a bi-objective minimization problem with a tradeoff between the number of sheets and the time required for placing all pieces. We solve this problem using multi-objective hyper-heuristics (MOHHs), whose main idea consists of finding a set of simple heuristics which can be combined to find a general solution for a wide range of problems, where a single heuristic is applied depending on the current condition of the problem, instead of applying a unique single heuristic during the whole placement process. The MOHHs are approximated after going through a learning process by mean of the NSGA-II, which evolves combinations of condition-action rules producing at the end a set of Pareto-optimal MOHHs. We tested the approximated MMOHHs on several sets of benchmark problems, having outstanding results for most of the cases. [ABSTRACT FROM AUTHOR]

This paper introduces a hybrid algorithm that combines local search and constraint satisfaction techniques with memetic algorithms for solving Course Timetabling hard problems. These problems require assigning a set of courses to a predetermined finite number of classrooms and periods of time, complying with a complete set of hard constraints while maximizing the consistency with a set of preferences (soft constraints). The algorithm works in a three-stage sequence: first, it creates an initial population of approximations to the solution by partitioning the variables that represent the courses and solving each partition as a constraint-satisfaction problem; second, it reduces the number of remaining hard and soft constraint violations applying a memetic algorithm; and finally, it obtains a complete and fully consistent solution by locally searching around the best memetic solution. The approach produces competitive results, always getting feasible solutions with a reduced number of soft constraints inconsistencies, when compared against the methods running independently. [ABSTRACT FROM AUTHOR]

The idea behind hyper-heuristics is to discover some combination of straightforward heuristics to solve a wide range of problems. To be worthwhile, such combination should outperform the single heuristics. This paper presents a GA-based method that produces general hyper-heuristics for the dynamic variable ordering within Constraint Satisfaction Problems. The GA uses a variable-length representation, which evolves combinations of condition-action rules producing hyper-heuristics after going through a learning process which includes training and testing phases. Such hyper-heuristics, when tested with a large set of benchmark problems, produce encouraging results for most of the cases. There are instances of CSP that are harder to be solved than others, this due to the constraint and the conflict density [4]. The testebed is composed of hard problems randomly generated by an algorithm proposed by Prosser [18]. [ABSTRACT FROM AUTHOR]

This paper introduces the QFCS, a new approach to fuzzy learning classifier systems. QFCS can solve the multistep reinforcement learning problem in continuous environments and with a set of continuous vector actions. Rules in the QFCS are small fuzzy systems. QFCS uses a Q-learning algorithm to learn the mapping between inputs and outputs. This paper presents results that show that QFCS can evolve rules to represent only those parts of the input and action space where the expected values are important for making decisions. Results for the QFCS are compared with those obtained by Q-learning with a high discretization to show that the new approach converges in a way similar to how Q-learning does for one-dimension problems with an optimal solution, and for two dimensions QFCS learns suboptimal solutions while it is difficult for Q-learning to converge due to that high discretization. [ABSTRACT FROM AUTHOR]

Breast cancer is one of the main causes of death in women and early diagnosis is an important means to reduce the mortality rate. The presence of microcalcification clusters are primary indicators of early stages of malignant types of breast cancer and its detection is important to prevent the disease. This chapter presents a procedure for the classification of microcalcification clusters in mammograms using sequential difference of gaussian filters (DoG) and three evolutionary artificial neural networks (EANNs) compared against a feedforward artificial neural network (ANN) trained with backpropagation. It is shown that the use of genetic algorithms (GAs) for finding the optimal weight set for an ANN, finding an adequate initial weight set before starting a backpropagation training algorithm and designing its architecture and tuning its parameters, results mainly in improvements in over-all accuracy, sensitivity and specificity of an ANN, compared with other networks trained with simple backpropagation. [ABSTRACT FROM AUTHOR]

The presence of microcalcification clusters in digital mammograms is a primary indicator of early stages of malignant types of breast cancer and its detection is important to prevent the disease. This paper uses a procedure for the classification of microcalcification clusters in mammograms using sequential Difference of Gaussian filters (DoG) and feedforward Neural Networks (NN). Three methods using class separability, forward sequential search and genetic algorithms for feature selection are compared. We found that the use of Genetic Algorithms (GAs) for selecting the features from microcalcifications and microcalcification clusters that will be the inputs of a feedforward Neural Network (NN) results mainly in improvements in overall accuracy, sensitivity and specificity of the classification. [ABSTRACT FROM AUTHOR]

Highly dynamic environments with uncertainty make inadequate long or rigid plans, because they are frequently dismissed by the arrival or new unexpected situations. In these environments, most approaches eliminate planning altogether, and evaluate just the current situation. We are interested in on-line planning, where execution and planning are interleaved, and short plans are continuously re-evaluated. Now, the plan evaluation itself could be an important issue. We have proposed in our recent work to evaluate plans taking into account the quantity and quality of future options, not just the single best future option. In this paper we present a detailed evaluation of real-time planning performance, changing the importance given to the current situation, to the best future option, and to the set of future options respective evaluations, in the context of the simulated soccer Robocup competition. Our results show that a well-tuned combination of the mentioned factors could outperform any of them alone. [ABSTRACT FROM AUTHOR]

The idea behind hyper-heuristics is to discover some combination of straightforward heuristics to solve a wide range of problems. To be worthwhile, such combination should outperform the single heuristics. This paper presents a GA-based method that produces general hyper-heuristics that solve two-dimensional cutting stock problems. The GA uses a variable-length representation, which evolves combinations of condition-action rules producing hyper-heuristics after going through a learning process which includes training and testing phases. Such hyper-heuristics, when tested with a large set of benchmark problems, produce outstanding results (optimal and near-optimal) for most of the cases. The testebed is composed of problems used in other similar studies in the literature. Some additional instances of the testbed were randomly generated. [ABSTRACT FROM AUTHOR]

Breast cancer is one of the main causes of death in women and early diagnosis is an important means to reduce the mortality rate. The presence of microcalcification clusters are primary indicators of early stages of malignant types of breast cancer and its detection is important to prevent the disease. This paper proposes a procedure for the classification of microcalcification clusters in mammograms using sequential difference of gaussian filters (DoG) and three evolutionary artificial neural networks (EANNs) compared against a feedforward artificial neural network (ANN) trained with backpropagation. We found that the use of genetic algorithms (GAs) for finding the optimal weight set for an ANN, finding an adequate initial weight set before starting a backpropagation training algorithm and designing its architecture and tuning its parameters, results mainly in improvements in overall accuracy, sensitivity and specificity of an ANN, compared with other networks trained with simple backpropagation. [ABSTRACT FROM AUTHOR]

We propose a general affective behavior model integrated to an intelligent tutoring system with the aim of providing the students with a suitable response from a pedagogical and affective point of view. The affective behavior model integrates the information from the student cognitive state, student affective state, and the tutorial situation, to decide the best pedagogical action. The affective model is implemented as a decision network with a utility measure on learning. For the construction of the affective behavior model, we are using personality questionnaires and emotions models. An initial evaluation of the model is presented, based on questionnaires applied to experienced teachers. We present the initial results of this evaluation. [ABSTRACT FROM AUTHOR]

Question Classification (QC) is usually the first stage in a Question Answering system. This paper presents a multilingual SVM-based question classification system aiming to be language and domain independent. For this purpose, we use only surface text features. The system has been tested on the TREC QA track questions set obtaining encouraging results. [ABSTRACT FROM AUTHOR]

In this paper we present the performance evaluation of a texture segmentation approach for synthetic textured images. Our segmentation approach uses a Bayesian inference procedure using co-ocurrence properties over a set of randomly sampled points in the image. We developed an exhaustive performance test for this approach that compares segmentation results to the "ground truth" images under a varying number of sampled points, in the neighborhood of each pixel used to classify it in the test images. We show our preliminary results that let us to choose the optimal number of points to analyze in the neighborhood of each pixel to assign a texture label. This method can be easily applied to segment outdoor real textured images. [ABSTRACT FROM AUTHOR]

Feature weighting is a more and more important step in clustering because data become more and more complex. An embedded local feature weighting method has been proposed in [1]. In this paper, we present a new method based on the same cost function, but performed through a genetic algorithm. The learning process can be performed through an evolutionary approach or through a cooperavive coevolutionary approach. Moreover, the genetic algorithm can be combined with the original Weighting K-means algorithm in a Lamarckian learning paradigm. We compare hill-climbing optimization versus genetic algorithms, evolutionary versus coevolutionary approaches, and Darwinian versus Lamarckian learning on different datasets. The results seem to show that, on the datasets where the original algorithm is efficient, the proposed methods are even better. [ABSTRACT FROM AUTHOR]

Open learning environments often involve simulation where learners can experiment with different aspects and parameters of a given phenomenon to observe the effects of these changes. These are desirable in virtual laboratories. However, an important limitation of open learning environments is the effectiveness for learning, because it strongly depends on the learner ability to explore adequately. We have developed a semi-open learning environment for a virtual robotics laboratory based on simulation, to learn through free exploration, but with specific performance criteria that guide the learning process. We proposed a generic architecture for this environment, in which the key element is an intelligent tutoring system coupled to a virtual laboratory. The tutor module combines the performance and exploration behaviour of a student in several experiments, to decide the best way to guide his/her. We present an evaluation with an initial group of 20 students. The results show how this semi-open leraning environment can help to accelerate and improve the learning process. [ABSTRACT FROM AUTHOR]

Industrial CNC-milling centers demand adaptive control systems for better product quality. Surface roughness of machined parts is a key indicator of product quality, as it is closely related to functional features of parts such as fatigue life, friction, wear, etc. However, on-line control systems for surface roughness are not yet ready for industrial use. One of the main reasons is the absence of sensors that provide measurements reliably and effectively in a hostile machining environment. One potential solution is to combine readings from several different kinds of sensors in an intelligent sensor-fusion monitoring system. We implemented such a system and compared three modelling approaches for sensor-fusion: multiple regression, artificial neural networks (ANNs), and a new probabilistic approach. Probabilistic approaches are desirable because they can be extended beyond simple prediction to provide confidence estimates and diagnostic information as to probable causes. While our early experimental results with aluminum show that the ANN approach has the greatest predictive power over a variety of operating conditions, our probabilistic approach performs well enough to justify continued research given its many additional benefits. [ABSTRACT FROM AUTHOR]

This paper proposes a fuzzy logic based autonomous flight controller for UAVs (unmanned aerial vehicles). Three fuzzy logic modules are developed for the control of the altitude, the speed, and the roll angle, through which the altitude and the latitude-longitude of the air vehicle are controlled. The implementation framework utilizes MATLAB's standard configuration and the Aerosim Aeronautical Simulation Block Set which provides a complete set of tools for rapid development of detailed 6 degree-of-freedom nonlinear generic manned/unmanned aerial vehicle models. The Aerosonde UAV model is used in the simulations in order to demonstrate the performance and the potential of the controllers. Additionally, Microsoft Flight Simulator and FlightGear Flight Simulator are deployed in order to get visual outputs that aid the designer in the evaluation of the controllers. Despite the simple design procedure, the simulated test flights indicate the capability of the approach in achieving the desired performance. [ABSTRACT FROM AUTHOR]

Input-Output data modelling using fully tuned radial basis function networks(RBF) for a tilt rotor aircraft experimental platform is presented in this paper. The behavior of the four degree-of-freedom platform exemplifies a high order nonlinear system with significant cross-coupling between longitudinal, latitudinal directional motions, and tilt rotor nacelles rolling movement. This paper develops a practical algorithm coupled with model validity tests for identifying nonlinear autoregressive moving average model with exogenous inputs(NARMAX). It is proved that input-output data modelling using fully tuned algorithm is suitable for modelling novelty configuration air vehicles. A procedure for system modelling was proposed in the beginning of this paper and the subsequent sections provided detailed descriptions on how each stage in the procedure could be realized. The effectiveness of this modelling procedure is demonstrated through the tilt rotor aircraft platform. The estimated model can be utilized for nonlinear flight simulation and control studies. [ABSTRACT FROM AUTHOR]

In this work a stable task space neuro algorithm for set-point control of robot manipulators with uncertain parameters is proposed. A depart from current approaches is the fact that a Wavelet Neural Network with on-line real-time learning seeks to explicitly compensate both the unknown gravity terms and the mismatch between the true and the estimated Jacobian matrix and the fact that it does not need velocity measurements. Linear position filtering is used to estimated the robot joint velocity in the control law and the properties of the Wavelet Neural Network are employed for avoiding velocity measurements in the learning rule. It is shown that all the closed loop signals are uniformly ultimately bounded. Experimental results in a two degrees of freedom robot are presented to evaluate the proposed controller. [ABSTRACT FROM AUTHOR]

The features of a novel dynamical discrete-time algorithm for robust adaptive learning in feed-forward neural networks and its application to the neuro-adaptive nonlinear feedback control of systems with uncertain dynamics are presented. The proposed approach makes a direct use of variable structure systems theory. It establishes an inner sliding motion in terms of the neurocontroller parameters, leading the learning error toward zero. The outer sliding motion concerns the controlled nonlinear system, the state tracking error vector of which is simultaneously forced towards the origin of the phase space. It is shown that there exists equivalence between the two sliding motions. The convergence of the proposed algorithm is established and the conditions are given. Results from a simulated neuro-adaptive control of Duffing oscillator are presented. They show that the implemented neurocontroller inherits some of the advantages of the variable structure systems: high speed of learning and robustness. [ABSTRACT FROM AUTHOR]

The main goal of this work is to study the reliability of fuzzy logic based systems. Three different configurations were compared to support this research. The context used was to guide a simulated robot through a virtual world populated with obstacles. In the first configuration, the system controls only the rotation angle of the robot. In the second one, there is an additional output that controls its step. In the third one, improvements were included in the decision process that controls the step and the rotation angle. In order to compare the performance of these approaches, we studied the controller stability based on the removal of rules. We measured two parameters: processing time and the amount of step necessary to reach the goal. This research shows that simplicity and easiness of the design of fuzzy controllers don't compromise its efficiency. Our experiments confirm that fuzzy logic based systems can properly perform under adverse conditions. [ABSTRACT FROM AUTHOR]

This paper presents a new approach to improve performance in wireless ad-hoc networks with the DSR protocol using a Behavior-Based Architecture. Four levels of competence based on strategies to improve the cache were implemented for the Behavior-Based Architecture: sort routes, prefer fresher routes, selection, and disperse traffic. A conflict solver was implemented to solve counter level commands. Three different activation criteria for the conflict solver were developed, generating instances of the architecture. Two metrics were used to evaluate the performance of the ad-hoc network: end-to-end delay and dropped packet average for voice and data services, respectively. Six scenarios were analyzed showing the improvement of the network performance with respect to the original DSR protocol. [ABSTRACT FROM AUTHOR]

This paper presents a dynamic optimization approach based on the differential evolution (DE) strategy which is applied to the concurrent optimal design of a continuously variable transmission (CVT). The structure-control integration approach is used to state the concurrent optimal design as a dynamic optimization problem which is solved using the Constraint Handling Differential Evolution (CHDE) algorithm. The DE strategy is compared with the sequential approach. The results presented here demonstrate that the DE strategy is less expensive than the sequential approach from the computational implementation point of view. [ABSTRACT FROM AUTHOR]

Two communities work in parallel in model-based diagnosis: FDI and DX. In this work an integration of the FDI and the DX communities is proposed. Only relevant information for the identification of the minimal diagnosis is used. In the first step, the system is divided into clusters of components, and each cluster is separated into nodes. The minimal and necessary set of contexts is then obtained for each cluster. These two steps automatically reduce the computational complexity since only the essential contexts are generated. In the last step, a signature matrix and a set of rules are used in order to obtain the minimal diagnosis. The evaluation of the signature matrix is on-line, the rest of the process is totally off-line. [ABSTRACT FROM AUTHOR]

We present a remote operated robot application based on a standard CORBA distributed system to control a MOTOMAN 6-DOF arm manipulator. The robot is based on XRC2001 robot controller. Current approach could be extented to other Yaskawa controllers (e.g. ERC, ERCII, MRC, MRCII) without major changes. The main idea is to define a set of generic IDL interfaces that can be used to integrate commercial proprietary libraries hiding the intricate of low level components. The challenge is to create a remote client-server application which facilitates the integration of one or several arm manipulators based on mentioned controllers independently from computer system or different platforms. Keywords: robotic, distributed systems, CORBA. [ABSTRACT FROM AUTHOR]

Robots can be greatly benefited from the integration of artificial senses in order to adapt to changing worlds. To be effective in complex unstructured environments robots have to perceive the environment and adapt accordingly. In this paper, it is introduced a biology inspired multimodal architecture called M2ARTMAP which is based on the biological model of sensorial perception and has been designed to be a more versatile alternative to data fusion techniques and non-modular neural architectures. Besides the computational overload compared to FuzzyARTMAP, M2ARTMAP reaches similar performance. This paper reports the results found in simulated environments and also the observed results during assembly operations using an industrial robot provided with vision and force sensing capabilities. [ABSTRACT FROM AUTHOR]

Industrial manufacturing cells involving fixtureless environments require more efficient methods to achieve assembly tasks. This paper introduces an approach for Robotic Fixtureless Assembly (RFA). The approach is based on the Fuzzy ARTMAP neural network and learning strategies to acquire the skill from scratch without knowledge about the assembly system. The vision system provides the necessary information to accomplish the assembly task such as pose, orientation and type of component. Different ad-hoc input vectors were used as input to the assembly and the vision systems through several experiments which are described. The paper also describes the task knowledge acquisition and the followed strategies to solve the problem of automating the peg-in-hole assembly using 2D images. The approach is validated through experimental work using an industrial robot. [ABSTRACT FROM AUTHOR]

Intelligent service robots (ISRs) adapt to unpredictable environments by determining how to resolve the problems occurred in a troubled situation. As a way to successfully and continuously provide services, we envisage the software system embedded in a robot to dynamically reconfigure itself using new components selected from component repositories. This paper describes a component selection mechanism, which is an essential function to support such dynamic reconfiguration. We adopt a semantically-based component selection mechanism in which situational information around ISRs is represented as critical semantic information for the service robots to select software components. [ABSTRACT FROM AUTHOR]

Advanced robotic systems require an end effector capable of achieving considerable gripping dexterity in unstructured environments. A dexterous end effector has to be able of dynamic adaptation to novel and unforeseen situation. Thus, it is vital that gripper controller is able to learn from its perception and experience of the environment. An attractive approach to solve this problem is intelligent control, which is a collection of complementary 'soft computing' techniques within a framework of machine learning. Several attempts have been made to combine methodologies to provide a better framework for intelligent control, of which the most successful has probably been that of neurofuzzy modelling. Here, a neurofuzzy controller is trained using the actor-critic method. Further, an expert system is attached to the neurofuzzy system in order to provide the reward signal and failure signal. Results show that the proposed framework permits a transparent-robust control of a robotic end effector. [ABSTRACT FROM AUTHOR]

Robotic end effectors are used over a diverse range of applications where they are required to grip with optimal force to avoid the object be either dropped or crushed. The slipping state can be easily detected by the output of the slip sensor. When the output has a non-zero value, the object is slipping. Conversely, detecting the deformation (crushing) state is more difficult, especially in an unstructured environment. Current proposed methodologies are ad hoc and specialised to the particular object or objects to be handled. Consequently, the gripper can only manipulate prior known objects, constraining the gripper application to a small set of predetermined objects. Accordingly, in this paper, it is proposed a hybrid approach of fuzzy and expert systems that permits to detect when an unknown object is being deformed. To determinate when the gripped object is being deformed, the fuzzy/expert system uses information from three sensors: applied force, slip rate and finger position. Several objects of different characteristics were used to prove the effectiveness of the proposed approach. [ABSTRACT FROM AUTHOR]

A new reference walking trajectory for a planar five-link biped, considering both the SSP and the DSP, is presented firstly. A new combined controller to generate walking gaits following the reference trajectory is designed subsequently. The controller of five-link biped is consisted of PD controller and a fuzzy wavelet neural network controller. The scalable and shiftable coefficients of the wavelet function and weights of the network can be acquired by training the network by back-propagation algorithm online. The simulation results of the reference trajectory show that the proposed reference trajectory have good stability, repeatability and continuity during both SSP and the DSP, and when given the different initial conditions, the compatible trajectories can be achieved correspondingly. The simulation results of the trained controller show that the controller can generate the walking gaits to track the reference trajectory as close as possible. [ABSTRACT FROM AUTHOR]

For autonomous mobile robots following a planned path, self-localization is a very important task. Cumulative errors derived from the different noisy sensors make it absolutely necessary. Absolute robot localization is commonly made measuring relative distance from the robot to previously learnt landmarks on the environment. Landmarks could be interest points, colored objects, or rectangular regions as posters or emergency signs, which are very useful and not intrusive beacons in human environments. This paper presents an active localization method: a visual planning function selects from a free collision path and a set of planar landmarks, a subset of visible landmarks and the best combination of camera parameters (pan, tilt and zoom) for positions sampled along the path. A visibility measurement and some utility measurements were defined in order to select for each position, the camera modality and the subset of landmarks that maximize these local criteria. Finally, a dynamic programming method is proposed in order to minimize saccadic movements all over the trajectory. [ABSTRACT FROM AUTHOR]

We propose a framework for reactive motion and sensing planning based on critical events. A critical event amounts to crossing a critical curve, which divides the environment. We have applied our approach to two different problems: i) object finding and ii) pursuit-evasion. We claim that the proposed framework is in general useful for reactive motion planning based on information provided by sensors. We generalize and formalize the approach and suggest other possible applications. [ABSTRACT FROM AUTHOR]

This paper shows the performance of randomized low-discre-pancy sequences compared with others low-discrepancy sequences. We used two motion planning algorithms to test this performance: the expansive planner proposed in [1], [2] and SBL [3] . Previous research already showed that the use of deterministic sampling outperformed PRM approaches [4], [5], [6]. Experimental results show performance advantages when we use randomized Halton and Sobol sequences over Mersenne-Twister and the linear congruential generators used in random sampling. Keywords: Sampling-based motion planning, deterministic sampling, randomized Halton sequence. [ABSTRACT FROM AUTHOR]

In highly dynamic environments with uncertainty the elaboration of long or rigid plans is useless because the constructed plans are frequently dismissed by the arrival or new unexpected situations; in these cases, a "second-best" plan could rescue the situation. We present a new real-time planning method where we take into consideration the number and quality of future options of the next action to choose, in contrast to most planning methods that just take into account the intrinsic value of the chosen plan or the maximum valued future option. We apply our method to the Robocup simulated soccer competition, which is indeed highly dynamic and involves uncertainty. We propose a specific architecture for implementing this method in the context of a player agent in the Robocup competition, and we present experimental evidence showing the potential of our method. [ABSTRACT FROM AUTHOR]

A prospective paradigm, named Growing Functional Modules (GFM) has been introduced in a recent publication. Based on the epigenetic approach, the GFM paradigm is conceived to automatically generate "artificial brains" that are able to build, through interaction, their own representation of their environments. The present application consists in designing an artificial brain for a simple virtual mushroom shaped robot named hOnGo. This paper describes this initial implementation and its preliminary results. [ABSTRACT FROM AUTHOR]

This work presents an infant cry automatic recognizer development, with the objective of classifying three kinds of infant cries, normal, deaf and asphyxia from recently born babies. We use extraction of acoustic features such as LPC (Linear Predictive Coefficients) and MFCC (Mel Frequency Cepstral Coefficients) for the cry's sound waves, and a genetic feature selection system combined with a feed forward input delay neural network, trained by adaptive learning rate back-propagation. We show a comparison between Principal Component Analysis and the proposed genetic feature selection system, to reduce the feature vectors. In this paper we describe the whole process; in which we include the acoustic features extraction, the hybrid system design, implementation, training and testing. We also show the results from some experiments, in which we improve the infant cry recognition up to 96.79% using our genetic system. We also show different features extractions that result on vectors that go from 145 up to 928 features, from cry segments of 1 and 3 seconds respectively. Keywords: Feature Selection, Evolutionary Strategies, Classification, Infant Cry Analysis, Pattern Recognition, Hybrid System. [ABSTRACT FROM AUTHOR]

The market demand for dental implants is growing at a significant pace. In practice, some dental implants do not succeed. Important questions in this regard concern whether machine learning techniques could be used to predict whether an implant will be successful and which are the best techniques for this problem. This paper presents a comparative study on machine learning techniques for prediction of success of dental implants. The techniques compared here are: (a) constructive RBF neural networks (RBF-DDA), (b) support vector machines (SVM), (c) k nearest neighbors (kNN), and (d) a recently proposed technique, called NNSRM, which is based on kNN and the principle of structural risk minimization. We present a number of simulations using real-world data. The simulations were carried out using 10-fold cross-validation and the results show that the methods achieve comparable performance, yet NNSRM and RBF-DDA produced smaller classifiers. [ABSTRACT FROM AUTHOR]

In this paper we describe a human organ selection process in which agents argue over whether a given donor's organ is viable for transplantation. This process is framed in the CARREL System; an agent-based organization designed to improve the overall transplant process. We formalize an argumentation based framework that enables CARREL agents to construct and assess arguments for and against the viability of a donor's organ for a given potential recipient. We believe that the use of argumentation has the potential to increase the number of human organs that current selection processes make available for transplantation. [ABSTRACT FROM AUTHOR]

An integrated method using the adaptive segmentation of brain tissues in Magnetic Resonance Imaging (MRI) images is proposed in this paper. Firstly, we give a template of brain to remove the extra-cranial tissues. Subsequently, watershed algorithm is applied to brain tissues as an initial segmenting method. Normally, result of classical watershed algorithm on gray-scale textured images such as tissue images is over-segmentation. The following procedure is a merging process for the over-segmentation regions using fuzzy clustering algorithm (Fuzzy C-Means). But there are still some regions which are not partitioned completely, particularly in the transitional regions between gray matter and white matter. So we proposed a rule-based re-segmentation processing approach to partition these regions. This integrated scheme yields a robust and precise segmentation. The efficacy of the proposed algorithm is validated using extensive experiments. [ABSTRACT FROM AUTHOR]

A Point Distribution Model (PDM) of the prostate has been constructed and used to automatically outline the contour of the gland in transurethral ultrasound images. We developed a new, two stages, method: first the PDM is fitted, using a multi-population genetic algorithm, to a binary image produced from Bayesian pixel classification. This contour is then used during the second stage to seed the initial population of a simple genetic algorithm, which adjusts the PDM to the prostate boundary on a grey level image. The method is able to find good approximations of the prostate boundary in a robust manner. The method and its results on 4 prostate images are reported. Keywords: Boundary segmentation, Genetic algorithms, Point distribution models. [ABSTRACT FROM AUTHOR]