By Topic

Foundations of Computational Intelligence, 2007. FOCI 2007. IEEE Symposium on

Date 1-5 April 2007

Filter Results

Displaying Results 1 - 25 of 101
  • Message From the Program Co-Chairs

    Publication Year: 2007 , Page(s): nil1 - nil2
    Save to Project icon | Request Permissions | PDF file iconPDF (60 KB)  
    Freely Available from IEEE
  • IEEE Symposium on Foundations of Computational Intelligence (FOCI'07)

    Publication Year: 2007 , Page(s): nil3 - nil8
    Save to Project icon | Request Permissions | PDF file iconPDF (87 KB)  
    Freely Available from IEEE
  • Rough-Neuro-Fuzzy Systems for Classification

    Publication Year: 2007 , Page(s): 1 - 8
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (293 KB) |  | HTML iconHTML  

    In the paper we present flexible neuro-fuzzy systems and a method for their reduction. The method is based on the concept of the weighted triangular norms. Moreover, a rough-neuro-fuzzy classifier working in the case of missing features is described. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Evolutionary Multiobjective Design of Fuzzy Rule-Based Systems

    Publication Year: 2007 , Page(s): 9 - 16
    Cited by:  Papers (6)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (282 KB) |  | HTML iconHTML  

    The main advantage of fuzzy rule-based systems over other non-linear models such as neural networks is their high interpretability. Fuzzy rules can be usually interpreted in a linguistic manner because they are described by linguistic values such as small and large. Fuzzy rule-based systems have high accuracy as well as high interpretability. A large number of tuning methods have been proposed to improve their accuracy. Most of those tuning methods are based on learning algorithms of neural networks and/or evolutionary optimization techniques. Accuracy improvement of fuzzy rule-based systems, however, is usually achieved at the cost of interpretability. This is because the accuracy improvement often increases the complexity of fuzzy rule-based systems. Thus one important issue in the design of fuzzy rule-based systems is to find a good tradeoff between the accuracy and the complexity. The importance of finding a good accuracy-complexity tradeoff has been pointed out in some studies in the late 1990s. Recently evolutionary multiobjective optimization algorithms were used to search for various fuzzy rule-based systems with different accuracy-complexity tradeoffs. Users are supposed to choose a final model based on their preference from the obtained fuzzy rule-based systems. Some users may prefer a simple one with high interpretability. Other users may prefer a complicated one with high accuracy. In this paper, we explain evolutionary multiobjective approaches to the design of accurate and interpretable fuzzy rule-based systems. We also suggest some future research directions related to the evolutionary multiobjective design of fuzzy rule-based systems. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Evolutionary Algorithms in the Presence of Noise: To Sample or Not to Sample

    Publication Year: 2007 , Page(s): 17 - 24
    Cited by:  Papers (2)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1236 KB) |  | HTML iconHTML  

    In this paper, we empirically analyze the convergence behavior of evolutionary algorithms (evolution strategies - ES and genetic algorithms A) for two noisy optimization problems which belong to the class of functions with noise induced multi-modality (FNIMs). Although, both functions are qualitatively very similar, the ES is only able to converge to the global optimizer state for one of them. Additionally, we observe that canonical GA exhibits similar problems. We present a theoretical analysis which explains the different behaviors for the two functions and which suggests to resort to resampling strategies to solve the problem. Although, resampling is an inefficient way to cope with noisy optimization problems, it turns out that depending on the properties of the problem, (moderate) resampling might be necessary to guarantee convergence to the robust optimizer View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • When the Plus Strategy Outperforms the Comma Strategyand When Not

    Publication Year: 2007 , Page(s): 25 - 32
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (247 KB) |  | HTML iconHTML  

    Occasionally there have been long debates on whether to use elitist selection or not. In the present paper the simple (1, lambda) EA and {1 + lambda) EA operating on {0, l}n are compared by means of a rigorous runtime analysis. It turns out that only values for lambda that are logarithmic in n are interesting. An illustrative function is presented for which newly developed proof methods show that the (1, lambda) EA - where lambda is logarithmic in n - outperforms the (1 + lambda) EA for any lambda. For smaller offspring populations the (1, lambda) EA is inefficient on every function with a unique optimum, whereas for larger lambda the two randomized search heuristics behave almost equivalently. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • On the Influence of Phenotype Plasticity on Genotype Diversity

    Publication Year: 2007 , Page(s): 33 - 40
    Cited by:  Papers (2)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (609 KB) |  | HTML iconHTML  

    A large body of research has investigated the advantages of combining phenotype adaptation and genotype adaptation. The hybridization of genetic search and local search methods, often known as memetic algorithms, and the influence of learning on evolution, i.e., the Baldwin effect and the hiding effect, have been widely studied. However, most work assumes a stationary environment, and thus overlooks potentially advantages or disadvantages that can arise from phenotype plasticity only in changing environments. We show that a process with two levels of adaptation allows the system to operate on two different levels of diversity at the same time, which can be of great advantage under certain environmental conditions View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A Probabilistic Model of MOSAIC

    Publication Year: 2007 , Page(s): 41 - 46
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (695 KB) |  | HTML iconHTML  

    Humans can generate accurate and appropriate motor commands in various and even uncertain environments. MOSAIC (MOdular Sellection And Identification for Control) was formerly proposed for describing such human ability, but it includes some complex and heuristic procedures which make the model's understandability hard. In this article, we present an alternative and probabilistic model of MOSAIC (p-MOSAIC) as a mixture of normal distributions, and an online EM-based learning method for its predictors and controllers. Theoretical consideration shows that the learning rule of p-MOSAIC corresponds to that of MOSAIC except for some points mostly related to the controller learning. Experimental studies using synthetic datasets have shown some practical advantages of p-MOSAIC. One is that the learning rule of p-MOSAIC makes the estimation of 'responsibility' stable. Another is that p-MOSAIC realizes accurate control and robust parameter learning in comparison to the original MOSAIC especially in noisy environments, due to the direct incorporation of the noise into the model View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Learning Bayesian Network Structures with Discrete Particle Swarm Optimization Algorithm

    Publication Year: 2007 , Page(s): 47 - 52
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (941 KB) |  | HTML iconHTML  

    A novel structure learning algorithm of Bayesian networks (BNs) using particle swarm optimization (PSO) is proposed. For searching in structure spaces efficiently, a discrete PSO algorithm is designed in term of the characteristics of BNs. Firstly, fitness function is given to evaluate the structure of BN. Then, encoding and operations for PSO are designed to provide guarantee of convergence. Finally, experimental results show that this PSO based learning algorithm outperforms genetic algorithm based learning algorithm in convergence speed and quality of obtained structures View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A Functional-Link-Based Fuzzy Neural Network for Temperature Control

    Publication Year: 2007 , Page(s): 53 - 58
    Cited by:  Papers (10)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (311 KB) |  | HTML iconHTML  

    This study presents a functional-link-based fuzzy neural network (FLFNN) structure for temperature control. The proposed FLFNN controller uses functional link neural networks (FLNN) that can generate a nonlinear combination of the input variables as the consequent part of the fuzzy rules. An online learning algorithm, which consists of structure learning and parameter learning, is also presented. The structure learning depends on the entropy measure to determine the number of fuzzy rules. The parameter learning, based on the gradient descent method, can adjust the shape of the membership function and the corresponding weights of the FLNN. Simulation result of temperature control has been given to illustrate the performance and effectiveness of the proposed model View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Granular Computing in Actor-Critic Learning

    Publication Year: 2007 , Page(s): 59 - 64
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (172 KB) |  | HTML iconHTML  

    The problem considered in this paper is how to guide actor-critic learning based an information granules that reflect knowledge about acceptable behavior patterns. The solution to this problem stems from approximation spaces, which were introduced by Zdzistaw Pawlak starting in the early 1980s and which provide a basis for perception of objects that are imperfectly known. It was also observed by Ewa Orlowska En 1982 that approximation spaces serve as a formal counterpart of perception, or observation. In our case, approximation spaces provide a ground for deriving pattern-based behaviours as well as information granules that can be used to influence the policy structure of an actor in a beneficial way. This paper includes the results of a recent study of swarm behavior by collections of biologically-inspired bots carried out in the context of an artificial ecosystem. This ecosystem has an ethological basis that makes it possible to observe and explain the behavior of biological organisms that carries over into the study of actor-critic learning by interacting robotic devices. The contribution of this article is a framework for actor-critic learning defined in the context of approximation spaces and information granulation View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Definability of Approximations for a Generalization of the Indiscernibility Relation

    Publication Year: 2007 , Page(s): 65 - 72
    Cited by:  Papers (2)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1318 KB) |  | HTML iconHTML  

    We discuss a generalization of the indiscernibility relation, i.e., a relation R that is not necessarily reflexive, symmetric, or transitive. On the basis of granules, defined by R, we introduce the idea of definability. Twelve different basic definitions of approximations are discussed. Since four of these approximations do not satisfy, in general, the inclusion property, four additional modified approximations are introduced. Furthermore, eight other approximations are constructed by duality. The main objective is to study definability of approximations. We study definability of all approximations for reflexive, symmetric, or transitive relations. In particular, for reflexive relations the set of these twenty four approximations is reduced, in general, to the set of fourteen approximations View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Hybrid Optimisation Method Using PGA and SQP Algorithm

    Publication Year: 2007 , Page(s): 73 - 80
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1456 KB) |  | HTML iconHTML  

    This paper investigates the hybridisation of two very different optimisation methods, namely the parallel genetic algorithm (PGA) and sequential quadratic programming (SQP) algorithm. The different characteristics of genetic-based and traditional quadratic programming-based methods are discussed and to what extent the hybrid method can benefit the solving of optimisation problems with nonlinear complex objective and constraint functions. Experiments show the hybrid method effectively combines the robust and global search property of parallel genetic algorithms with the high convergence velocity of the sequential quadratic programming algorithm, thereby reducing computation time, maintaining robustness and increasing solution quality View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Opposition-Based Differential Evolution (ODE) with Variable Jumping Rate

    Publication Year: 2007 , Page(s): 81 - 88
    Cited by:  Papers (7)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1018 KB) |  | HTML iconHTML  

    In this paper, a time varying jumping rate (TVJR) model for opposition-based differential evolution (ODE) has been proposed. According to this model, the jumping rate changes linearly during the evolution based on the number of function evaluations. A test suite with 15 well-known benchmark functions has been employed to compare performance of the DE and ODE with variable jumping rate settings. Results show that a higher jumping rate is more desirable during the exploration than during the exploitation. Details for the proposed approach and the conducted experiments are provided View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Large-scale Sensor Networks as Collective and Frustrated Systems

    Publication Year: 2007 , Page(s): 89 - 94
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (160 KB) |  | HTML iconHTML  

    This article presents a large-scale analysis of a distributed sensing model for systemized and networked sensors. In the system model, a data center acquires binary information from a bunch of L sensors which each independently encode their noisy observations of an original bit sequence, and transmit their encoded sequences to the data center at a combined data rate R, which is strictly limited. Supposing that the sensors use independent quantization techniques, we show that the performance can be evaluated for any given finite R when the number of sensors L goes to infinity. The analysis shows how the optimal strategy for the distributed sensing problem changes at critical values of the data rate R or the noise level p View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Balancing the sticks with fluctuation and delay: Human vs Machine

    Publication Year: 2007 , Page(s): 95 - 99
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (3386 KB) |  | HTML iconHTML  

    We compared the similarities and differences in stick balancing for the human fingertip and that by a servo-controlled machine. The motion of the stick in both cases exhibited a swinging or hunting behavior, which appears to be related to feedback delay. However, human stick balancing appears also to be affected by psychological factors, such as attention, which are not present in machine control systems. We discuss how machine control systems compare with human stick balancing View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Nonequilibrium phase transitions in stochastic systems with and without time delay: controlling various attractors with noise

    Publication Year: 2007 , Page(s): 100 - 105
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (178 KB) |  | HTML iconHTML  

    Statistical behavior of ensembles of nonlinearly coupled elements driven by external noise is studied on the basis of nonlinear Fokker-Planck equations. The models incorporate two kinds of noise, the Langevin noise and the colored noise introduced in the coupling strength, and time delays. Various types of nonequilibrium phase transitions including Hopf bifurcations and transitions between limit cycle and chaos are shown to occur as the noise level is changed. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Behavioral Partitioning in a Hierarchical Mixture of Experts using K-Best-Experts Algorithm

    Publication Year: 2007 , Page(s): 106 - 112
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (989 KB) |  | HTML iconHTML  

    In recent years methods for combining multiple experts (multi expert systems, MES) have been used to solve different problems of classification and regression. In particular hierarchical mixture of experts (HME) has been widely studied. This paper presents a novel method which divides the problem space into behaviorally portioned subsets using k-best-experts algorithm and then uses the HME structure to assign an expert to each subset. The gates used in the HME structure are support vector machines which are trained to route each problem to the best fitting expert. The method is implemented and tested on the DELVE framework and is compared with other similar methods View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Why Intervals? Why Fuzzy Numbers? Towards a New Justification

    Publication Year: 2007 , Page(s): 113 - 119
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (8492 KB)  

    The purpose of this paper is to present a new characterization of the not of all intervals (and of the corresponding set of fuzzy numbers). This characterization is based on several natural properties useful in mathematical modeling; the main of these properties is the necessity to be able to combine (fuse) several pieces of knowledge View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Fuzzy Partial-Order Relations for Intervals and Interval Weighted Graphs

    Publication Year: 2007 , Page(s): 120 - 127
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1667 KB) |  | HTML iconHTML  

    Weighted graphs have been broadly employed in various kinds of applications. Weights associated with edges in a graph are constants mostly in the literature. However, in real world applications, these weights may vary within ranges rather than fixed values. To model such kind of uncertainty or variability, we propose interval-valued weighted graphs in this study. In solving practical graph applications such as finding shortest paths and minimum spanning trees for interval weighted graphs, it is necessary to be able to compare interval valued weights. However, two general intervals can not be ordered reasonably in binary logic. In this paper, we establish fuzzy partial-order relations for intervals. These relations are continuous, except only at a single point in a special case. After studying the properties of the fuzzy partial order relations, we applied the interval partial order to extend well-known shortest path and minimum spanning tree algorithms for interval weighted graphs. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Task Scheduling on Flow Networks with Temporal Uncertainty

    Publication Year: 2007 , Page(s): 128 - 135
    Cited by:  Papers (2)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1651 KB) |  | HTML iconHTML  

    This study was motivated by the large-scale evacuations and aftermath relief efforts caused by hurricanes Katrina and Rita in 2005. During large-scale natural disasters, both demands and uncertainties on transportation networks can be pushed to the extreme. The objective of this study is to optimally schedule tasks on a flow network that has temporal uncertainty modeled with interval valued time costs. We apply a fuzzy partial order relation for intervals to extend the Edmonds-Karp max-flow min-cost algorithm in this paper. Then, using the greedy approach, we propose task optimal schedule algorithms on a flow network with temporal uncertainties by utilizing its full capacity with the least possibility of delay in terms of the fuzzy partial order relations of intervals. In addition to the task scheduling algorithms, simple case studies are also provided in this paper View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • The Use of Interval Methods in Signal Processing and Control for Systems Biology

    Publication Year: 2007 , Page(s): 136 - 142
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (139 KB) |  | HTML iconHTML  

    The development of approaches for understanding the complex dynamics of biological systems is a growing research area in electrical engineering, particularly in the fields of signal processing and controls. The focus of our research is the exploitation of the parallels between engineering and biology through the development of optimization and identification methods. Specifically, this research consists of developing methods for the estimation of unmeasured states, the identification of parameters of kinetic models and the validation of biochemical models. This work falls under the general research topic of systems biology. We explore the use of interval analysis in developing numerical algorithms for optimization and validation of systems biology problems. A major attribute of this method is that convergence to global minima is guaranteed. This paper includes a development of an adaptive interval optimization method based on the branch-and-bound method known as smooth interval branch-and-bound. One potential impact of this research is the development of more accurate models of biological systems. This will aid in the design of drugs for cancer and disease treatment and aid in the study of how they propagate View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Performance Optimization of Adaptive Resonance Neural Networks Using Genetic Algorithms

    Publication Year: 2007 , Page(s): 143 - 148
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1824 KB) |  | HTML iconHTML  

    We present a hybrid clustering system that is based on the adaptive resonance theory 1 (ART1) artificial neural network (ANN) with a genetic algorithm (GA) optimizer, to improve the ART1 ANN settings. As a case study, we will consider text clustering. The core of our experiments will be the quality of clustering, multi-dimensional domain space of ART1 design parameters has many possible combinations of values that yield high clustering quality. These design parameters are hard to estimate manually. We proposed GA to find some of these sets. Results show better clustering and simpler quality estimator when compared with the existing techniques. We call this algorithm genetically engineered parameters ART1 or ARTgep View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • On the BMDGAs and Neural Nets

    Publication Year: 2007 , Page(s): 149 - 155
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1210 KB) |  | HTML iconHTML  

    We analyze a bivariate marginal distribution genetic model in case of infinite populations and provide relations between the associated infinite population genetic system and the neural networks. A lower bound on population size is exhibited stating that the behaviour of the finite population system, in case of sufficiently large sizes, can be suitably approximated by the behaviour of the corresponding infinite population system for a number of transitions exponentially greater than that suggested by Vose's analysis. The infinite population system is analyzed by showing that, conversely to what happens in the univariate case, the fitness is not a Lyapunov function for its asynchronous variant. The attractors (with binary components) of the infinite population genetic system are characterized as equilibrium points of a discrete (neural network) system that can be considered as a variant of a Hopfield's network; it is shown that the fitness is a Lyapunov function for the variant of the discrete Hopfield's net. The genetic algorithm based on the proposed infinite population system is experimentally compared with the (neural) network algorithm for the max-cut problem. Our main result can be summarized by stating that the relation between marginal distribution genetic systems and neural nets is much more general than that already shown elsewhere for the univariate models. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Artificial immune systems based novelty detection with CNN-UM

    Publication Year: 2007 , Page(s): 156 - 161
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (369 KB) |  | HTML iconHTML  

    In this paper, we show that the earlier presented immune response inspired algorithmic framework in the work of Gy. Cserey et al. (2006, 2004) for spatial-temporal target detection applications using CNN technology by T. Roska and L.O. Chua (1993, 2002) and T. Roska (2002) can be implemented on the latest CNN-UM chip (Acel6k) by A. Rodriguez-Vazquez (2004) and Bi-i system by A. Zarandy and C. Rekcezky (2005). The implementation of the algorithm is real-time and able to detect novelty events in image flows reliably, running 10000 templates/s with video-frame (25 frame/s) speed and on image size of 128 times 128. Besides that some results of the implementation of this AIS model and its application for natural image flows are shown, the realized adaptation and mutation methods are also introduced. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.