By Topic

Adaptive Processes, 1968. Seventh Symposium on

Date 16-18 Dec. 1968

Filter Results

Displaying Results 1 - 25 of 36
  • [Front cover and table of contents]

    Publication Year: 1968 , Page(s): c1
    Save to Project icon | Request Permissions | PDF file iconPDF (462 KB)  
    Freely Available from IEEE
  • Conditional optimization of nonlinear adaptive control systems

    Publication Year: 1968 , Page(s): 11
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (613 KB)  

    A design algorithm for adaptive control of systems described by nonlinear integral operator equations is considered. Kulikowski's original approach is extended to include bounded input functions. The limited range of performance criteria considered previously is similarly extended making use of the concepts of nonlinear functional analysis. The plant operator is assumed to be only partially known, and the optimization problem is formulated as a conditional-minimization problem. The solution is obtained in the form of an iteration and two such iteration techniques are described. The first one is based on the principle of contraction mapping; the second on Newton's modified method. The identification procedure is briefly mentioned and an analog implementation scheme is presented. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • The application of parametric sensitivity coefficient vectors to the design of parameter tracking and self-adaptive multivariable control systems

    Publication Year: 1968 , Page(s): 12
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (208 KB)  

    A method of designing multivariable adaptive control systems using sensitivity methods is discussed. By means of a Lyapunov type of synthesis procedure, stability of adaptation is almost assured. Computational aspects of the sensitivity vectors that are employed in the adaptive control are also briefly examined. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • On the characteristics of the parameter-perturbation process dynamics

    Publication Year: 1968 , Page(s): 13
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (273 KB)  

    First Page of the Article
    View full abstract»

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

    Publication Year: 1968 , Page(s): 14
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (193 KB)  

    A system is described which employs an adaptive modeler to estimate the state and future trajectory of the unknown plant. Control is automatically synthesized to minimize a performance index involving this estimated future. The effects of perfect and imperfect modeling and of optimal and suboptimal control laws are considered. Experimental results are presented. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Theory for a type of digital predictive compensation

    Publication Year: 1968 , Page(s): 15
    Cited by:  Papers (2)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (228 KB)  

    First Page of the Article
    View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A learning controller model for on-line optimization of a class of stochastic control systems

    Publication Year: 1968 , Page(s): 16
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (284 KB)  

    In this paper, a learning controller model is applied to the on-line optimization of a general class of discrete stochastic control systems. Stochastic properties of the plant are unknown to the designer. Using the plant's output, the controller iteratively applys a reinforcement learning algorithm to learn the optimum control policy while minimizing an expected performance index. A technique that accelerates learning and improves performance is discussed. Computer simulation results of the successful application of the learning controller model to several different cases are presented. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • On the use of a priori information in learning control systems

    Publication Year: 1968 , Page(s): 17
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (305 KB)  

    First Page of the Article
    View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Adaptive control of a reflective satellite communication system

    Publication Year: 1968 , Page(s): 21
    Cited by:  Papers (2)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (441 KB)  

    First Page of the Article
    View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • The application of pattern recognition techniques to a remote sensing problem

    Publication Year: 1968 , Page(s): 22
    Cited by:  Papers (5)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (2626 KB)  

    Pattern recognition techniques are being applied to the analysis of data gathered from a multiband optical-mechanical scanner mounted on an aerospace platform. The purpose of the system is to provide automated techniques for making surveys of earth resources such as agricultural crop status, forest inventories, bodies of water, etc. This paper describes techniques used in the research including aspects of categorizer design, feature selection algorithms, and other methods suitable for carrying out research in a high data volume environment. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • The information content measure as a performance criterion for feature selection

    Publication Year: 1968 , Page(s): 23
    Cited by:  Papers (7)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (570 KB)  

    The use of performance measures for selecting features in pattern recognition systems is reviewed. An approximation to the information content measure is derived. The approximation reduces the computation required to calculate the measure. The accuracy of the approximation depends directly on the nature of the patterns and their features. Computational requirements for the approximate measure are specified. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • An application of pattern recognition for social systems simulation

    Publication Year: 1968 , Page(s): 24
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (667 KB)  

    The use of the techniques developed for adaptive learning and pattern recognition have been applied to the social system discussed in, Intergroup Conflict and Cooperation - The Robbers' Cave Experiment by Sherif et al. This was an experiment in social psychology designed to trace over a period of time the formation and functioning of attitudes of one group toward another in an experimental situation. Two methods of simulation were attempted, the first a sequential learning system employing a matched filter machine; the second a parallel self-adaptive machine bearing the acronym GELISIMA GEeralized LInear-element SImulation MAchine. The simulation models of the second kind provided accurate replication of the social system suitable for hypothesis generation, variable manipulation and control, internal structure variations and other parameter manipulation. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A survey of pattern recognition

    Publication Year: 1968 , Page(s): 25
    Cited by:  Papers (3)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1321 KB)  

    This paper surveys the developments in the field of pattern classification to describe the current state-of-the-art. The author divides the pattern classification problem into deterministic and statistical approaches although, in many instances, both converge to the same result. The approaches are further subdivided into: Adalines and Madalines, linear discriminant functions, mathematical programming, mode-seeking, nearest neighbor, Bayes and minimax, statistical criteria, and fuzzy sets subclasses. A major effort is devoted to show relationships among procedures. The conditions under which procedures are equivalent are discussed. Such relationships are summarized through a graph. The problems of supervision, adaptive property, sample size, and sequential analysis are discussed. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • On mode estimation in pattern recognition

    Publication Year: 1968 , Page(s): 31
    Cited by:  Papers (3)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (719 KB)  

    A procedure for determining the modes of a continuous univariate probability density function (p.d.f.) is suggested. Inherent in this procedure is the use of a new nonparametric estimate of the p.d.f. An extension of this method to the multidimensional case is posed and some results of the procedure applied to real data problems are presented. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Unsupervised learning, minimum risk pattern classification for dependent hypotheses and dependent measurements

    Publication Year: 1968 , Page(s): 32
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (465 KB)  

    A recursive Bayes optimal solution is found for the problem of sequential, multicategory pattern recognition, when unsupervised learning is required. An unknown parameter model is developed which, for the pattern classification problem, allows for (i) both constant and time-varying unknown parameters, (ii)partially unknown probability laws of the hypotheses and time-varying parameter sequences, (iii) dependence of the observations on past as well as present hypotheses and parameters, and most significantly, (iv) sequential dependencies in the observations arising from either (or both) dependency in the pattern or information source (context dependence) or in the observation medium (sequential measurement correlation), these dependencies being up to any finite Markov orders. For finite parameter spaces, the solution which is Bayes optimal (minimum risk) at each step is found and shown to be realizable in recursive form with fixed memory requirements. The asymptotic properties of the optimal solution are studied and conditions established for the solution (in addition to making best use of available data at each step) to converge in performance to operation with knowledge of the (unobservable constant unknown parameters. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • An empirical Bayes decision problem

    Publication Year: 1968 , Page(s): 33
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (65 KB)  

    The empirical Bayes decision problem is considered. Let {ai}, i=1,..,N be a sequence of Markov dependent random variables, ai?? {1,...,m} where ai denotes the category of the ith sample also called the state of nature. Let pl j be the elements of the transition matrix of the Markov process and consider that the initial probabilities are equal to the steady state probabilities of the Markov chain. Let xN = {x1,...,xN} be a sequence of random observations where xi has probability density function fa i(xi). Suppose that the receiver does not know which state of nature is acting after the reception of the sample xi and after N observations, it is desired to partition the received samples into m sets with minimum probability of misclassification with respect to the true partition induced by the states of nature. Such a problem may arise in recognition of written characters Ref. [4] and in receiving signals over a noisy channel with intersymbol interference Ref. [3]. In the present work it is assumed that the fj(xi) are unknown, it is only known that fj(xi) belong to a family F of pdf's with known functional form. It is assumed that the probability transition matrix of the Markov chain is unknown. It is shown that if the family F of pdf's satisfies certain identifiability and differentiality conditions, then by using moment estimates of the unknown quantities, a decision function t?? can be determined such that the corresponding risk converges to the optimal Bayes risk. The present work extends the results obtained in Ref. [4] by considering that the transition probabilities and the pdf's are unknown. The work of Ref. [3] is extended by showing file convergence of the risk corresponding to t?? to the optimal risk, without requiring that the signal to noise ratio converge to zero. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A non-parametric method for feature selection

    Publication Year: 1968 , Page(s): 34
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (210 KB)  

    A non-parametric feature selection technique is proposed. It is hoped that a finite number of classes is represented by some finite number of unknown probability structures which are distributed in a finite discrete measurement space. No assumptions of statistical independence between pattern measurements will be made. The proposed non-parametric feature selection criterion is based on the direct estimation of the minimal expected error rates for a given data set of training samples and is independent from the classification technique used. The properties of the proposed feature section are demonstrated using data from agricultural remote sensing. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Randomly generated nonlinear transformations for pattern recognition

    Publication Year: 1968 , Page(s): 35
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (227 KB)  

    A general method is proposed to find non-linear transformations for discrete data in information processing problems. The main feature of the method is random perturbation of the data subject to constraints which ensure that in the transformed space the problem is in some sense simpler. The technique has been applied to pattern recognition by finding a nonlinear transformation of the feature space such that all classes become linearly separable. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Recursive algorithms for pattern classification using misclassified samples

    Publication Year: 1968 , Page(s): 36
    Cited by:  Papers (2)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (291 KB)  

    We consider the samples x(i) belonging to one of the two non-overlapping classes, ??1 and ??0, which possess a separating function f(x). The observed membership of pattern x(i) is represented by the variable z(i) which can assume only one of two values, ?? 1, or z(i) = [sgn f(x(i))]??(i) where ??(i) is the measurement noise and E(??) is known. Thus the membership of the training samples may be erroneous. Using only the available sample pairs {x(i),z(i)}, i=1,2,..., we will obtain either a separating function or an optimal approximation to the separating function f(x). View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A sequential algorithm for piecewise linear classification functions

    Publication Year: 1968 , Page(s): 37
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (257 KB)  

    A sequential algorithm for designing piecewise linear classification functions without a priori knowledge of pattern class distributions is described. The algorithm combines, under control of a performance criterion, adaptive error correcting linear classifier design procedures and clustering techniques. An error rate criterion is used to constrain the classification function structure so as to minimize design calculations and to increase recognition throughput for many classification problems. Examples from the literature are used to evaluate this approach relative to other classification algorithms. View full abstract»

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

    Publication Year: 1968 , Page(s): 38
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (184 KB)  

    Two new rapid adaption procedures are presented in this paper. The first one, Multi-Pattern Maximum Adaption Procedure, corrects all errors simultaneously, instead of one by one (Ref. 2-5), exactly to the unity of the worst dot product, X.W. The second one, Multi-Pattern Random Adaption Procedure, simultaneously adapts every pattern with undesired output to the unity of any one of the wrong dot products. Both procedures converge to a solution weight much faster than the increment, relaxation and modified relaxation adaption procedures. The results of some illustrative examples solved by computer simulation are tabulated for comparison. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Some reflections on pattern recognition using interactive graphics

    Publication Year: 1968 , Page(s): 41
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (157 KB)  

    Workers in pattern recognition will have to cope, initially, with great complexity to solve most real world pattern-recognition problems. In this paper we seek to describe how we feel the interactive graphic computer can be used to aid the worker in pattern recognition. Our goal in examining interactive graphics is, in addition to doing the old things better, to use the new medium to do things differently. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • SARF a signature analysis research facility based on interactive computer graphics

    Publication Year: 1968 , Page(s): 42
    Cited by:  Papers (2)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (93 KB)  

    The Signature Analysis Research Facility (SARF), based on the utilization of interactive computer graphics, provides pattern recognition and general signature analysis capabilities. SARF is operational at AC Electronics - Defense Research Laboratories (AC-DRL) in Goleta, California. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Multi-sensor imagery processing

    Publication Year: 1968 , Page(s): 43
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (77 KB)  

    In considering applications of satellite sensors to exploitation of earth resources, it becomes apparent that imagery from a variety of sensors will be required to accommodate the various disciplines and the environmental constraints. To attempt crop surveys, for example, it appears that both the visual and the infrared spectrum will be utilized. This may be augmented by radar information to assist in defining water boundaries, regions, land with high water content, or to permit penetration over areas of continuous or seasonal cloud cover. One immediately becomes concerned by the lack of commonality in the data acquired by these sensors. The infrared imagery may be obtained by utilizing an optical line scanner with 20-30 channels covering the 0.3?? to 30?? band. The photographic imagery might be obtained by a panchromatic camera with 20-50 lines/mm resolution. The radar data could be provided by a side looking radar with several hundred foot resolution. The registration and resolution compatibility problems appear insurmountable, if one wishes to simultaneously utilize all the imagery to achieve the data classification. However, examining this in the light of information extraction as opposed to data processing one sees an immediate reconciliation. A specific sensor can be used as a primary sensor in establishing an initial categorization. Upon command the other sensory data can then be interrogated to provide supporting or verifying information. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Some comments on interactive aids in statistical signal analysis and pattern recognition

    Publication Year: 1968 , Page(s): 44
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (110 KB)  

    First Page of the Article
    View full abstract»

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