Abstract: A physical neural network based on nanotechnology, including methods thereof. Such a physical neural network generally includes one or more neuron-like nodes, which are formed from a plurality of interconnected nanoconnections formed from nanoconductors. Each neuron-like node sums one or more input signals and generates one or more output signals based on a threshold associated with the input signal. The physical neural network also includes a connection network formed from the interconnected nanoconnections, such that the interconnected nanoconnections used thereof by one or more of the neuron-like nodes are strengthened or weakened according to an application of an electric field.

Abstract: The invention relates to a method of searching a predetermined number of objects which are closest to an example, in a hierarchical object structure which is obtained by using a hierarchical object classification method. The proposed method consists of going through the hierarchical structure starting from the root, descending as far as the first node closest to the searched example of which a number of leaves lower than the number of lacking results depend, and then going back up to the brother node of this node which is closest to the searched example, so as to add other results and so on and so forth until said predetermined number is arrived at. Application: MPEG-7 descriptions.

Abstract: A system for developing and implementing empirically derived algorithms to generate decision rules to predict subject noncompliance and fraud with research protocols in clinical trials allows for the identification of complex patterns of variables that detect or predict subject noncompliance and fraud with research protocol in the clinical trial. The present invention can also be used to monitor subject compliance with the research protocol to determine preferred actions to be performed. Optionally, the invention may provide a spectrum of noncompliance, from minor noncompliance needing only corrective feedback, to significant noncompliance requiring subject removal from the clinical trial.

Abstract: A neural network system includes a feedforward network comprising at least one neuron circuit for producing an activation function and a first derivative of the activation function and a weight updating circuit for producing updated weights to the feedforward network. The system also includes an error back-propagation network for receiving the first derivative of the activation function and to provide weight change data information to the weight updating circuit.

Abstract: A method and system for generating a management solution for a client having an organization. The method comprises the steps of providing a pre-defined solution of documentation, executable code, and installation aids, having pre-determined processes, roles and skills, and tool function; assessing the impact of implementing said solution within said organization of said client; building a training plan of needed training in response to said impact; and defining a client data processing infrastructure needed to implement said pre-defined solution at said client.

Abstract: Method and system for computer assisted performance support include remote electronic monitoring and mentoring to provide for supervision, control and instruction of the worker while the worker performs directed tasks. Activity data representative of worker activities are collected at the work site and can be accessed by a supervisor located remote from the work site. The remote supervisor generates and transmits to the worker feedback based on the activity data. The remote supervisor can evaluate an activity trail of a worker's performance of directed tasks and development of skills through educational training to permit timely identification and correction of worker deficiencies. The ability of the remote supervisor to transmit feedback to the worker while the worker is performing directed tasks enhances rapid development of the worker's skills and workplace safety and productivity.

Abstract: A system and method in accordance with the present invention determines in real-time the portions of a set of characters from a data or character stream which satisfies one or more predetermined regular expressions. A Real-time Deterministic Finite state Automaton (RDFA) ensures that the set of characters is processed at high speeds with relatively small memory requirements. An optimized state machine models the regular expression(s) and state related alphabet lookup and next state tables are generated. Characters from the data stream are processed in parallel using the alphabet lookup and next state tables, to determine whether to transition to a next state or a terminal state, until the regular expression is satisfied or processing is terminated. Additional means may be implemented to determine a next action from satisfaction of the regular expression.

Abstract: A method and apparatus that allow business users to dynamically add, modify, and delete business rules and associated constraints, and then to apply these changes in a very efficient manner without needing to recompile and/or restart e-business applications, is disclosed. According to one embodiment, hybrid use of the business rules and constraints with multiple decision-support engines provides the practical solution for the described problems. Rules are used to define the generic search space only, without needing to describe all special cases. Then an optimization (not rule-based) engine can automatically create and analyze all possible branches of the tree specified by the rules to select the best one. Because one universal rule engine is not sufficient to handle online decision support, in one embodiment there is a family of intelligent engines (including the rule engine) that can resolve the optimization problem generated automatically based on the applicable dynamic rules.

Abstract: The method and system described herein presents a technique for performance analysis that helps users understand the communication behavior of their message passing applications. The method and system described herein may automatically classifies individual communication operations and reveal the cause of communication inefficiencies in the application. This classification allows the developer to quickly focus on the culprits of truly inefficient behavior, rather than manually foraging through massive amounts of performance data. Specifically, the method and system described herein trace the message operations of Message Passing Interface (MPI) applications and then classify each individual communication event using a supervised learning technique: decision tree classification. The decision tree may be trained using microbenchmarks that demonstrate both efficient and inefficient communication.

Abstract: A computer-implemented method for flexibly and efficiently representing and applying business rules in a transaction-processing relational database management system (RDBMS) environment. The method includes providing a deterministic rule-based expert-system shell. A late-binding mechanism within the RDBMS environment is also provided. An extensible data-maintenance mechanism is created for the rule-based expert-system shell. The extensible data-maintenance mechanism maintains sets of approval rules governing business transactions generated by other transaction-processing applications. The data-maintenance mechanism uses late binding to make the sets of rules and rule components stored in the data-maintenance mechanism arbitrarily extensible. A rule-processing engine applies the sets of approval rules stored in the extensible data-maintenance mechanism to business transactions originating in transaction-processing applications.

Abstract: A codifying and storing method for membership functions representing a membership degree of fuzzy variables defined within a universe of discourse which is discretized into a finite number of points is provided. The membership functions are quantized into a finite number of levels corresponding to a finite number of membership degrees and are stored by means of a characteristic value of each sub-set of values of fuzzy variables having for their image the same value of the membership degree corresponding to one of said levels. Also provided is a method for calculating the value of the membership degree of a fuzzy variable defined within a universe of discourse discretized into a finite number of points with reference to a membership function thereof, as well as to a circuit for calculating the membership degree of a fuzzy variable with reference to a membership function thereof.

Abstract: In a first aspect, a pipelined hardware implementation of a neural network circuit includes an input stage, two or more processing stages and an output stage. Each processing stage includes one or more processing units. Each processing unit includes storage for weighted values, a plurality of multipliers for multiplying input values by weighted values, an adder for adding products outputted from product multipliers, a function circuit for applying a non-linear function to the sum outputted by the adder, and a register for storing the output of the function circuit.

Abstract: A neural processing module is disclosed which combines a weighted synapse array that performs “primitive arithmetic” (products and sums) in parallel with a weight change architecture and a data input architecture that collectively maximize the use of the weighted synapse array by providing it with signal permutations as frequently as possible. The neural processing module is used independently, or in combination with other modules in a planar or stacked arrangement.

Abstract: Power industry boiler tube failures are a major cause of utility forced outages in the United States, with approximately 41,000 tube failures occurring every year at a cost of $5 billion a year. Accordingly, early tube leak detection and isolation is highly desirable. Early detection allows scheduling of a repair rather than suffering a forced outage, and significantly increases the chance of preventing damage to adjacent tubes. The instant detection scheme starts with identification of boiler tube leak process variables which are divided into universal sensitive variables, local leak sensitive variables, group leak sensitive variables, and subgroup leak sensitive variables, and which may be automatically be obtained using a data driven approach and a leak sensitivity function. One embodiment uses artificial neural networks (ANN) to learn the map between appropriate leak sensitive variables and the leak behavior.

Abstract: A system and method for automatically evaluating a decisional rule containing a task and a condition which must be fulfilled before the task can be performed and for automatically performing the task when or whenever the condition is fulfilled in which a decisional rule is entered into a computer, parsed, and then processed such that automatic and continuing iterative evaluations of whether the condition is fulfilled are provided until the condition is fulfilled and then the task associated with the decisional rule is automatically performed and whereby further processing is resumed only after the condition is fulfilled.

Abstract: There is disclosed a pattern identifying neural network comprising at least an input and an output layer, the output layer having a plurality of principal nodes, each principal node trained to recognize a different class of patterns, and at least one fuzzy node trained to recognize all classes of patterns recognized by the principal nodes but with outputs set out at levels lower than the corresponding outputs of the principal nodes.

Abstract: The performance of a given computer task is optimized by utilizing a plurality of intelligent agents suited to perform the computer task but having varied degrees of domain knowledge. Based upon an objective criteria that may be determined for a given situation, one of the plurality of intelligent agents may be selected and dispatched to perform the task, thereby optimizing the performance of the computer task for a wide variety of situations.

Abstract: A genetic algorithm executing device, in which search of a genetic algorithm is improved, sets a distribution of characters at each position of character strings represent candidates of solution, obtain the degree of disorder at each position by the distributions of characters, set a larger mutation rate at the position of which degree of disorder is higher. This device performs a mutation operation on the character strings according to the mutation rates. This device generates the new character strings by selecting positions of which degrees of disorder are higher than a predetermined value, and changing the character at the selected positions.

Abstract: A classification neural network for piecewise linearly separating an input space to classify input patterns is described. The multilayered neural network comprises an input node, a plurality of difference nodes in a first layer, a minimum node, a plurality of perceptron nodes in a second layer and an output node. In operation, the input node broadcasts the input pattern to all of the difference nodes. The difference nodes, along with the minimum node, identify in which vornoi cell of the piecewise linear separation the input pattern lies. The difference node defining the vornoi cell localizes input pattern to a local coordinate space and sends it to a corresponding perceptron, which produces a class designator for the input pattern.

Abstract: A system for processing data includes a memory coupled to a processor. The memory stores a correlithm object that defines a point in a particular space, wherein the particular space is defined by a plurality of dimensions and including a plurality of points. The correlithm object is associated with each point using a metric. The memory further stores data associated with the correlithm object. The processor applies the metric to the correlithm object and each of the plurality of points in the particular space to generate a plurality of values defining a distribution having a mean and a standard of deviation such that the ratio of the mean to the standard of deviation increases with the number of dimensions of the particular space.