Abstract: Estimation sections which have beforehand learned a relationship between known connection data pertaining to connection design and unknown connection data pertaining to connection design for the known connection data estimate the unknown connection data for the known connection data in accordance wit an input of the known connection data, on the basis of the result of learning. The respective estimation sections are formed from a multilayer feedforward neural network in which layers constituted of a plurality of neurons are coupled together in a direction in which the layer runs from an input layer to an output layer by way of an intermediate layer.

Abstract: A multi-criteria event detection system, comprising a plurality of sensors, wherein each sensor is capable of detecting a signature characteristic of a presence of an event and providing an output indicating the same. A processor for receiving each output of the plurality of sensors is also employed. The processor includes a probabilistic neural network for processing the sensor outputs. The probabilistic neural network comprises a nonlinear, nor-parametric pattern recognition algorithm that operates by defining a probability density function for a plurality of data sets that are each based on a training set data and an optimized kernel width parameter. The plurality of data sets includes a baseline, non-event, first data set; a second, event data set; and a third, nuisance data set.

Abstract: A system and method for generating a custom learning object. The system and method generate the custom learning object based on a knowledge object and a set of user characteristics for a user. In one embodiment, the knowledge object is converted to a set of knowledge atoms. Each knowledge atom is then mapped to a container defining an output format. One or more containers are combined to define the custom learning object.

Abstract: A computer-implemented method and system for building a neural network is disclosed. The neural network predicts at least one target based upon predictor variables defined in a state space. First, an input data set is retrieved that includes the predictor variables and at least one target associated with the predictor variables for each observation. In the state space, a number of points is inserted in the state space based upon the values of the predictor variables. The number of points is less than the number of observations. A statistical measure is determined that describes a relationship between the observations and the inserted points. Weights and activation functions of the neural network are determined using the statistical measure.

Abstract: An intelligent control system based on an explicit model of cognitive development (Table 1) performs high-level functions. It comprises up to O hierarchically stacked neural networks, Nm, . . . , Nm+(O?1), where m denotes the stage/order tasks performed in the first neural network, Nm, and O denotes the highest stage/order tasks performed in the highest-level neural network. The type of processing actions performed in a network, Nm, corresponds to the complexity for stage/order m. Thus N1 performs tasks at the level corresponding to stage/order 1. N5 processes information at the level corresponding to stage/order 5. Stacked neural networks begin and end at any stage/order, but information must be processed by each stage in ascending order sequence. Stages/orders cannot be skipped. Each neural network in a stack may use different architectures, interconnections, algorithms, and training methods, depending on the stage/order of the neural network and the type of intelligent control system implemented.

Abstract: The systems and methods for modeling of a propagation and a containment process described herein enable the modeling of containment processes by mathematically mapping a containment process onto a topological model. The systems and methods described herein can model the effects of various degrees of data fusion and automation that are present in specific locations throughout the containment process. In addition, human and automated responses to propagated defects within the containment process can also be modeled. The method for modeling a containment process includes building a topological model of a containment process having a plurality of containment levels, with each containment level having a containment node. A behavior is then assigned to each containment node. The behavior is configured to model defect propagation based upon a first predetermined level of data fusion in the containment node.

Abstract: Non-linear model with disturbance rejection. A method for training a non linear model for predicting an output parameter of a system is disclosed that operates in an environment having associated therewith slow varying and unmeasurable disturbances. An input layer is provided having a plurality of inputs and an output layer is provided having at least one output for providing the output parameter. A data set of historical data taken over a time line at periodic intervals is generated for use in training the model. The model is operable to map the input layer through a stored representation to the output layer. Training of the model involves training the stored representation on the historical data set to provide rejection of the disturbances in the stored representation.

Abstract: A method, system and computer program product are provided for scaling, or dimensionally reducing, multi-dimensional data sets that scale well for large data sets. The invention scales multi-dimensional data sets by determining one or more non-linear functions between a sample of points from the multi-dimensional data set and a corresponding set of dimensionally reduced points. Thereafter, these one or more non-linear functions are used to non-linearly map additional points. The additional points may be members of the original multi-dimensional data set or may be new, previously unseen points. In an embodiment, the determination of the non-linear relationship between the sample of points from the multi-dimensional data set and the corresponding set of dimensionally reduced points is performed by a self-learning system such as a neural network. The additional points are mapped using the self-learning system in a feed-forward/predictive manner.

Abstract: A data processing system program to develop, train, and implement a neural network for identifying customers who represent a bad debt risk is disclosed. A feature vector is applied to a neural network to generate outputs that approximate the relative likelihood that customers who are the subjects of the records used to generate the feature vector will be a bad debt risk. Statistical values relating categorical attributes of the customers to the likelihood of their becoming a bad debt risk are substituted for the categorical attributes, and the attributes are normalized before the feature vector is applied to the network. In one embodiment the customers are customers of a long distance service provider.

Abstract: A method for predicting a time to failure of a component in a system is presented. The method comprises obtaining a set of data measurements related to the component. The set of data measurements are representative of a plurality of parameters including a plurality of leading parameters. The method comprises generating a prediction model based upon the leading parameters considered in combination. The prediction model is then used to predict the time to failure of the component based on a set of real-time measurements, wherein the plurality of parameters are processed to predict the time to failure for the component. Finally, a confidence level for the predicted time to failure is determined based upon the plurality of parameters.

Abstract: A system and method is provided for monitoring the operating condition of a pump by evaluating fault data encoded in the instantaneous current of the motor driving the pump. The data is converted to a frequency spectrum which is analyzed to create a fault signature having fault attributes relating to various fault conditions associated with the pump. The fault signature is then input to a neural network that operates in conjunction with a preprocessing and post processing module to perform decisions and output those decisions to a user interface. A stand alone module is also provided that includes an adaptive preprocessing module, a one-shot unsupervised neural network and a fuzzy based expert system to provide a decision making module that operates with limited human supervision.

Abstract: A modified large margin perceptron learning algorithm (LMPLA) uses asymmetric margin variables for relevant training documents (i.e., referred to as “positive examples”) and non-relevant training documents (i.e., referred to as “negative examples”) to accommodate biased training sets. In addition, positive examples are initialized to force at least one update to the initial weighting vector. A noise parameter is also introduced to force convergence of the algorithm.

Abstract: An adaptive learning method for automated maintenance of a neural net model is provided. The neural net model is trained with an initial set of training data. Partial products of the trained model are stored. When new training data are available, the trained model is updated by using the stored partial products and the new training data to compute weights for the updated model.

Abstract: A method of training neural systems and estimating regression coefficients of regression models with respect to an error criterion is disclosed. If the error criterion is a risk-averting error criterion, the invented method performs the training/estimation by starting with a small value of the risk-sensitivity index of the risk-averting error criterion and gradually increasing it to ensure numerical feasibility. If the error criterion is a risk-neutral error criterion such as a standard sum-of-squares error criterion, the invented method performs the training/estimation first with respect to a risk-averting error criterion associated with the risk-neutral error criterion. If the result is not satisfactory for the risk-neutral error criterion, further training/estimation is performed either by continuing risk-averting training/estimation with decreasing values of the associated risk-averting error criterion or by training/estimation with respect to the given risk-neutral error criterion or by both.

Abstract: The present invention is directed to a method and mechanism for encoding multiple virtual tables into one or more source tables. An aspect of the invention is directed to the access of a virtual table virtual table without requiring separate entries for the virtual table in the meta-data of a database system. Another aspect of the invention is directed to a virtual table that has different column signatures than its underlying source table(s).

Abstract: A method learns a binary classifier for classifying samples into a first class and a second class. First, a set of training samples is acquired. Each training sample is labeled as either belonging to the first class or to the second class. Pairs of dyadic samples are connected by projection vectors such that a first sample of each dyadic pair belonging to the first class and a second sample of each dyadic pair belonging to the second class. A set of hyperplanes are formed so that the hyperplanes have a surface normal to the projection vectors. One hyperplane from the set of hyperplanes is selected that minimizes a weighted classification error. The set of training samples is then weighted according to a classification by the selected hyperplane. The selected hyperplanes are combined into a binary classifier, and the selecting, weighting, and combining are repeated a predetermined number of iterations to obtain a final classifier for classifying test samples into the first and second classes.

Abstract: The present invention provides a test bitstream generator and method for generating test bitstreams to test a bitstream decoder that is arranged to decode bitstreams generated in accordance with a predefined syntax. The method comprises the steps of generating test code from the syntax, the test code being arranged when executed to generate a test bitstream dependent on values assigned to a plurality of variables, each variable having a number of interesting values. The method then involves executing the test code, including the step of, for each of said variables, assigning that variable one of its interesting values, in order to thereby generate a test bitstream dependent on the interesting value assigned to each variable. It has been found that this technique provides an automated process for extensively testing a bitstream decoder.

Abstract: A system, software module, and computer program product for performing neural network based data mining that improved performance in model building, good integration with the various databases throughout the enterprise, flexible specification and adjustment of the models being built, and flexible model arrangement and export capability.

Abstract: The present invention relates to a system and methodology to facilitate multiattribute adjustments and control associated with messages and other communications and informational items that are directed to a user via automated systems. An interface, specification language, and controls are provided for defining a plurality of variously configured groups that may attempt to communicate respective items. Controls include the specification of priorities and preferences as well as the modification of priorities and preferences that have been learned from training sets via machine learning methods. The system provides both a means for assessing parameters used in the control of messaging and communications and for the inspection and modification of parameters that have been learned autonomously.

Abstract: The present disclosure provides a job flow system for use in a manufacturing environment, such as a semiconductor fab. The job flow system includes a plurality of sequence-related jobs associated with the manufacturing and a computer-controlled Petri Net structure. The Petri Net structure includes a plurality of agents associated with each of the sequence-related jobs. The Petri Net structure also includes a plurality of application processes to be performed by the agents, one or more description files, and a PN Center for loading the one or more description files and activating a first agent to perform one or more of the application processes in response to the one or more description files and in response to process status information from application processes.

Abstract: A student neural network that is capable of receiving a series of tutoring inputs from one or more teacher networks to generate a student network output that is similar to the output of the one or more teacher networks. The tutoring inputs are repeatedly processed by the student until, using a suitable method such as back propagation of errors, the outputs of the student approximate the outputs of the teachers within a predefined range. Once the desired outputs are obtained, the weights of the student network are set. Using this weight set the student is now capable of solving all of the problems of the teacher networks without the need for adjustment of its internal weights. If the user desires to use the student to solve a different series of problems, the user only needs to retrain the student by supplying a different series of tutoring inputs.

Abstract: A stress/strain curve is established by means of neural networks 1 to N and 4. To that end, parameters are input into the input 50, from which the neural networks 1 to N respectively establish the principal components of characteristic points. The curve type is selected on the basis of the output of the neural network 4. The principal components of the characteristic points of the corresponding curve type are then inverse-transformed. The stress/strain curve is then calculated by the generator 59 on the basis of the inverse transformation.

Abstract: Method and apparatus for training a system model with gain constraints. A method is disclosed for training a steady-state model, the model having an input and an output and a mapping layer for mapping the input to the output through a stored representation of a system. A training data set is provided having a set of input data u(t) and target output data y(t) representative of the operation of a system. The model is trained with a predetermined training algorithm which is constrained to maintain the sensitivity of the output with respect to the input substantially within user defined constraint bounds by iteratively minimizing an objective function as a function of a data objective and a constraint objective. The data objective has a data fitting learning rate and the constraint objective has constraint learning rate that are varied as a function of the values of the data objective and the constraint objective after selective iterative steps.

Abstract: A CAI system for vehicle repair workers is disclosed. The CAI system for supporting an instruction for one group consisting of a plurality of training participants includes a storage device for storing attribute data of each participant, a keyboard for inputting degree-of-master basic data of each participant for every plurality of instruction items, and a control unit for processing the data, the control unit evaluates a degree of master of each participant for every instruction item, calculates an overall evaluation of the concerned instruction item of the target group, selects the instruction item exhibiting a low degree of master by comparing the overall evaluations of the concerned instruction item, and selects which instruction target person should be instructed with respect to the selected instruction item. The CAI system is capable of efficiently enhancing a working efficiency of one working group consisting of the plurality of workers.

Abstract: A system and process for tracking an object state over time using particle filter sensor fusion and a plurality of logical sensor modules is presented. This new fusion framework combines both the bottom-up and top-down approaches to sensor fusion to probabilistically fuse multiple sensing modalities. At the lower level, individual vision and audio trackers can be designed to generate effective proposals for the fuser. At the higher level, the fuser performs reliable tracking by verifying hypotheses over multiple likelihood models from multiple cues. Different from the traditional fusion algorithms, the present framework is a closed-loop system where the fuser and trackers coordinate their tracking information. Furthermore, to handle non-stationary situations, the present framework evaluates the performance of the individual trackers and dynamically updates their object states.

Abstract: A computer based system and method provides color solutions to a customer. A solution request from an operator located at a remote location is received and delivered to a central location over a computer network. A composite solution database is searched to determine a color solution as a function of the solution request.

Abstract: Digitized image data are input into a processor where a detection component identifies the areas (objects) of particular interest in the image and, by segmentation, separates those objects from the background. A feature extraction component formulates numerical values relevant to the classification task from the segmented objects. Results of the preceding analysis steps are input into a trained learning machine classifier which produces an output which may consist of an index discriminating between two possible diagnoses, or some other output in the desired output format. In one embodiment, digitized image data are input into a plurality of subsystems, each subsystem having one or more support vector machines. Pre-processing may include the use of known transformations which facilitate extraction of the useful data. Each subsystem analyzes the data relevant to a different feature or characteristic found within the image.

Abstract: A system and method for converting a color formula from compositions such as paints, pigments, or dye formulations, is provided. The input to the system is a first color formula. The system includes an input device for entering a plurality of color formula values and an artificial intelligence conversion model coupled to the input device. The conversion model produces an output signal for communicating a second color formula. The artificial intelligence model may be embodied in a neural network. More specifically, the conversion model may be a back propagation neural network.

Abstract: Power spectrum data is calculated from a digital audio signal D10. A part of power spectrum data is extracted from thus calculated power spectrum data. Classification is made based on the extracted part of power spectrum data. And the digital audio signal D10 is converted by a predicting method that corresponds to the classified class. Thereby, conversion further adapted to the characteristic of the digital audio signal D10 can be performed.

Abstract: The current invention is directed to a system for adaptive medical decision support. The invented system provides a system that allows users to efficiently enter, access, and analyze medical information, without disrupting patient-doctor interactions or medical facility course of business; which assists in all stages of medical assessment and treatment; and which is tailored to the particular medical practice or specialty and taking into account the developing habits, preferences, performance, and individual patient histories, of an individual user. The invention provides a learning capacity configured to learn previously presented data and decisions and predict data or decisions based on data that it receives from the user, thereby adapting its operations to the developing habits, preferences, performance, and individual patient histories of an individual user.

Abstract: A computer based system and method provides color solutions to a customer over a computer network. A solution request is received from an operator. A composite solution database is searched to determine a color solution as a function of the solution request.

Abstract: An adaptive process controller drives a process variable to be substantially equivalent to a set point and adapts the controller gain, the controller reset, and/or the controller rate, based on model free adaptation. The adaptive controller combines a controller gain computed from an oscillation index with a controller gain computed from a steady state estimate and that adapts the controller reset/rate by forcing the ratio of two of the controller proportional, integral or derivative terms to be equal to a predetermined value.

Abstract: A method for training a neural network in order to optimize the structure of the neural network includes identifying and eliminating synapses that have no significant influence on the curve of the risk function. First and second sending neurons are selected that are connected to the same receiving neuron by respective first and second synapses. It is assumed that there is a correlation of response signals from the first and second sending neurons to the same receiving neuron. The first synapse is interrupted and a weight of the second synapse is adapted in its place. The output signals of the changed neural network are compared with the output signals of the unchanged neural network. If the comparison result does not exceed a predetermined level, the first synapse is eliminated, thereby simplifying the structure of the neural network.

Abstract: The present invention is directed to a system, method, and apparatus for orienting images. A neural net is trained with images of known orientation and an indicator indicating such known orientation. Images of unknown orientation are then input to the neural net and the orientation is determined based on the output of the neural net.

Abstract: The invention relates to a method of transmitting teaching materials by means of an electronic data network, wherein a person to be taught is first shown the teaching material on a client computer system. After the teaching material has been displayed, the person to be taught has the possibility of undergoing a test on the displayed teaching material. The invention also describes a computer network with a server computer system as well as a server computer system.

Abstract: The learning rate used for updating the weights of a self-ordering feature map is determined by a process that injects some type of perturbation into the value so that it is not simply monotonically decreased with each training epoch. For example, the learning rate may be generated according to a pseudorandom process. The result is faster convergence of the synaptic weights.

Abstract: A system and method for electronic communication management comprises a universal data model, a modeling engine, and an adaptive knowledge base. The modeling engine includes a natural language processor and a statistical modeler. A communication is translated from its native format into the universal data model. The modeling engine determines the intent of the communication using the natural language processor and the statistical modeler. A response is generated, either automatically or by an agent. An audit module analyzes each response and provides feedback to the modeling engine and the adaptive knowledge base. The modeling engine uses the feedback to update models in the adaptive knowledge base. The modeling engine supports various application specific modules.

Abstract: A process for modeling numerical data for forecasting a phenomenon relates to constructing a model by processing and learning on collected data. The fit and robustness of the model are evaluated and the model parameters are adjusted to select an optimal model in the form of a Dth order polynomial. A trade-off between learning accuracy and learning stability is controlled by adding to a covariance matrix a perturbation in the form of the product of a scalar ? times a matrix H or in the form of a matrix H dependent on a vector of k parameters ?=(?1, ?2, . . . ?k). A data partition step can divide the data into a first subset for constructing the model and a second subset for adjusting the value of the model parameters according to a validity criterion obtained from data that was not used to construct the model.

Abstract: A technique for machine learning, such as supervised artificial neural network learning includes receiving data and checking the dimensionality of the read data and reducing the dimensionality to enhance machine learning performance using Principal Component Analysis methodology. The technique further includes specifying the neural network architecture and initializing weights to establish a connection between read data including the reduced dimensionality and the predicted values. The technique also includes performing supervised machine learning using the specified neural network architecture, initialized weights, and the read data including the reduced dimensionality to predict values. Predicted values are then compared to a normalized system error threshold value and the initialized weights are revised based on the outcome of the comparison to generate a learnt neural network having a reduced error in weight space.

Abstract: A system is disclosed that provides a goal based learning system utilizing a rule based expert training system to provide a cognitive educational experience. The system provides the user with a simulated environment that presents a business opportunity to understand and solve optimally. Mistakes are noted and remedial educational material presented dynamically to build the necessary skills that a user requires for success in the business endeavor. The system utilizes an artificial intelligence engine driving individualized and dynamic feedback with synchronized video and graphics used to simulate real-world environment and interactions. Multiple “correct” answers are integrated into the learning system to allow individualized learning experiences in which navigation through the system is at a pace controlled by the learner.

Abstract: A method and apparatus for producing a keyword dictionary using a computer includes inputting a user's necessity or a user's non-necessity for each of pieces of information data, wherein one or more keywords are attached to the information data. A summed necessity value relating to the user's necessities inputted for the pieces of information data and a summed non-necessity value relating to the user's non-necessities inputted for the pieces of information data are calculated. A difference between the summed necessity value and the summed non-necessity value for each of the keywords is calculated, and a predication value predicting a user's necessity degree for each of the keywords according to the differences, respectively, relating to the keywords, is calculated. Each prediction value calculated is assigned to the corresponding keyword.

Abstract: A system and method for historical database training of a support vector machine (SVM). The SVM is trained with training sets from a stream of process data. The system detects availability of new training data, and constructs a training set from the corresponding input data. Over time, many training sets are presented to the SVM. When multiple presentations are needed to effectively train the SVM, a buffer of training sets is filled and updated as new training data becomes available. Once the buffer is full, a new training set bumps the oldest training set from the buffer. The training sets are presented one or more times each time a new training set is constructed. A historical database of time-stamped data may be used to construct training sets for the SVM. The SVM may be trained retrospectively by searching the historical database and constructing training sets based on the time-stamped data.

Abstract: A computer-implemented method and system for building a neural network is disclosed. The neural network predicts at least one target based upon predictor variables defined in a state space. First, an input data set is retrieved that includes the predictor variables and at least one target associated with the predictor variables for each observation. In the state space, a number of points is inserted in the state space based upon the values of the predictor variables. The number of points is less than the number of observations. A statistical measure is determined that describes a relationship between the observations and the inserted points. Weights and activation functions of the neural network are determined using the statistical measure.

Abstract: Apparatus for adding new learning tasks to an incremental supervised learner provides a flexible incremental representation of all encountered training examples, thereby permitting state representations for new learning tasks to take advantage of incremental training already completed by encoding all past training examples as negative examples for a hypothetical learning task. The state representation of the hypothetical learning task is copied as the initial state representation for a new learning task to be initiated, and is initialized with negative training examples of all previously presented training examples, thereby permitting the learning task to efficiently incorporate the previous examples.

Abstract: A method and system are described for end-user transaction recognition based on server data such as sequences of remote procedure calls (RPCs). The method may comprise machine-learning techniques for pattern recognition such as Bayesian classification, feature extraction mechanisms, and a dynamic-programming approach to segmentation of RPC sequences. The method preferably combines information-theoretic and machine-learning approaches. The system preferably includes a learning engine and an operation engine. A learning engine may comprise a data preparation subsystem (feature extraction) and a Bayes Net learning subsystem (model construction). The operation engine may comprise transaction segmentation and transaction classification subsystems.

Abstract: A method and a system are presented, which assist classifiers in gathering information in a cost-effective manner by determining which piece of information, if any, to gather next. The system includes an explicit system, an implicit system, a classifier, and a profit module. A feature set is inputted into the explicit system, which uses the feature set to determine tests to perform to gather information useful for classifying the system state. The relative profit of each test performed is determined by the profit module and the profit determined is used to determine which test or tests to select for a particular feature set. The results of the explicit system, which is generally an exhaustive or semi-exhaustive search algorithm, are used to train the implicit system. The implicit system is then able to process decisions much faster than the explicit system when circumstances require time-critical operation.

Abstract: The present invention relates to a method for determining the precompensated pattern of exposure doses of an electron beam required per pattern position to obtain a desired pattern in a coating on a substrate, comprising the steps of: determining a smearing function of the electron beam; determining a precompensated pattern with the smearing function and the desired pattern, wherein the determination is performed such that exposure doses contain almost exclusively positive values and the exposure doses are smooth relative to each other.

Abstract: A system and method which partitions a parameter estimation model, a fault detection model, and a fault classification model for a process surveillance scheme into two or more coordinated submodels together providing improved diagnostic decision making for at least one determined operating mode of an asset.

Abstract: A computerized method and system for solving non-linear Boolean equations is disclosed comprising at least partially solving a Boolean function; developing at least one inference regarding said Boolean function and saving said inference to a state machine; and accessing said inference from said state machine to develop at least one heuristic for determining whether said Boolean function is satisfiable.

Abstract: Power spectrum data is calculated from a digital audio signal D10. A part of power spectrum data is extracted from thus calculated power spectrum data. Classification is made based on the extracted part of power spectrum data. And the digital audio signal D10 is converted by a predicting method that corresponds to the classified class. Thereby, conversion further adapted to the characteristic of the digital audio signal D10 can be performed.