Abstract: A method is provided herein for determining and recognizing types of vehicles passing a check point. The method takes advantage of an EM algorithm which is up-loaded into a CPU and which processes data of the vehicles which drive past a checkpoint, the data being representative of essential characteristics of vehicles to produce an output model of the traffic volumes of the various types of vehicles. This model enables the forecasting of future road maintenance costs and the planning and designing of future road networks.

Abstract: An occupancy estimator calculates an occupancy estimate (x) of a region based on sensor data (z) provided by one or more sensor devices and a model-based occupancy estimate generated by an occupant traffic model (f). The occupant traffic model (f) is based on predicted movement of occupants throughout a region. The occupancy estimation system includes an occupancy estimator algorithm (20) that combines the sensor data (z) and the model-based occupancy estimate generated by the occupant traffic model (f) to generate an occupancy estimate (x) for the region.

Abstract: A Self-Organizing Process Control Architecture is introduced with a Sensing Layer, Control Layer, Actuation Layer, Process Layer, as well as Self-Organizing Sensors (SOS) and Self-Organizing Actuators (SOA). A Self-Organizing Sensor for a process variable with one or multiple input variables is disclosed. An artificial neural network (ANN) based dynamic modeling mechanism as part of the Self-Organizing Sensor is described. As a case example, a Self-Organizing Soft-Sensor for CFB Boiler Bed Height is presented. Also provided is a method to develop a Self-Organizing Sensor.

Abstract: Interactive database viewing is provided. A server associated with a database accesses a stored result set from a database query. The server presents a portion of the result set in a list view to a client device remote from the server and the cache memory, where the portion presented is based on filter selection based on fields of the data. In response to changes to the filter selection, the server can dynamically alter the list view without having to again access the database. If a change to the filter selection requests data outside the cached result set, the server accesses the database and updates the cached result set. Inline editing of the data in the list view can be applied back to the database by the server.

Abstract: A method and system for identifying a relevance of a relation between at least two entities includes receiving at least one item of information relating to one or more of the entities and determining whether a proximity between the at least two entities exists. A level of the proximity between the entities is identified. The relevance of the proximity between the entities is determined based on the level of the proximity and the at least one item of information received.

Abstract: A system for estimating a strain of a component and method of estimating strain is provided. The system includes a signal generator configured to transmit a signal toward the component. A sensor is coupled to the component and configured to receive the signal and to generate a reflected signal. The system includes a fiber Bragg grating filter coupled to the sensor and configured to filter the reflected signal and to generate a filtered signal. A detector is coupled to the filter and configured to convert the filtered signal to a time domain signal. The system includes an artificial neural network coupled to the detector and configured to process the time domain signal to facilitate estimating the strain of the component.

Abstract: The position calculation of prior art position sensitive detector systems relies on a known geometry pattern of individual electrodes and the distribution of the charge parts. A heuristic estimation is made in order to calculate an initial coordinate of irradiation. In contrast, the present invention allows one to calculate the position of an incident particle in terms of direct mapping of the measured detector response into position coordinates detector surface. The device for estimating the space coordinates of an irradiation position onto a detector comprises a position sensitive detector; an irradiation source; means for measuring the response of detector generated upon irradiation by irradiation source; and an artificial neural network structure provided such that the measured detector response is the input to the artificial neural network structure and the initial space coordinates of irradiation are the output of the artificial neural network structure.

Abstract: Embodiments of the invention relate to producing spike-timing dependent plasticity in an ultra-dense synapse cross-bar array for neuromorphic systems. An aspect of the invention includes when an electronic neuron spikes, an alert pulse is sent from the spiking electronic neuron to each electronic neuron connected to the spiking electronic neuron. When the spiking electronic neuron sends the alert pulse, a gate pulse is sent from the spiking electronic neuron to each electronic neuron connected to the spiking electronic neuron. When each electronic neuron receives the alert pulse, a response pulse is sent from each electronic neuron receiving the alert pulse to the spiking electronic neuron. The response pulse is a function of time since a last spiking of the electronic neuron receiving the alert pulse. In addition, the combination of the gate pulse and response pulse is capable increasing or decreasing conductance of a variable state resistor.

Abstract: A method and a system for standardizing data used for monitoring an aeroengine, and including: operating over time to collect time-series measurements from the aeroengine; calculating from the time-series measurements a set of indicators Y=(y1, . . . , yj, . . . , ym) specific to elements of the engine; identifying from the time-series measurements an exogenous data set X=(x1, . . . , xn) representative of external context acting on the set of indicators Y; defining a conditional multidimensional model simultaneously handling the indicators of the set of indicators Y while taking account of the exogenous data set X to form a set of estimators ?=(?1, . . . , ?j, . . . , ?m) corresponding to the set of indicators Y=(y1, . . . , yj, . . . , ym); and normalizing each estimator ?j as a function of a reference value for the corresponding indicator yj and of a difference between each estimator ?j and corresponding indicator yj so as to form a set of standardized values {tilde over (Y)}=({tilde over (y)}1, . . .

Abstract: A computationally implemented method includes, but is not limited to: selecting at least one hypothesis from a plurality of hypotheses relevant to a user, the selection of the at least one hypothesis being based, at least in part, on at least one reported event associated with the user; and presenting one or more advisories related to the hypothesis. In addition to the foregoing, other method aspects are described in the claims, drawings, and text forming a part of the present disclosure.

Abstract: Real-time condition-based analysis is performed on a machine for providing diagnostic and prognostic outputs indicative of machine status includes a signal processor for receiving signals from sensors adapted for measuring machine performance parameters. The signal processor conditions and shapes at least some of the received signals into an input form for a neural network. A fuzzy adaptive resonance theory neural network receives at least some of the conditioned and shaped signals, and detects and classifies a state of the machine based upon the received conditioned and shaped signals, and upon a predetermined ontology of machine states, diagnostics, and prognostics. The neural network can also determine from the machine state a health status thereof, which can comprise an anomaly, and output a signal representative of the determined health status. A Bayesian intelligence network receives the machine state from the neural network and determines a fault probability at a future time.

Abstract: In various embodiments, a method for compressed transmission of data packet header fields in a packet-oriented data stream may comprise: estimating in advance a data packet header field in a packet-oriented data stream from at least one preceding data packet header field; forming a piece of comparison information which indicates the difference between the data packet header field and the data packet header field estimated in advance using the transmitter neural network; transmitting the comparison information as a compressed data packet header field from the transmitter to a receiver; estimating in advance the data packet header field from at least one already transmitted data packet header field in the packet-oriented data stream using the receiver neural network; and generating the data packet header field from the data packet header field estimated in advance using the neural network of the receiver and from the transmitted piece of comparison information.

Abstract: Methods, systems, and apparatus, including computer programs encoded on one or more computer storage devices, for training and retraining predictive models. A series of training data sets for predictive modeling can be received, e.g., over a network from a client computing system. The training data included in the training data sets is different from initial training data that was used with multiple training functions to train multiple trained predictive models stored in a predictive model repository. The series of training data sets are used with multiple trained updateable predictive models obtained from the predictive model repository and multiple training functions to generate multiple retrained predictive models. An effectiveness score is generated for each of the retrained predictive models. A first trained predictive model is selected from among the trained predictive models included in the predictive model repository and the retrained predictive models based on their respective effectiveness scores.

Abstract: A gas concentration estimation device of a coal-burning boiler adapted to estimate the concentration of the gas component included in an exhaust gas emitted from a coal-burning boiler using a neural network, including: a process database section adapted to store process data of a coal-burning boiler; a filtering processing section adapted to perform filtering processing for extracting data suitable for learning of a neural network from the process data stored in the process database section; a neural-network learning processing section adapted to perform learning processing of the neural network based on the data extracted by the filtering processing section and suitable for learning of the neural network; and a neural-network estimation processing section adapted to perform estimation processing of the CO concentration or the NOx concentration in the exhaust gas emitted from the coal-burning boiler based on the learning processing of the neural-network learning processing section.

Abstract: A method comprises: partitioning a region of interest into a plurality of soft bin regions that span the region of interest; estimating an integral over each soft bin region of a function defined over the region of interest; and outputting a value equal to or derived from the sum of the estimated integrals over the soft bin regions spanning the region of interest. The method may further comprise: integrating a Bayesian theorem function using the partitioning, estimating, and outputting operations, and classifying an object to be classified using a classifier trained using the Bayesian machine learning. The method may further comprise performing optimal control by iteratively minimizing a controlled system cost function to determine optimized control inputs using the partitioning, estimating, and outputting with the function equal to the controlled system cost function having the selected control inputs, and controlling the controlled system using the optimized control inputs.

Abstract: The invention comprises methods and devices of forecasting future computational needs of an application based on input load, where the application comprises a plurality of processes executed on a plurality of computing devices. The method of the invention proceeds by monitoring at least a computational load characteristic of at least a first process executed on a first computing device and a second process executed on a second computing device. A mathematical relationship between input load and the computational load characteristic is established, and future computational needs are forecasted based on the established mathematical relationship.

Abstract: Embodiments relate to systems and methods for training a self-learning network using interpolated input sets based on a target output. A database management system can store sets of operational data, such as financial, medical, climate or other information. A user can input or access a set of target data, representing an output which a user wishes to be generated from an interpolated set of input data. The interpolation engine can generate a conformal interpolation function and input sets that map to the set of target output data. After interpolation, the interpolation engine can transmit the interpolated inputs, along with the set of target output data and other information, to a self-learning network such as a neural or fuzzy logic network. The self-learning network can be trained to converge to the target output based on the interpolated input results as generated by the interpolation engine, thus reproducing the desired interpolation function.

Abstract: The present invention provides a method of real-time crystal peak tracking for avalanche-photodiode (APD) detectors on positron emission tomography (PET) scanners that satisfies the need to compensate for the significant gain drifting due to thermal variations in APD detectors on PET scanners.

Abstract: A method and a system for standardizing data used for monitoring an aeroengine, and including: operating over time to collect time-series measurements from the aeroengine; calculating from the time-series measurements a set of indicators Y=(y1, . . . , yj, . . . , ym) specific to elements of the engine; identifying from the time-series measurements an exogenous data set X=(x1, . . . , xn) representative of external context acting on the set of indicators Y; defining a conditional multidimensional model simultaneously handling the indicators of the set of indicators Y while taking account of the exogenous data set X to form a set of estimators ?=(?1, . . . , ?j, . . . , ?m) corresponding to the set of indicators Y=(y1, . . . , yj, . . . , ym); and normalizing each estimator ?j as a function of a reference value for the corresponding indicator yj and of a difference between each estimator ?j and corresponding indicator yj so as to form a set of standardized values {tilde over (Y)}=({tilde over (y)}1, . . .

Abstract: The position calculation of prior art position sensitive detector systems relies on a known geometry pattern of individual electrodes and the distribution of the charge parts. A heuristic estimation is made in order to calculate an initial coordinate of irradiation. In contrast, the present invention allows one to calculate the position of an incident particle in terms of direct mapping of the measured detector response into position coordinates detector surface. The device for estimating the space coordinates of an irradiation position onto a detector comprises a position sensitive detector; an irradiation source; means for measuring the response of detector generated upon irradiation by irradiation source; and an artificial neural network structure provided such that the measured detector response is the input to the artificial neural network structure and the initial space coordinates of irradiation are the output of the artificial neural network structure.

Abstract: Techniques for identifying potential communication uses of various systems are disclosed. Identifying potential communication uses of a computing system can improve the manner in which the computing system is used by allowing more intelligent decisions and better choices to be made regarding its communication use. By way of example, communication applications (or tasks or services) that are likely (or more likely) to be used by a person in a particular situation can be identified as potential communication use of a particular device. Such potential uses can, for example, be made more assessable (or more readily available) and/or effectively recommended (or automatically initiated), thereby allowing a person to more conveniently use the device. By way of example, identifying communication applications or tasks that are likely to be used by a person in a particular situation for various reasons (e.g., preferences and/or habits of a person in a particular situation) as potential communication use of a system (e.

Abstract: Techniques for identifying potential uses of computing systems are disclosed. A potential use of a computing system can be identified by considering the context that effectively represents a situation for the computing system and further considering a known situation to be a match for the situation. As such, a potential state of use of the computing system can be identified based on one or more known states of use of the computing system. A matching situation can, for example, be determined, based on pattern of use data that effectively associates a state of use with a situation in which the use may or has occurred. Identifying a potential state of use, among other things, allows making various applications, tasks, and/or services more accessible and/or effectively recommending what is likely to be used for a given situation without requiring supervised training or direct feedback from a user.

Abstract: A method, computer program product, and system are disclosed for dynamically provisioning clusters of middleware appliances. In one embodiment, the method includes referencing a resource measurement from a plurality of middleware appliances. The middleware appliances process one or more service domains and the resource measurement includes processing resources consumed by each middleware appliance for each of the one or more service domains. The method may also include determining an implementation plan based on a performance goal and one or more resource calculations. The implementation plan specifies service domain instances to activate and service domain instances to deactivate on the plurality of middleware appliances. The method may also include dynamically enabling and disabling the service domain instances on the plurality of middleware appliances based on the implementation plan.

Abstract: A system for estimating long term characteristics of a battery includes a learning data input unit for receiving initial characteristic learning data and long term characteristic learning data of a battery to be a learning object; a measurement data input unit for receiving initial characteristic measurement data of a battery to be an object for estimation of long term characteristics; and an artificial neural network operation unit for receiving the initial characteristic learning data and the long term characteristic learning data from the learning data input unit to allow learning of an artificial neural network, receiving the initial characteristic measurement data from the measurement data input unit and applying the learned artificial neural network thereto, and thus calculating long term characteristic estimation data from the initial characteristic measurement data of the battery and outputting the long term characteristic estimation data.

Abstract: The American legal system, judges and lawyers are continually researching an ever-expanding body of past judicial opinions, or case law, for the ones most relevant to resolution of new disputes. To facilitate these searches, some companies collect and publish the judicial opinions of courts across the United States in both paper and electronic forms, with some of the cases containing references to prior cases from other courts that have previously ruled on all or part of the same dispute. Identifying the prior cases is problematic, because, for example, conventional computer text-matching not only suggests too many non-prior cases, but also misses too many actual prior cases. Accordingly, the present inventors devised systems, methods, and software that generally facilitate identification of one or more documents that are related to a given document, and particularly facilitate identification of prior cases for a given case.

Abstract: A neural network process for improving wind retrievals from wind profiler measurements is described. In this invention, a neural network is trained to retrieve (missing or incomplete) upper level winds from ground based wind profiler measurements. Radiosonde measurements in conjunction with wind profiler ground measurements for specific geographical locations are used as training sets for the neural network. The idea is to retrieve timely and spatially continuous upper level wind information from (fragmented or incomplete) wind profiler measurements.

Abstract: System, method and device for reducing the time required for solution of problems in the NP complexity class to polynomial time. Within satisfaction problems or problems reducible to a satisfaction problem, the invention tracks the sources of implications and identifies proximal parameterizations of conditional contradictions and subsequently avoids those contradictory conditions. The action is completed in less time than is incurred by existing methods and thus provides a performance improvement to the devices, software, or processes which address such problems.

Abstract: Embodiments of the invention relate to producing spike-timing dependent plasticity in an ultra-dense synapse cross-bar array for neuromorphic systems. An aspect of the invention includes when an electronic neuron spikes, an alert pulse is sent from the spiking electronic neuron to each electronic neuron connected to the spiking electronic neuron. When the spiking electronic neuron sends the alert pulse, a gate pulse is sent from the spiking electronic neuron to each electronic neuron connected to the spiking electronic neuron. When each electronic neuron receives the alert pulse, a response pulse is sent from each electronic neuron receiving the alert pulse to the spiking electronic neuron. The response pulse is a function of time since a last spiking of the electronic neuron receiving the alert pulse. In addition, the combination of the gate pulse and response pulse is capable increasing or decreasing conductance of a variable state resistor.

Abstract: Apparatus and methods to perform downhole fluid analysis using an artificial neural network are disclosed. A disclosed example method involves obtaining a first formation fluid property value of a formation fluid sample from a downhole fluid analysis process. The first formation fluid property value is provided to an artificial neural network, and a second formation fluid property value of the formation fluid sample is generated by means of the artificial neural network.

Abstract: A virtual analyzer is provided to estimate either an attribute of a reactant applied during performance of, or an amount of a reactant exhausted by, a process having multiple process parameters (MPPs) that is performed to control an amount of a pollutant emitted into the air. The virtual analyzer includes an interface which receives signals corresponding to attributes of the MPPs. If the process is a wet flue gas desulfurization (WFGD) process, the signals include a signal corresponding to a measured pH level of the applied reactant. If the process is a selective catalytic reduction (SCR) process, the signals include a signal corresponding to a measured amount of the reactant exhausted by the process.

Abstract: A method and mechanism that protect a person with a property interest in a piece of real property against a loss of expected future market value of the real property. A policyholder may first select an expected future market value for the real property and a time period in which the expected future market value is to be attained (e.g., 10 years). An insurer may model relevant conditions, factors and events that effect property values over time for a particular geographic region. A premium may be determined based on the selected expected future market value and the determination of expected future market value made by the insurer. Upon a conveyance of the real property at the time period or window specified, a loss may be determined if the fair market value of the real property is less than the selected expected future market value.

Abstract: It is difficult to select parameter values for constraint programming problem solvers which will yield good performance. Automated tuning of such problem solvers on a per problem instance basis may be used and this involves learning a function for predicting the runtime of a problem solver depending on parameter values of the problem solver and features of the problem instance being solved. However, it takes time for such prediction functions to be learnt, either during operation of a problem solver or offline, using specified examples. To address this, information about such a prediction function is shared between two or more problem solvers to improve performance. A sharing system may be used to receive prediction function information and send this to problem solvers.

Abstract: A method for creating a risk estimate for a tool includes creating a plurality of source patterns from tool data and maintenance data related to a plurality of tools. The method also includes creating a risk model from the plurality of source patterns, the risk model including a plurality of example stressors each having an associated risk value. The method also includes creating at least one stress pattern from tool data related to the tool and comparing the at least one stress pattern to the risk model to create a risk estimate for the tool.

Abstract: When a patient enters a medical situation, healthcare professionals can use various amounts of information in evaluating the situation. However, different information can be beneficial dependent on the medical situation. Moreover, personnel can historically use specific information types regardless of the situation. An artificial neuron network is employed to pre-fetch information that personnel likely will want prior to a request from the personnel. In addition, the artificial neuron network can be trained based on results of presented information.

Abstract: A method and system of creating an approximate kernel matrix to train a kernel machine. In one embodiment of the invention, a set of kernel machine training data is partitioned into a set of partitioned training data based on a set of partition parameters. The set of partition parameters includes one or more of, an axis-aligned grid location, or an axis-aligned grid resolution in one embodiment of the invention. A partition matrix that approximates a kernel machine kernel matrix is created from the set of partitioned training data and the partition matrix is used to train a kernel machine in one embodiment of the invention.

Abstract: A recognizing apparatus includes a training vector input unit configured to enter training vectors for training, a weak classifier generator configured to obtain weak classifiers based on the value of an element of a dimension common to the training vectors using a learning method, a classifier integrator configured to obtain non-linear mappings for each dimension of the training vectors by combining the weak classifiers, a test vector input unit configured to input a test vector to be classified, a non-linear transformer configured to a transformed vector by transforming the values of the elements of the test vector using the respective non-linear mappings, and a score calculator configured to obtain a classification score by summing the value of the respective elements of the transformed vector and recognize the test vector using the classification score.

Abstract: An empirical ensemble based virtual sensor system (VS) for the estimation of an amount of a gas (G) resulting from a combustion process (CP) comprising two or more empirical models (NN1, NN2, . . . , NNn). The amount of gas (G) is estimated in each of the empirical models (NN1, NN2, . . . , NNn), and a combination function (f) combines the results from the empirical models (NN1, NN2, . . . , NNn) to provide a combined estimate for the amount of gas (G) that is more accurate than the estimated amount of gas from each of the individual empirical models (y1, y2, . . . , ym). The total performance of the virtual sensor system (VS) may be increased by increasing the number of empirical models (y1, y2, . . . , ym).

Abstract: An unsupervised classification approach is improved by imposing some order into the treatment of the records and their attributes, which otherwise would be treated as random variables. A method is provided to identify particular attributes that are most associated with the “good” records within each of the plurality of groups of records within a data set. Based on a supervised scoring method, the records of the data set are processed to indicate their measure of “goodness”. There are various ways by which the records can be processed to indicate a bias during unsupervised clustering processing.

Abstract: In general, the present invention provides a method, system and program product for approximating/estimating computer resource consumption of a computer system. Specifically, under the present invention, a more efficient or reduced computer work gradient matrix (hereinafter “matrix”) is first built. This occurs by creating load measurements for a set of computer resource metrics of the computer system to analyze dependencies between different computer resource metrics. Then, a correlation matrix between the set of computer resource metrics is created based on the dependencies. The set of computer system resource metrics in the correlation matrix is thereafter clustered into a set of clusters, and a reduced matrix is built based thereon. Once the reduced matrix is built, it can be restored to a “full” matrix using linear transformation or the like.

Abstract: A method and system are provided for characterizing a reservoir of interest by comparing measured downhole fluid analysis measurement data with predicted downhole fluid analysis measurement data for the corresponding depth within the reservoir. The downhole fluid analysis measurement data may comprise the results of compositional analysis, gas-oil ratio measurements, and spectrophotometry measurements. The compositional analyses may be delumped to characterize the compositional components of the downhole fluid and equations of state may be used to predict compositional gradients and fluid properties with depth. The method and system enable the user to characterize a reservoir as to its states of compartmentalization and equilibrium.

Abstract: A system and method is disclosed which integrates a machine learning solution into a large scale, distributed transaction processing system using a supporting architecture comprising a combination of computer hardware and software. Methods of using a system comprising such supporting architecture provide application designers access to the functionality included in a machine learning solution, but might also provide additional functionality not supported by the machine learning solution itself.

Abstract: The invention is directed to a method for training at least one learning-capable system comprising the steps of providing a predetermined training data set corresponding to a predetermined number of subjects comprising a predetermined input data set and a predetermined outcome data set, augmenting the input data set and/or the outcome data set, and training each learning-capable system using the augmented input data set and/or the augmented outcome data set.

Abstract: A method and apparatus are disclosed for modeling a system to estimate values and associated uncertainties for a first set of variables describing the system. A second set of system variables is selected, where the second set is directly or indirectly causally related to the first set of variables. Data is obtained or estimated for each variable in the second set and the quality of selected data is appraised. A network is formed with nodes including both sets of variables and the quality appraisals, having directional links connecting interdependent nodes, the directional links honoring known causality relationships. A Bayesian Network algorithm is used with the data and quality information to solve the network for the first set of variables and their associated uncertainties.

Abstract: A method for predicting a future occurrence of an event involves obtaining a history of prior occurrences of the event. A plurality of variables is created that are associated with the event. Weights are assigned to each variable. An artificial neural network is accessed and trained with the history of past occurrences of the event by comparing an output of the artificial neural network to the past occurrence of the event. The weights are adjusted until the output corresponds to the past occurrence of the event.

Abstract: Methods of determining values of keywords in an internet search are described. According to one aspect of the invention, a method comprises steps of receiving keywords entered for a plurality of searches; detecting converted transactions associated with the plurality of searches; analyzing the converted transactions; and determining values associated with the keywords based upon the converted transactions. According to other aspects of the invention, methods for recommending subsets of keywords and for recommending keywords based upon converted transactions and click through rates are disclosed. A database search system is also disclosed.

Abstract: A neural network is used to combine one or more estimates of a physiologic parameter with one or more associated signal quality metrics, creating a more accurate estimate of said physiologic parameter, as well as a second estimate of the accuracy of said physiologic parameter estimate.

Abstract: A method of creating a non-invasive predictor of both physiologic and imposed patient effort from airway pressure and flow sensors attached to the patient using an adaptive mathematical model. The patient effort is commonly measured via work of breathing, power of breathing, or pressure-time product of esophageal pressure and is important for properly adjusting ventilatory support for spontaneously breathing patients. The method of calculating this non-invasive predictor is based on linear or nonlinear calculations using multiple parameters derived from the above-mentioned sensors.

Abstract: The present invention is a method and an apparatus for improved regression modeling to address the curse of dimensionality, for example for use in data analysis tasks. In one embodiment, a method for analyzing data includes receiving a set of exemplars, where at least two of the exemplars include an input pattern (i.e., a point in an input space) and at least one of the exemplars includes a target value associated with the input pattern. A function approximator and a distance metric are then initialized, where the distance metric computes a distance between points in the input space, and the distance metric is adjusted such that an accuracy measure of the function approximator on the set of exemplars is improved.

Abstract: A neural model for simulating a scorecard comprises a neural network for transforming one or more inputs into an output. Each input of the neural model has a squashing function applied thereto for simulating a bin of the simulated scorecard. The squashing function includes a control variable for controlling the steepness of the response to the squashing function's input so that during training of the neural model the steepness can be controlled. The output of the neural model represents the score of the simulated scorecard. The neural network is trained to behave like a scorecard by providing plurality of example values to the inputs of the neural network. Each output score produced is compared to an expected score to produce an error value. Each error value is back-propagated to adjust the neural network transformation to reduce the error value. The steepness of each squashing function is controlled using the respective control variable to affect the response of each squashing function.

Abstract: A system that facilitates composition of schema mappings. A general algorithm is provided for composing a broad class of mappings, where one or both mappings are not functions, such as constraints between two schemas and the inverse of functions. A composition component performs composition on schema mappings of disparate data sources, which schema mappings can be expressed by at least one of full, embedded, and second-order dependencies, wherein the second-order dependencies need not be in source-to-target form. The algorithm for composition further includes a procedure that tests whether the composition algorithm will terminate.