Abstract: The invention relates to a method and means for performing a diagnosis of a medical condition and, in particular, an allergy associated with the upper respiratory tract, using an artificial neural network.

Abstract: A system that can identify, create, update and/or process a workflow based upon a current, past or future activity is disclosed. A ‘workflow’ can be defined as an activity flow that includes interaction with, or assignment of work to, people, devices, or services by a single individual or a group of individuals. Once a workflow is determined in accordance with the innovation, the system can inform other users or groups that are performing, or intend to perform, a similar or like activity. In establishing the workflow, the innovation can operate in an ad hoc or authored manner. As well, the system can employ a combination of either ad hoc or authored mechanisms in establishment of the workflow.

Abstract: Various embodiments for intelligent dump suppression in a computing environment following an error are provided. A plurality of historical information is considered in view of a current alert level to generate an output decision. The current alert level is one of an available plurality of alert levels configurable by a user. The current alert level is selectable by the user for a predetermined data collection restrictiveness. Data capture is performed according to the output decision.

Abstract: A system for designing a free form reflector includes a user input interface (1), a free form reflector design unit (2), and a free form reflector output unit (3). The user input interface is configured for receiving various data associated with a desired free form reflector, via an input device. The free form reflector design unit is installed in a computer and configured for generating an optimum free form surface according to the input data by performing a non-uniform rational basis splines (NURBS) algorithm, a merit evaluation function, and a differential evolution (DE) algorithm. The free form reflector output module is configured for generating a free form reflector according to the optimum free form surface and outputting the free form reflector, in the form of a computer-aided design (CAD) drawing, to a display and/or a printer. A related method is also disclosed.

Abstract: A STM network 11 for temporarily storing input pattern vectors is formed in Phases 1 and 2, and then layered LTM networks 2 to L are formed successively by assigning output vectors provided by the STM network 11 as input vectors. In phase 4, a LTM network 1 for intuitive outputs to which input pattern vectors are applied directly is formed by taking the parameters of comparatively highly activated centroids among centroids in the LTM networks 2 to L. In phase 5, the parameters of the comparatively highly activated centroids among the centroids in the LTM networks 2 to L are fed back as the parameters of the centroids in the STM network. In phase 3, the LTM networks 2 to L are reconstructed at a particular time or in a fixed period by giving the centroid vectors of the LTM networks 2 to L again as input pattern vectors to the STM network 11.

Abstract: A method and apparatus that detects a multiplicity of normal data sets, each of which includes values of n parameters, for each of the operation modes of an object having a plurality of operation modes. Self-organizing maps are provided for each operation mode using the normal data sets. Abnormal data sets representing virtual abnormal states are created by modifying the values of the n parameters of each of the multiple normal data sets so that as many abnormal data sets as the number of deviation vectors are created for each of the multiple normal data sets. Abnormal operation mode proportion vectors are then created by selecting a self-organizing map from the above noted self-organizing maps which has the highest similarity degree to each of the abnormal data sets.

Abstract: A method and system for grouping spam email messages are described. In one embodiment, the method includes receiving probe email messages indicative of spam and modifying the probe email messages to reduce noise. The method further includes comparing the probe email messages using fuzzy logic to identify similar email messages, and creating groups of similar email messages. Each of the created groups pertains to a distinct spam attack.

Abstract: Memory networks and methods are provided. Machine intelligence is achieved by a plurality of linked processor units in which child modules receive input data. The input data are processed to identify patterns and/or sequences. Data regarding the observed patterns and/or sequences are passed to a parent module which may receive as inputs data from one or more child modules. the parent module examines its input data for patterns and/or sequences and then provides feedback to the child module or modules regarding the parent-level patterns that correlate with the child-level patterns. These systems and methods are extensible to large networks of interconnected processor modules.

Abstract: A method and a system for discovering knowledge from text documents are disclosed, which involve extracting from text documents semi-structured meta-data, wherein the semi-structured meta-data includes a plurality of entities and a plurality of relations between the entities; identifying from the semi-structured meta-data a plurality of key entities and a corresponding plurality of key relations; deriving from a domain knowledge base a plurality of attributes relating to each of the plurality of entities relating to one of the plurality of key entities for forming a plurality of pairs of key entity and a plurality of attributes related thereto; formulating a plurality of patterns, each of the plurality of patterns relating to one of the plurality of pairs of key entity and a plurality of attributes related thereto; analyzing the plurality of patterns using an associative discoverer; and interpreting the output of the associative discoverer for discovering knowledge.

Abstract: The present invention provides systems and methods for dynamic detection and prevention of electronic fraud and network intrusion using an integrated set of intelligent technologies. The intelligent technologies include neural networks, multi-agents, data mining, case-based reasoning, rule-based reasoning, fuzzy logic, constraint programming, and genetic algorithms. The systems and methods of the present invention involve a fraud detection and prevention model that successfully detects and prevents electronic fraud and network intrusion in real-time. The model is not sensitive to known or unknown different types of fraud or network intrusion attacks, and can be used to detect and prevent fraud and network intrusion across multiple networks and industries.

Abstract: The present invention relates to the system, method and program for the pharmacokinetic parameter prediction of peptide sequence by the mathematical model. The present invention is comprising the steps of acquiring a variety of peptide sequence having specific features by the experimental technique; acquiring, on the basis of the sequence, a variety of peptide sequences lacking the specific features; storing the acquired peptide sequences as each set respectively, followed by randomly extracting peptide sequences in the constant ratio to divide into a training set and a test set of mathematical model; allowing individual peptide sequence descriptor values and an activity value; training the set of training peptide by mathematical model; predicting pharmacokinetic parameter of the set of test peptide by the trained mathematical model; and validating the trained mathematical model.

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: Systems, methods and apparatus are provided through which in some embodiments an autonomic entity manages a system by generating one or more stay alive signals based on the functioning status and operating state of the system. In some embodiments, an evolvable synthetic neural system is operably coupled to one or more evolvable synthetic neural systems in a hierarchy. The evolvable neural interface receives and generates heartbeat monitor signals and pulse monitor signals that are used to generate a stay alive signal that is used to manage the operations of the synthetic neural system. In another embodiment an asynchronous Alice signal (Autonomic license) requiring valid credentials of an anonymous autonomous agent is initiated. An unsatisfactory Alice exchange may lead to self-destruction of the anonymous autonomous agent for self-protection.

Abstract: Systems and methods for determining semantically related terms using an active learning framework such as Transductive Experimental Design are disclosed. Generally, to enhance a keyword suggestion tool, an active learning module trains a model to predict whether a term is relevant to a user. The model is then used to present the user with terms that have been determined to be relevant based on the model so that an online advertisement service provider may more efficiently provide a user with terms that are semantically related to a seed set.

Abstract: A probabilistic boosting tree framework for computing two-class and multi-class discriminative models is disclosed. In the learning stage, the probabilistic boosting tree (PBT) automatically constructs a tree in which each node combines a number of weak classifiers (e.g., evidence, knowledge) into a strong classifier or conditional posterior probability. The PBT approaches the target posterior distribution by data augmentation (e.g., tree expansion) through a divide-and-conquer strategy. In the testing stage, the conditional probability is computed at each tree node based on the learned classifier which guides the probability propagation in its sub-trees. The top node of the tree therefore outputs the overall posterior probability by integrating the probabilities gathered from its sub-trees. In the training stage, a tree is recursively constructed in which each tree node is a strong classifier. The input training set is divided into two new sets, left and right ones, according to the learned classifier.

Abstract: In a distributed computing system, an artificial intelligence system may be employed to configure the network variables. A metric describing the overall system performance may be derived during network operation or simulation and compared to an ideal metric describing the same distributed system performance. The difference between the derived metric and the ideal metric may then be used with an artificial intelligence system to modify the network variables to evolve the system toward the ideal performance standard.

Abstract: An objective function is dynamically chosen from a pool of available objective functions, and a prediction model is dynamically chosen from a pool of available prediction models. Points of inflection are determined for the chosen objective function, based on past values of a metric, to obtain a set of equations that can be solved to obtain model parameters associated with the chosen prediction model. The equations are solved to obtain the model parameters, and a future value of the metric is predicted based on (i) at least some of the past values of the metric and (ii) the chosen prediction model, with the obtained associated model parameters.

Abstract: The present invention relates to an emotion controlled system for processing multimedia data comprising a multimedia system for presenting multimedia content to a user an emotion model means for determining the emotional state of the user during the presentation of the multimedia content and an editing unit for changing said multimedia content in accordance with the emotional state of the user in order to present the changed multimedia content by the multimedia system. The present invention further relates to a method for executing the steps on this system.

Abstract: According to the present invention, there is provided a method to minimize human intervention during decision making process while controlling the electrical power system by identifying the initiating element that cause a tripping of the transmission overhead lines and capable of identifying the potential future protection system failures that can initiate a cascading of tripping or total national blackout. A method of producing flashover analysis signal as a protection system analysis comprising processing a neutral current, three phase current profile, three phase voltage profile, and a plurality of digital signal of a transmission line using an artificial neural network to calculate pickup time, reset time, DEF confirmation time or total fault clearance time. A method of producing flashover analysis signal comprising as a flashover signature analysis to identify the cause of the flashover as a current transformer explosion, tree encroachment, crane, lightning strike or polluted insulator.

Abstract: In a hierarchical neural network having a module structure, learning necessary for detection of a new feature class is executed by a processing module which has not finished learning yet and includes a plurality of neurons which should learn an unlearned feature class and have an undetermined receptor field structure by presenting a predetermined pattern to a data input layer. Thus, a feature class necessary for subject recognition can be learned automatically and efficiently.

Abstract: A method for projecting wafer product overlay error of the present invention is disclosed, the steps of the method comprises:(a) sample equipment overlay error data, equipment condition data, and actual wafer product overlay error data; (b) establish a neural network, the equipment overlay error data and the equipment condition data are inputs of the neural network, the generated output of the neural network is projected wafer product overlay error data, and the actual wafer product overlay error data is the target output of the neural network; and (c) set a mean square error target, train the neural network continuously until the mean square error of the neural network is no longer bigger than the mean square error target. Additionally a method for projecting wafer product critical dimension is also presented in the present invention.

Abstract: Associative memories include associative memory cells. A respective cell includes a sensor input, a prior association representation, a next association representation and an associative output. The cells are serially interconnected to form a linear array, such that the sensor inputs, the prior association representations and the next association representations of the serially connected cells are arranged in a sequence from distal to proximal cells based on affinities of associations among the series of sensor inputs. A respective cell also includes processing logic. The processing logic is responsive to the associated sensor input being active, to send a measure of the next association representation to an adjacent proximal cell and/or to send a measure of prior association representation to an adjacent distal cell.

Abstract: Generation of a meta-model for scatterometry analysis of a sample diffracting structure having unknown parameters. A training set comprising both a spectral signal evaluation and a derivative of the signal with respect to at least one parameter across a parameter space is rigorously computed. A neural network is trained with the training set to provide reference spectral information for a comparison to sample spectral information recorded from the sample diffracting structure. A neural network may be trained with derivative information using an algebraic method wherein a network bias vector is centered over both a primary sampling matrix and an auxiliary sampling matrix. The result of the algebraic method may be used for initializing neural network coefficients for training by optimization of the neural network weights, minimizing a difference between the actual signal and the modeled signal based on a objective function containing both function evaluations and derivatives.

Abstract: Missing data is addressed in a medical decision support system. The classifier applied to the patient record with missing data is obtained as a function of the available data. For example, one of a plurality of different classifiers is selected based on the features available in the patient record to be classified. The different classifiers are developed using different feature sets. The classifier developed using a feature set closest to or a sub-set of the features available in the patient record is selected for classifying the patient record. As another example, features in a training set corresponding to features available in the patient record are used to build a classifier. The classifier is applied to the patient record by inputting the available features of the patient record.

Abstract: A learning device includes: a plurality of learning modules, each of which performs update learning to update a plurality of model parameters of a pattern learning model that learns a pattern using input data; model parameter sharing means for causing two or more learning modules from among the plurality of learning modules to share the model parameters; and sharing strength updating means for updating sharing strengths between the learning modules so as to minimize learning errors when the plurality of model parameters are updated by the update learning.

Abstract: In a method for managing a data center, data for a training set of data is collected. The data includes heat loads generated by a plurality of electronic components and temperatures at multiple locations in the data center and under varying power profiles. A machine learning application is implemented to develop a model of a thermal topology of the data center using the training set of data and the model is implemented to predict heat profiles corresponding to a plurality of power profiles that are outside of the training set of data.

Abstract: In a procedure for regulating a combustion process in an installation while air is being supplied, material is converted by the combustion process, with at least one flame being formed, and the state variables (s(t)) describing the state of the system in the installation are determined using at least one observation device that images the flame, as well as other sensors, and are evaluated in a computer, whereupon any appropriate actions (ai) that may be needed are selected in order to control adjusting devices for the supply of material and/or air, wherein during setpoint regulation to achieve setpoints (s0) of the state variables and/or stability of the combustion process a changeover is occasionally made from setpoint control to disturbance control.

Abstract: Architecture that employs machine-based learning to automatically categorize data on a per-user basis. Auto-tagging reduces the burden on infoworkers by creating a machine learning model to learn from user tagging behavior or preferences. Once this information is obtained, a trained model for this specific user is used to assign tags to incoming data, such as emails. The architecture finds particular applicability to compliance and message retention policies that otherwise would mandate extra work for the infoworker. The architecture learns the tagging behavior of a user and uses this learned behavior to automatically tag data based on the user's prior tagging habits. A regression algorithm is employed to process the training data according to an n-dimensional framework for prediction and application of the tag(s) to the incoming messages.

Abstract: A system for detecting banking frauds in historical data and future transactions from a user supplied specimen set of fraudulent transactions, said specimen set of transactions defining one type of fraud identified by the user, said system comprises: means (301) to accept at least one set of banking transactions from the user and means to accept a type of fraud associated with each said set of transactions from the user (FIG. 3, Step 1); means (302) to run a set of atomic clue detectors on each said transaction for each said specimen (FIG. 3, Step 2); means (303) to store the output of said clue detectors for each said transactions for each said specimen fraudulent transactions (FIG. 3, Step 3); means (303) to compare the output of each said clue detector with a pre-defined threshold (FIG. 3, Step 3); means (304, 305) to assign weight to each said clue detector (FIG. 3, Step 4 and 5); means (306) to combine the clue detectors and their said weights into one fraud scenario (FIG.

Abstract: A Q Framework, or QFX for short, is provided for performing efficient tree transformation in a generalized manner that achieves preservation of action semantics for FSTs that support action information in their representations across a diverse set of types of representations for FSTs. Among other features, the QFX also enables the preservation of ordered and unordered nest information while performing tree transformation, supports the transformation of non-deterministic data structures to a deterministic data structure and enables intersection operations on machines having action semantics.

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 hierarchy of computing modules is configured to learn a cause of input data sensed over space and time, and is further configured to determine a cause of novel sensed input data dependent on the learned cause. Further, the hierarchy has a first level of computing modules and a second level of at least one computing module, where at least one of the computing modules in the first level is configured to receive a portion of the novel sensed input data, and where the computing module in the first level is further capable of determining a possible cause of the novel sensed input data dependent on analyzing only a subset of the portion of the novel sensed input data.

Abstract: A configuration that minimizes cost subject to the constraints is selected. A Simple Genetic Algorithm (SGA) is modified to incorporate the handling of constraints. The SGA is further modified to incorporate Optimal-Storage-Profiling to produce an increased number of fit individuals in each generation, developing a policy which will consider configurations of all storage-consumption levels and favor those configurations whose storage-consumption indicate they more likely to be strong candidates. An ideal-distribution of configurations, based on their storage-consumption, for each generation is developed. Different elitist policies are also incorporated to achieve greater scalability without sacrificing the quality of the solution.

Abstract: A hierarchy of computing modules is configured to learn a cause of input data sensed over space and time, and is further configured to determine a cause of novel sensed input data dependent on the learned cause. At least one of the computing modules has a sequence learner module configured to associate sequences of input data received by the computing module to a set of causes previously learned in the hierarchy.

Abstract: A method, apparatus and computer-program product for performing a system analysis of a system is disclosed.

Abstract: We develop a system consisting of a neural architecture resulting in classifying regions corresponding to users' keystroke patterns. We extend the adaptation properties to classification phase resulting in learning of changes over time. Classification results on login attempts of 43 users (216 valid, 657 impersonation samples) show considerable improvements over existing methods.

Abstract: Techniques are provided to determine user-interest features and user-interest parameter weights for a user-interest model. The user-interest features are pre-determined and/or determined dynamically. Pre-determined user-interest features are based on user-interest profiles, prior user activities, documents listed in a resume, reading or browsing patterns and the like. Dynamically determined user-interest features include features learned from an archive of user activities using statistical analysis, machine learning and the like. User-interest parameter weights are pre-determined and/or dynamically determined. Pre-determined user-interest parameter weights include parameter weights manually entered by a user indicating the relevant importance of a user-interest feature and parameter weights previously learned from an archive of the user's past activities.

Abstract: A hierarchy of computing modules is configured to learn a cause of input data sensed over space and time, and is further configured to determine a cause of novel sensed input data dependent on the learned cause. The hierarchy is further configured to associate a first pattern in the input data and a second pattern in the input data to a same possible cause of the input data.

Abstract: Methods of creating and using robust neural network ensembles are disclosed. Some embodiments take the form of computer-based methods that comprise receiving a set of available inputs; receiving training data; training at least one neural network for each of at least two different subsets of the set of available inputs; and providing at least two trained neural networks having different subsets of the available inputs as components of a neural network ensemble configured to transform the available inputs into at least one output. The neural network ensemble may be applied as a log synthesis method that comprises: receiving a set of downhole logs; applying a first subset of downhole logs to a first neural network to obtain an estimated log; applying a second, different subset of the downhole logs to a second neural network to obtain an estimated log; and combining the estimated logs to obtain a synthetic log.

Abstract: An integrated light and fragrance system automatically controls (150) aromatic effects (125) based on a user's control (111) of lighting effects (115). The system accepts one or more sets of lighting-fragrance correlations (165) from which to determine a preferred aromatic effect (161, 162) from a selected lighting effect (151, 152). Optionally, these sets of lighting-fragrance correlations (165) are provided by third-party vendors (310) who employ the skills of expert ambiance designers. In an alternative embodiment, the system may also use these sets of lighting-fragrance correlations (165) to control lighting effects (115) based on a user's control (121) of the fragrance effects (125). In each of these embodiments, the user merely controls (111, 121) a desired first effect (115, 125), and a suitable second effect (125, 115) is created to enhance the first effect (115, 125).

Abstract: Techniques are presented for assigning reputations to email senders. In one implementation, real-time statistics and heuristics are constructed, stored, analyzed, and used to formulate a sender reputation level for use in evaluating and controlling a given sender's connection to an message transfer agent or email recipient. A sender with an unfavorable reputation may be denied a connection before resources are spent receiving and processing email messages from the sender. A sender with a favorable reputation may be rewarded by having safeguards removed from the connection, which also saves system resources. The statistics and heuristics may include real-time analysis of traffic patterns and delivery characteristics used by an email sender, analysis of content, and historical or time-sliced views of all of the above.

Abstract: A diversified, self-organizing map (SOM) system and method creates a number of special-purpose SOMs by filtering and training from a SOM Database which contains user preference data entries that include a wide range of fields or attributes of user preferences. Each special-purpose SOM is trained with a filtered subset of user preference data for fields and attributes related to its special purpose. Two or more special-purpose SOMs are harnessed inter-cooperatively together to provide recommendations of preferred items in response to queries. Multiple SOMs can be maintained at different websites and harnessed together through a global SOM interface. The system can function more efficiently than a single large SOM using a monolithic database with single-type data entries of large dimensionality.

Abstract: A system and method of computer data analysis using neural networks. In one embodiment of the invention, the system and method includes generating a data representation using a data set, the data set including a plurality of attributes, wherein generating the data representation includes: modifying the data set using a training algorithm, wherein the training algorithm includes growing the data set; and performing convergence testing, wherein convergence testing checks for convergence of the training algorithm, and wherein the modifying of the data set is repeated until convergence of the training algorithm occurs; and displaying one or more subsets of the data set using the data representation. In one embodiment, the data representation is a knowledge filter that includes a representation of an input data set. The representation may be constructed during a training process. In one exemplary embodiment, the training process uses unsupervised neural networks to create the data representation.

Abstract: A method, system, API, GUI, and computer readable media for managing object-based clusters is provided. The method provides a computer executable methodology for discovering, monitoring, and managing object-based clusters. The system provides a computer-based system for facilitating interactions with heterogeneous cluster solutions. The system includes computer components for detecting clusters and supervising detected clusters and/or components.

Abstract: Pattern classification systems and methods are disclosed. The pattern classification systems and methods employ one or more classification networks that can parse multiple patterns simultaneously while providing a continuous feedback about its progress. Pre-synaptic inhibition is employed to inhibit feedback connections to permit more flexible processing. Various additional improvements result in highly robust pattern recognition systems and methods that are suitable for use in research, development, and production.

Abstract: Enhancements to Bayesian prediction models for network location traffic provide increased accuracy in web traffic predictions. The enhancements include implementing user advertising target queries to determine preferred edges of a Bayesian model, employing hierarchical data structures to cleanse training data for a Bayesian model, and/or augmenting existing data with new training data to enhance a previously constructed Bayesian model. Preferred edge enhancements for the Bayesian model utilize target attribute derived preferred edges and/or explicitly specified preferred edges. Training data is cleansed utilizing tag hierarchies that can employ parent-child relationships, ancestor relationships, and/or network location specific parameters. New training data can also be employed to adjust probabilities in a previously constructed Bayesian model. The new training data can be weighted differently than data represented by the previously constructed Bayesian model.

Abstract: A method of combinatorial multimodal optimization uses a genetic algorithm to find simultaneous global optimal solutions to combinatorial problems. Each individual within the population is associated not only with a fitness value but with a fitness vector, using which the persistence of all of the best individuals into the next generation can be guaranteed. Phenotype as well as genotype analysis is an integral part of the method.

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: This invention is a method of training a mean-field Bayesian data reduction algorithm (BDRA) based classifier which includes using an initial training for determining the best number of levels. The Mean-Field BDRA is then retrained for each point in a target data set and training errors are calculated for each training operation. Cluster candidates are identified as those with multiple points having a common training error. Utilizing these cluster candidates and previously identified clusters as the identified target data, the clusters can be confirmed by comparing a newly calculated training error with the previously calculated common training error for the cluster. The method can be repeated until all cluster candidates are identified and tested.

Abstract: Logging systems and methods are disclosed to reduce usage of radioisotopic sources. Some embodiments comprise collecting at least one output log of a training well bore from measurements with a radioisotopic source; collecting at least one input log of the training well bore from measurements by a non-radioisotopic logging tool; training a neural network to predict the output log from the at least one input log; collecting at least one input log of a development well bore from measurements by the non-radioisotopic logging tool; and processing the at least one input log of the development well bore to synthesize at least one output log of the development well bore. The output logs may include formation density and neutron porosity logs.