Abstract: A system described herein includes a trainer component that receives an estimated gradient of cost that corresponds to a first ranker component with respect to at least one training point and at least one query. The trainer component builds a second ranker component based at least in part upon the received estimated gradient. The system further includes a combiner component that linearly combines the first ranker component and the second ranker component.

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: A behavior control apparatus collects information of a mobile object in relation to an action space of the mobile object, and acquires a position and an orientation of a human. The behavior control apparatus sets an exclusive area for the human based on the position and the orientation, and judges whether any information is to be notified to the human by the mobile object. If judging negatively, the behavior control apparatus determines a target position, a target orientation and a travel route of the mobile object such that the mobile object moves out of the exclusive area.

Abstract: The present invention provides a method for learning a policy used by a computing system to perform a task, such delivery of one or more objects by the computing system. During a first time interval, the computing system determines a first state, a first action and a first reward value. As the computing system determines different states, actions and reward values during subsequent time intervals, a state description identifying the current sate, the current action, the current reward and a predicted action is stored. Responsive to a variance of a stored state description falling below a threshold value, the stored state description is used to modify one or more weights in the policy associated with the first state.

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: Online fraud prevention including receiving a rules set to detect fraud, mapping the rules set to a data set, mapping success data to members of the rules set, filtering the members of the rules set, and ordering members of the data set by giving priority to those members of the data set with a greater probability for being fraudulent based upon the success data of each member of the rule set in detecting fraud. Further, a receiver coupled to an application server to receive a rules set to detect fraud, and a server coupled to the application server, to map the rules set to a data set, and to map the success data to each members of the rules set. The server is used to order the various members of the data set by giving priority to those members of the data set with a greatest probability for being fraudulent.

Abstract: A custodian profile, e.g., a user profile, associated with a first content item, e.g., an associated web page in a social network, can be identified as a first content item. The first content item can be utilized to identify one or more second content items, e.g., advertisements, that are displayed when the first content item is presented, e.g., viewed by another user of the social network.

Abstract: A system and method for protecting identity fraud are disclosed. A system includes a detection subsystem to identify applications and/or accounts at risk of identity fraud, and a disposition subsystem to process data provided by the detection system and to determine whether identity fraud exists in the applications and/or accounts. According to an implementation, one or more neural network models are defined, each neural network model being configured to handle a class of cases related to the subject and a specific data configuration describing a case of the class. The one or more neural network models are run to generate data requests about the subject's identity, and the data requests are passed to a detection system that monitor transactions associated with the subject. Additional data associated with the transactions is requested until a threshold certainty is achieved or until available data or models are exhausted.

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 patent 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 system for controlling applications of a wireless mobile device includes a server for receiving data related to an adaptive user profile and for controlling operations of applications within the wireless mobile device. An adaptive neural/fuzzy logic control application implemented within the network server generates the adaptive user profile responsive to the received data. The adaptive user profile controls operations of the applications within the wireless mobile device and changes in real time responsive to the received data.

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 method for improving stacking schema for classification tasks, according to which predictive models are built, based on stacked-generalization meta-classifiers. Classifications are combined to build a new scheme from at least two layers and multiclass classification problems are converted into binary classification problems. One-against-all class binarization and regression learners are used for each class model and ensemble classifiers are improved using stacking. Accuracy differences, accuracy ratio, and runtime classification in multiclass datasets are also improved and the class of a value is then predicted.

Abstract: A welding system is provided, in which two electrodes are directed at a joint between two work pieces and the electrodes are energized with DC pulse or AC welding waveforms at a controlled waveform phase angle. The systems include a synchronizing controller to synchronize the welding waveforms, and a work point allocation system provides one or more work point values to the welding machines to provide synergic control of the welding according to a user selected system work point value or parameter. The systems and methods further provide for synchronized work point value modulation for the opposite sides of a dual fillet weld. The system and method further provide a high energy heat source that directs intense heat at a portion of the weld joint to improve weld penetration.

Abstract: A tool facilitating learning latent semantics for ranking (LLSR) tailored to the ranking task via leveraging relevance information of query-document pairs to learn a tailored latent semantic space such that other documents are better ranked for the queries in the subspace. The tool applying a learning latent semantics for ranking algorithm integrating LLSR, thereby enabling learning an optimal latent semantic space (LSS) for ranking by utilizing relevance information in the training process of subspace learning. The tool enabling an optimization of the LSS as a closed form solution and facilitating reporting the learned LSS.

Abstract: Providing for automated generation of tags (e.g., metadata descriptors) for items of e-mail or syndication formatted communication is described herein. By way of example, a system can include a filtering component that can generate one or more tags based on information relevant to content of the communication, a sender, or recipient, or combinations thereof. In addition, such tags can be automatically attached to a received item, or a presentation component can furnish the tags to a recipient (e.g., by way of a communication device user interface) for selection, whereby selected tags are associated with the item of communication. Accordingly, the subject innovation provides for improved classification and description of items of communication by automatic generation of descriptive and/or representative tags associated therewith.

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: Methods, apparatuses and systems directed to pattern identification and pattern recognition. In some particular implementations, the invention provides a flexible pattern recognition platform including pattern recognition engines that can be dynamically adjusted to implement specific pattern recognition configurations for individual pattern recognition applications. In some implementations, the present invention also provides for a partition configuration where knowledge elements can be grouped and pattern recognition operations can be individually configured and arranged to allow for multi-level pattern recognition schemes.

Abstract: A cross descriptor learning system, method and program product therefor. The system extracts descriptors from unlabeled exemplars. For each unlabeled exemplar, a cross predictor uses each descriptor to generate labels for other descriptor. An automatic label generator also generates labels for the same unlabeled exemplars or, optionally, for labeled exemplars. A label predictor results for each descriptor by combining labels from the cross predictor with labels from the automatic label generator.

Abstract: Systems and methods for a scalable artificial neural network, wherein the architecture includes: an input layer; at least one hidden layer; an output layer; and a parallelization subsystem configured to provide a variable degree of parallelization to the artificial neural network by providing scalability to neurons and layers. In a particular case, the systems and methods may include a back-propagation subsystem that is configured to scalably adjust weights in the artificial neural network in accordance with the variable degree of parallelization. Systems and methods are also provided for selecting an appropriate degree of parallelization based on factors such as hardware resources and performance requirements.

Abstract: Various amounts of information can be beneficial to different controllers configured upon an industrial control system. Information can be retained in a distributed directory such that controllers quickly learn information concerning other controllers. The distributed directory can be automatically constructed and populated with information from different controllers. When a module enters an industrial control system, information can automatically advertise to other units through use of the distributed directory.

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: A data processing apparatus includes: predicting means for calculating a prediction value of time series data with respect to input of the time series data using a prediction model for predicting the time series data; determining means for determining a target value of the time series data on the basis of the prediction value of the time series data; error calculating means for calculating an error of the prediction value relative to the target value; and retrieving means for retrieving error reduction data as input of the time series data to the prediction model for reducing the error of the prediction value.

Abstract: Intrusions in a system under surveillance are detected by matching the events occurring during operation of the system against a knowledge base including information on events which occurred during a learning phase. The detection technique includes the steps of: recording, during the learning phase, temporal data related to the events during the learning phase; identifying, as a function of the temporal data recorded, a dynamic part of the knowledge base; discovering patterns that cover the dynamic part of the knowledge base; and using, during the analysis phase, a regular expression match at least with respect to the dynamic part of the knowledge base.

Abstract: In some embodiments, a streaming message classification method dynamically allocates a stream of messages to a variable number of clusters (e.g. message categories), each containing messages which share a set of similar features. Incoming messages are compared to a collection of known spam clusters. New spam types are identified, and new clusters are created automatically and dynamically in order to accommodate the new spam types. Message clustering is performed in a hyperspace of message feature vectors using a modified k-means algorithm. Triangle inequality distance comparisons may be used to accelerate hyperspace distance calculations.

Abstract: Machine-readable media, methods, apparatus and system for obtaining and processing image features are described. In some embodiments, groups of training features derived from regions of training images may be trained to obtain a plurality of classifiers, each classifier corresponding to each group of training features. The plurality of classifiers may be used to classify groups of validation features derived from regions of validation images to obtain a plurality of weights, wherein each weight corresponds to each region of the validation images and indicates how important the each region of the validation images is. Then, a weight may be discarded from the plurality of weights based upon a certain criterion.

Abstract: Process control system for detecting abnormal events in a process having one or more independent variables and one or more dependent variables. The system includes a device for measuring values of the one or more independent and dependent variables, a process controller having a predictive model for calculating predicted values of the one or more dependent variables from the measured values of the one or more independent variables, a calculator for calculating residual values for the one or more dependent variables from the difference between the predicted and measured values of the one or more dependent variables, and an analyzer for performing a principal component analysis on the residual values. The process controller is a multivariable predictive control means and the principal component analysis results in the output of one or more scores values, T2 values and Q values.

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: Coalitions from interactions and adaptations of cognitive map agents are evolved using an algorithm. A population of agents are seeded with cognitive map variants characterizing different cultures or different affiliations. The algorithm evolves this population by modifying the cognitive maps using a modified Particle Swarm Optimization algorithm. The modifications include modification to weights of the cognitive map, and the structure of the cognitive map of the global best (gbest) in the neighborhood is imitated according to a weighted random selection, based on the commonality of the node characteristic in the neighborhood. The end results indicate whether a coalition is possible and what cognitive maps emerge. These results are visualized on a 2D grid and measured with a clustering metric.

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: 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: Identification of a determinative subset of features from within a group of features is performed by training a support vector machine using training samples with class labels to determine a value of each feature, where features are removed based on their the value. One or more features having the smallest values are removed and an updated kernel matrix is generated using the remaining features. The process is repeated until a predetermined number of features remain which are capable of accurately separating the data into different classes.

Abstract: A multiphase flow meter used in conjunction with an electrical submersible pump system in a well bore includes sensors to determine and transmit well bore pressure measurements, including tubing and down hole pressure measurements. The multiphase flow meter also includes at least one artificial neural network device to be used for outputting flow characteristics of the well bore. The artificial neural network device is trained to output tubing and downhole flow characteristics responsive to multiphase-flow pressure gradient calculations and pump and reservoir models, combined with standard down-hole pressure, tubing surface pressure readings, and the frequency applied to the electrical submersible pump motor.

Abstract: In one aspect, input data for a predictive model characterizing a level of risk for a data transaction is received that includes values for categorical variables and one or more of binary variables and continuous variables the predictive model. Thereafter, one or more of the categorical variables is associated with one of a plurality of keys. Each key having corresponding coefficients for at least a subset of the binary variables and the continuous variables and the coefficients being dependent on a value for the key. A composite value based on values for each of at least a subset of the binary variables and the continuous variables as calculated using the corresponding coefficients for each key can then be generated. Scoring of the data transaction using the binary variables, the continuous variables, and the composite variables can then be initiated by the predictive model. Related apparatus, systems, techniques and articles are also described.

Abstract: A novel and useful mechanism enabling a standard learning algorithm to generate rules for complex event processing (CEP) systems. The method creates rules that infer previously defined output events by creating input event feature vectors for each targeted output event. In addition, a method for automatically generating CEP system rules to infer output events which are anomalies (i.e. statistical outliers) of input event sequences is disclosed. Input feature vectors consisting of multiple input events and parameters for each targeted output event are then input into a standard learning algorithm to generate CEP system rules.

Abstract: Various neural-network based surrogate model construction methods are disclosed herein, along with various applications of such models. Designed for use when only a sparse amount of data is available (a “sparse data condition”), some embodiments of the disclosed systems and methods: create a pool of neural networks trained on a first portion of a sparse data set; generate for each of various multi-objective functions a set of neural network ensembles that minimize the multi-objective function; select a local ensemble from each set of ensembles based on data not included in said first portion of said sparse data set; and combine a subset of the local ensembles to form a global ensemble. This approach enables usage of larger candidate pools, multi-stage validation, and a comprehensive performance measure that provides more robust predictions in the voids of parameter space.

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: Techniques for altering application user-interface controls are provided. More particularly the customization of a new or existing user-interface control in an application user-interface is provided. In one aspect of the invention, a method for customizing user-interface controls of an existing application comprises the recording of a procedure description performed by a user in the application user-interface. A new or modified application user-interface control relating to the procedure description is then installed in the existing application.

Abstract: Disclosed are methods, systems, and computer-readable media for determining stability of a neural system. The method includes tracking a function world line of an N element neural system within at least one behavioral space, determining whether the tracking function world line is approaching a psychological stability surface, and implementing a quantitative solution that corrects instability if the tracked function world line is approaching the psychological stability surface.

Abstract: A sound classification system for adding and correcting a sound type is disclosed. When the add/correct command processor receives a command to add or correct a sound type, the data in the first database is stored in the second database, and the type adding/correcting device adds the feature of the sound to the first database, and re-calculates the statistical values. Besides, the classifier re-classifies the sample sounds, and the precision calculator calculates a ratio of accurate classification. When the ratio is high, the type adding/correcting device stores, in the feature database, the feature of the sound for which a type is to be added or corrected. When the ratio is low, the second database restores the data back to the first database.

Abstract: In one aspect, the invention is based on a process that combines information present in a joint distribution of the predictor variables and the variable (or variables) to be predicted. This information may be captured in the form of a table or other like data structure that includes a set of vectors (referred to as a “TAB”). The process uses the information in the TAB in conjunction with one or more rules. In one embodiment, a set of different rules are applied to the TAB to determine which rule in the set produces the most accurate predictions. The RULE that produces the most accurate predictions is then used in conjunction with observed information to make predictions.

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 pattern matching technique for high throughput network processing includes a simple yet powerful special purpose architecture and a set of novel string matching algorithms that can work in unison. The novel set of algorithms allow for bit-level partitioning of rules such that may be more easily implemented in hardware or software. The result is a device that maintains tight worst case bounds on performance, can be updated with new rules without interrupting operation, compiles in seconds instead of hours, and is ten times more efficient than the existing best known solutions in this area.

Abstract: Methods, systems, and apparatuses for generating relevance functions for ranking documents obtained in searches are provided. One or more features to be used as predictor variables in the construction of a relevance function are determined. The relevance function is parameterized by one or more coefficients. An ideal query error is defined that measures, for a given query, a difference between a ranking generated by the relevance function and a ranking based on a training set. According to a structured output learning framework, values for the coefficients of the relevance function are determined to substantially minimize an objective function that depends on a continuous upper bound of the defined ideal query error. The query error is determined using a structured output learning technique. The query error is defined as a maximum over a set of permutations.

Abstract: A method for scheduling and usage of bandwidth wherein data is transmitted to an artificial intelligence model for analysis and assigned a ranking. An event resource allocation model analyzes the ranking and determines how to compress the associated video. The event resource allocation model also determines when to compress the remotely stored video and transmit to a central data center where the video may be reviewed.

Abstract: A system improves recognition results. The system receives multimedia data and recognizes the multimedia data based on training data to generate documents. The system receives user augmentation relating to one of the documents or new documents from a user. The system supplements the training data with the user augmentation or new documents and retrains based on the supplemented training data.

Abstract: A system and method for protecting identity fraud are disclosed. A system includes a detection subsystem to identify applications and/or accounts at risk of identity fraud, and a disposition subsystem to process data provided by the detection system and to determine whether identity fraud exists in the applications and/or accounts. According to an implementation, one or more neural network models are defined, each neural network model being configured to handle a class of cases related to the subject and a specific data configuration describing a case of the class. The one or more neural network models are run to generate data requests about the subject's identity, and the data requests are passed to a detection system that monitor transactions associated with the subject. Additional data associated with the transactions is requested until a threshold certainty is achieved or until available data or models are exhausted.

Abstract: A technique is provided to coarse-class one or more customer characteristics used in a predictive model. A set of functions are used to represent partition points of the customer characteristic into smaller classes. Each of the final classes of the customer characteristic is represented separately in the predictive model. An initial set of functions may be established to provide an initial set of partitions points of the customer characteristic. The set of functions is then processed using a genetic algorithm to evolve the partition points to new values. Processing the set of partitions using the genetic algorithm may continue until a stopping criterion is reached.

Abstract: A learning control apparatus for an autonomous agent including a functional module having a function of multiple inputs and multiple outputs, the function receiving at least one variable and outputting at least one value, includes an estimating unit for estimating a causal relationship of at least one variable, a grouping unit for grouping at least one variable into a variable group in accordance with the estimated causal relationship, a determining for determining a behavior variable corresponding to each of the variable groups, and a layering unit for layering, in accordance with the variable group and the behavior variable, the function corresponding to each variable group, the function receiving the variable grouped into the variable group and outputting the behavior variable.

Abstract: A system and method for support vector machine plus (SVM+) computations include selecting a set of indexes for a target function to create a quadratic function depending on a number of variables, and reducing the number of variables to two in the quadratic function using linear constraints. An extreme point is computed for the quadratic function in closed form. A two-dimensional set is defined where the indexes determine whether a data point is in the two-dimensional set or not. A determination is made of whether the extreme point belongs to the two-dimensional set. If the extreme point belongs to the two-dimensional set, the extreme point defines a maximum and defines a new set of parameters for a next iteration. Otherwise, the quadratic function is restricted on at least one boundary of the two-dimensional set to create a one-dimensional quadratic function. The steps are repeated until the maximum is determined.

Abstract: A group of features that has been identified as “significant” in being able to separate data into classes is evaluated using a support vector machine which separates the dataset into classes one feature at a time. After separation, an extremal margin value is assigned to each feature based on the distance between the lowest feature value in the first class and the highest feature value in the second class. Separately, extremal margin values are calculated for a normal distribution within a large number of randomly drawn example sets for the two classes to determine the number of examples within the normal distribution that would have a specified extremal margin value. Using p-values calculated for the normal distribution, a desired p-value is selected. The specified extremal margin value corresponding to the selected p-value is compared to the calculated extremal margin values for the group of features.