Abstract: Embodiments of the invention provide a method, system, and computer readable medium for classifying network traffic based on application signatures generated during a training phase using a modified subspace clustering scheme based on feature vectors extracted from network flows in a training set generated by a particular application and applying the signatures to a new feature vector extracted in real-time from current network data. The newly extracted feature vector is projected into the subspaces and compared with the signatures.

Abstract: A method may use a genetic algorithm to varying prediction parameters in forecasting software to obtain optimal predictions is disclosed. The method identifies parameters that can be varied and by modifying the parameters, the predictions of the forecasting software improve. The method uses sample data to train and validate the forecast and the optimal forecasting parameters are determined.

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 for operating a plurality of processor to determine a sequence of states traversed by a finite state machine in processing an input stream that includes a plurality of characters. The method divides the input stream into a plurality of contiguous sections characterized by a beginning character and an end character. Each section is assigned to a different processor. Each processor determines an end state that the finite state machine would traverse after the end character in the section assigned to that processor has been processed. That end state is provided to a processor that did not process the section associated with that end state. Each processor determines a sequence of states that the finite state machine would traverse if the finite state machine processed the section assigned to that processor using one of said received end states.

Abstract: Some implementations provide techniques for selecting web pages for inclusion in an index. For example, some implementations apply regularization to select a subset of the crawled web pages for indexing based on link relationships between the crawled web pages, features extracted from the crawled web pages, and user behavior information determined for at least some of the crawled web pages. Further, in some implementations, the user behavior information may be used to sort a training set of crawled web pages into a plurality of labeled groups. The labeled groups may be represented in a directed graph that indicates relative priorities for being selected for indexing.

Abstract: Provided is an information processing apparatus. The information processing apparatus includes a predictor construction unit. The predictor construction unit creates, by machine learning based on a genetic algorithm and by combining processing functions prepared in advance, feature quantity extraction formulae for extracting, from an observation value observed before a predetermined time, feature quantities of the observation value, and creates a prediction formula for predicting an observation value at the predetermined time based on the feature quantities calculated by the feature quantity extraction formulae. The information processing apparatus further includes a prediction unit. The prediction unit predicts an observation value at a time t from an observation value observed before the time t, by using the prediction formula created by the predictor construction unit.

Abstract: A method, system and computer software product are provided for process automation. The method includes monitoring one or more workstations to obtain multiple sets of character and graphical information including user actions, times and durations associated with the user actions, user-entered data, and retrieved character and graphical screen data. The method further includes identifying functional events from the sets, defining focal states as respective historic patterns of the functional events, and generating a set of facilitating scripts associated with respective ones of the focal states. The facilitating scripts each provide one or more automatic actions in the workstations. At least a portion of the set of facilitating scripts are then executed.

Abstract: A method for pattern recognition performed by a physical computing system includes, with the physical computing system, structuring data as a bi-partite graph, a set of instance nodes within the graph representing instances within the data and a set of property nodes within the graph representing properties of the instances, edges between the instance nodes and the property nodes representing values of the properties, assigning a transition probability function to each of the instance nodes and to each of the property nodes, and applying a random walker to the graph, the random walker utilizing the transition probability functions.

Abstract: Systems and methods (e.g., utilities) for use in providing automated, lightweight collection of online, open source data which may be content-based to reduce website source bias. In one aspect, a utility is disclosed for use in extracting content of interest from at least one website or other online data source (e.g., where the extracted content can be used in a subsequent search query). In other aspects, utilities are disclosed that are operable to perform various types of analyses on such extracted content and present graphical representations of such analyses on a display of a client device.

Abstract: The invention relates to a method, device and system for fusion of information originating from several sensors. The invention includes a mechanism for fusion of belief functions. To apply this mechanism, various information, knowledge and operations are modelled within the framework of the theory of belief functions: information provided by the sensors, knowledge regarding the propensity of the sensors to be in a given operating state, and merge operators for each operating state considered.

Abstract: Disclosed are an apparatus and method of deducing a user's intention using multimodal information. The user's intention deduction apparatus includes a first predictor to predict a part of a user's intention using at least one piece of motion information, and a second predictor to predict the user's intention using the predicted part of the user's intention and multimodal information received from at least one multimodal sensor.

Abstract: A method for solving an optimization problem includes: causing a cylindrical algebraic decomposition (CAD) processing unit to carry out a projection processing for a first expression that appears in a quantifier elimination problem equivalent to the optimization problem including objective functions and to generate second expressions that are projection factors of the first expression; calculating plural sets of values of the objective functions by generating plural value sets of variables in the objective functions and substituting the generated plural value sets of the variables into the objective functions; extracting points including non-dominated solutions in a space mapped by the objective functions, from plural points corresponding to the plural sets of values; calculating, for each second expression, an evaluation value concerning a distance between a corresponding second expression and each of the extracted points; and identifying a second expression whose evaluation value is minimum.

Abstract: A neuromorphic circuit performs functions representative of spiking timing dependent plasticity of a synapse.

Abstract: Certain aspects of the present disclosure present a technique for unsupervised training of input synapses of primary visual cortex (V1) simple cells and other neural circuits. The proposed unsupervised training method utilizes simple neuron models for both Retinal Ganglion Cell (RGC) and V1 layers. The model simply adds the weighted inputs of each cell, wherein the inputs can have positive or negative values. The resulting weighted sums of inputs represent activations that can also be positive or negative. In an aspect of the present disclosure, the weights of each V1 cell can be adjusted depending on a sign of corresponding RGC output and a sign of activation of that V1 cell in the direction of increasing the absolute value of the activation. The RGC-to-V1 weights can be positive and negative for modeling ON and OFF RGCs, respectively.

Abstract: A neural network has an array of interconnected processors, at least a first processor in the array operating in a pulse domain and at least a second processor in the array operating in a spike domain, and each said processor having: first inputs selectively coupled to other processors in the array of interconnected processors, each first input having an associated VCCS (a 1 bit DAC) coupled to a summing node, second inputs selectively coupled to inputs of the neural network, the second inputs having current generators associated therewith coupled to said summing node, a filter/integrator for generating an analog signal corresponding to current arriving at the summing node, and for processors operating in the pulse domain, an analog-to-pulse converter for converting an analog signal derived either directly from the filter/integrator or via a non-linear element, to the pulse domain, and providing the converted analog signal as an unquantized pulse domain signal at an output of each processor operating in the p

Abstract: A method of operation of a navigation system includes: extracting navigation-related web documents having a point of interest; generating formatting sequences of the navigation-related web documents; selecting a user-defined percentile representing reciprocal fraction of an expected number of clusters; calculating a threshold value for a first cluster with the threshold value to be equal to the user-defined percentile of a first normalized distribution of sample comparison values between the first cluster and formatting sequence samples from the formatting sequences, the first cluster is from the clusters; computing an associated comparison value between a first formatting sequence from the formatting sequences and the first cluster; grouping the first formatting sequence with the first cluster when the associated comparison value exceeds the threshold value for the first cluster; and generating a travel route for the point of interest related to the first cluster for displaying on a device.

Abstract: Some embodiments of a method to lazily enable truth maintenance in a rule engine have been presented. A rule engine can enable truth maintenance on an object type basis. In response to a working memory action involving a particular object type, the rule engine may check a truth maintenance parameter of the object type involved to determine if truth maintenance for the object type has to be enabled. If so, the rule engine may enable truth maintenance for the object type.

Abstract: A classifier training system trains unified classifiers for categorizing videos representing different categories of a category graph. The unified classifiers unify the outputs of a number of separate initial classifiers trained from disparate subsets of a training set of media items. The training process takes into account the relationships that exist between the various categories of the category graph by relating scores associated with related categories, thus enhancing the accuracy of the unified classifiers.

Abstract: Disclosed are an emotion model, and an apparatus and method for adaptively learning personality of the emotion model. The emotion model, which maintains personality information, creates emotion information according to the personality information and takes a predetermined behavior according to the emotion information. The personality information may change adaptively in correspondence to a user's response to the behavior performed by the emotion model. Accordingly, the emotion model may react adaptively to the user through interactions with the user.

Abstract: A system, method, and computer-readable instructions for real-time characters with learning capabilities. A plurality of rules are defined in a rules-based system, each of the rules defining a condition that determines a behavior of a virtual agent when the rule is triggered by the condition being satisfied so that upon triggering of multiple rules at the same time, each of the behaviors of the multiple rules whose conditions were satisfied are combined into a resultant behavior for the virtual agent. This resultant behavior is compared with a desired behavior to providing feedback in the form of rewards or punishments to each of the multiple rules based on their corresponding contribution to the resultant behavior as compared to the desired behavior.