Abstract: There is disclosed a method of monitoring an infrastructure comprising managed units, the method comprising the steps of: acquiring data associated with a first performance indicator from a first managed unit; determining a first quantized distribution function of at least a subset of pieces of data of the acquired data of the first managed unit; determining if the first quantized distribution function verifies one or a plurality of predefined rules describing particular distribution functions of performance indicators.

Abstract: A method for providing cognitive insights via a cognitive information processing system comprising: receiving data from a plurality of data sources; receiving and processing queries; and, bridging the queries into a cognitive graph.

Abstract: Systems and methods are provided relating to predictive analysis, and more specifically to predictive analysis of one or more data types in a data center environment. The data center environment itself includes a power manager in communication with at least one cabinet power distribution unit (CDU) that is in power-supplying communication with at least one electronic appliance in an electronic equipment rack. The predictive analysis approach estimates the rate of change over a future interval of time for at least one data type based on said historical data and predicts when said at least one data type will reach an associated user-defined threshold based on said rate of change. Results can be displayed graphically on an application program associated with the power manager.

Abstract: An exemplary embodiment of the present invention provides a method of processing network activity data. The method includes receiving network activity data and generating an event based on the network activity data. The method also includes generating a probability based at least in part on Bayesian statistics, the probability corresponding to a likelihood that the event caused or was caused by another event. The method also includes generating an event message corresponding to the event based on the probability.

Abstract: Storage devices and components, including memory components (e.g., non-volatile memory) can be trained by executable code that facilitates and/or performs reads and/or write requests to one or more storage sub-modules of a storage component (e.g., memory configured on a memory channel) made up of multiple storage components (e.g., DIMMs). The executable code can also train multiple storage components at the same time and/or in parallel.

Abstract: According to methods, apparatus and AI servers for determining an AI behavior, a protocol-requesting command sent by an application logic server can be received. The protocol-requesting command can contain an application identifier, a notification message, and current environment data. From a plurality of preset AI systems, an AI system corresponding to the application identifier can be found. The AI system can be formed by a preset plurality of components that include one or more classifying components. From the AI system, a classifying component corresponding to the notification message can be found. The classifying component can be mounted with at least one behavior component. From the classifying component corresponding to the notification message, a behavior component matching the current environment data can be obtained. An AI behavior can be determined based on the obtained behavior component matching the current environment data.

Abstract: A method, system, and computer program product for event detection using roles and relationships of entities are provided in the illustrative embodiments. A training event and a set of entities participating in the training event are identified in a training data. For a first entity in the set of entities, a first role occupied by the entity in the event is determined. A behavior attribute is assigned to the first role. A relationship of the first role with a second role corresponding to a second entity in the set of entities is determined. An event rule is constructed to detect an event corresponding to the training event in new data and comprising a plurality of roles, behavior attributes, and the relationship. The plurality of roles includes the first role and the second role, and the plurality of behavior attributes includes the behavior attribute assigned to the first role.

Abstract: Techniques and systems are described that enable improved gesture discernment from input devices, as well as simplified modeling and processing of gestures by application software layers. Given data (e.g., about movements, actions, or events) gathered from input devices, techniques and systems allow gestures to be discerned and inferred more formally and reliably, and processed more easily by an application layer. A gesture interpreter is provided that, in response to receiving an activation input data from an input device, instantiates a high-level Petri Net instance, executes the high-level Petri Net instance, and returns, to an application layer, an outcome gesture indicative of a terminal node in a path of the high-level Petri Net instance being traversed during the execution of the high-level Petri Net instance.

Abstract: Embodiments of the invention involve receiving a first set of data describing one or more first observations and a second set of data describing one or more second observations. The first set of data comprises at least two types of data and the second set of data comprises at least two types of data. At least one of the two types of data in the first data set are common with at least one of the two types of data in the second data set. The common types of data comprise common data to the first and second sets of data. The types of data that are not common comprise exclusive data for each of the first and second sets of data. A first multiple regression model is developed for the first data set. The common data for the first data set are set as independent variables and the exclusive data for the first data set are set as dependent variables. A second multiple regression model is developed for the second data set.

Abstract: The present invention relates to the identification of a person having risk for developing type 2 diabetes (T2D) by determining the presence or absence of specific genes, gene clusters, genera or species of bacteria in the person's gastrointestinal microbiota. More specifically the invention relates to a model to identify an individual having or at risk of developing type 2 diabetes (T2D) using metagenomic clusters (MGCs), wherein said model is characterized by using different metagenomic clusters for different population groups. Also provided is the use of such a model in the identification of a person having risk for developing type 2 diabetes (T2D).

Abstract: A system for developing and implementing empirically derived algorithms to generate decision rules to predict invalidity of subject reported data and fraud with research protocols in surveys allows for the identification of complex patterns of variables that detect or predict subject invalidity of subject reported data and fraud with the research protocol in the survey. The present invention may also be used to monitor invalidity of subject reported data within a research protocol to determine preferred actions to be performed. Optionally, the invention may provide a spectrum of invalidity, from minor invalidity needing only corrective feedback, to significant invalidity requiring subject removal from the survey. The algorithms and decision rules can also be domain-specific, such as detecting invalidity or fraud among subjects in a workplace satisfaction survey, or demographically specific, such as taking into account gender or age.

Abstract: A unified attractiveness prediction method is provided. The method includes receiving a plurality of videos and extracting at least one of metadata and view data from each of the plurality of received videos, wherein the metadata is information for describing video contents, and the view data is a total number of users who watch the video. The method also includes obtaining potential view amounts of the plurality of received videos and calculating impact factor scores of the plurality of received videos based on the potential view amounts, each impact factor score including a numerical score for indicating a degree of effectiveness of a corresponding video. Further, the method includes providing a video with a highest impact factor score based on the calculated impact factor scores.

Abstract: The subject disclosure is directed towards the use of Monte Carlo (MC) procedures for computing the value of information (VOI), including with long evidential sequences. An MC-VOI algorithm is used to output a decision as to balancing the value and costs of collecting information in advance of taking action by running prediction model-based simulations to determine execution paths through possible states, and processing the results of the simulations/paths taken into a final decision.

Abstract: Disclosed are methods and systems for classifying one or more human subjects in one or more categories indicative of a health condition of the one or more human subjects. The method includes categorizing one or more parameters of each of the one or more human subjects in one or more data views based on a data type of each of the one or more parameters. A data view corresponds to a first data structure storing a set of parameters categorized in the data view, associated with each of the one or more human subjects. The one or more data views are transformed to a second data structure representative of the set of parameters across the one or more data views. Thereafter, a classifier is trained based on the second data structure, wherein the classifier classifies the one or more human subjects in the one or more categories.

Abstract: A method predicting a network activity associated with a given network site is provided. The method can include receiving a request to predict a probability of network activity associated with the network site, analyzing historical data associated with the network site, and, based on the analysis, determining the probability of the network activity in future. The method can further include monitoring the network site, ascertaining evidence associated with the network activity, and, based on the evidence, adjusting treatment of the network site. Additionally, the method can include comparing the probability to a predetermined threshold probability and, based on the comparison, selectively taking an action concerning the network site.

Abstract: A system, method and computer program product for hierarchical sparse dictionary learning (“HiSDL”) to construct a learned dictionary regularized by an a priori over-complete dictionary, includes providing at least one a priori over-complete dictionary for regularization, performing sparse coding of the at least one a priori over-complete dictionary to provide a sparse coded dictionary, using a processor, updating the sparse coded dictionary with regularization using at least one auxiliary variable to provide a learned dictionary, determining whether the learned dictionary converges to an input data set, and outputting the learned dictionary regularized by the at least one a priori over-complete dictionary when the learned dictionary converges to the input data set. The system and method includes, when the learned dictionary lacks convergence, repeating the steps of performing sparse coding, updating the sparse coded dictionary, and determining whether the learned dictionary converges to the input data set.

Abstract: Features are disclosed for identifying randomized latent feature language modeling, such as a recurrent neural network language modeling (RNNLM). Sequences of item identifiers may be provided as the language for training the language model where the item identifiers are the words of the language. To avoid localization bias, the sequences may be randomized prior to or during the training process to provide more accurate prediction models.

Abstract: Apparatus and methods facilitate analysis of events associated with network and computer systems. Event data, such as security threats, are comparison matched with event rules of event rule sets associated with each network or computer system to determine whether the items are potentially significant. Additionally, the system-event data may be scored where the score is used for prioritizing system-event data as to their significance. Associated with the comparison matching are various analytics that further analyze event data for measuring and analyzing the system-event data according to various algorithms.

Abstract: The invention discloses an identification method of nonlinear parameter varying models (NPV) and belongs to the industrial identification field. The invention carries out identification tests and model identification for an identified object with nonlinear parameter varying characteristics. Firstly, the multi-input single-output nonlinear parameter varying model is identified through the steps of local nonlinear model tests, local nonlinear models identification, and operating point variable transition tests; after completing the identification of all the multi-input single-output nonlinear parameter varying models with respect to all the controlled variables, the completed multi-input multi-output nonlinear parameter varying models are built. The nonlinear parameter varying models of an identified object can be obtained by the identification method of the present invention with limited input/output data without detailed mechanism knowledge of the identified object.

Abstract: Exemplary practice of this invention implements a computer to model human decision-making within a comprehensive human-perception construct dichotomized as probabilistically perceptual and preferentially perceptual. Potential actions are identified. Each identified potential action is evaluated in consideration of (i) probabilities of success as perceived by the decision-maker, and (ii) preferences of the decision-maker relating to consistency, credibility, confidence, bias, and urgency. Decision-making is modeled on a continual basis whereby evaluation of at least one potential action is performed anew in each successive time-step. Evaluations of potential actions yield “goodness” values, which are compared to determine best potential actions. Threshold “goodness” parameters are established to filter out some potential actions and leave other potential actions for goodness comparison.