Abstract: Apparatuses, systems, and techniques to identify out-of-distribution input data in one or more neural networks. In at least one embodiment, a technique includes training one or more neural networks to infer a plurality of characteristics about input information based, at least in part, on the one or more neural networks being independently trained to infer each of the plurality of characteristics about the input information.

Abstract: A computer implemented method of detecting anomalous behavior within a computer network, the method including accessing data records each corresponding to an occurrence of communication occurring via the computer network and including a plurality of attributes of the communication; generating, for each of at least a subset of the data records, a training data item for a neural network, the training data item being derived from at least a portion of the attributes of the record and the neural network having input units and output units corresponding to items in a corpus of attribute values for communications occurring via the network; augmenting the training data by replicating each of one or more training data items responsive to one or more attributes of the data record corresponding to the training data item; training the neural network using the augmented training data so as to define a vector representation for each attribute value in the corpus based on weights in the neural network for an input unit cor

Abstract: A method for managing specialized hardware resources includes obtaining, by a disaster recovery (DR) virtual resource agent, a request for a DR environment for a set of virtual resources in a primary site, in response to the request: monitoring the primary site to obtain virtual workload information corresponding to the set of virtual resources, performing a workload analysis on the set of virtual resources in the primary site using the virtual workload information to obtain a virtual resource mapping of each virtual resource in the primary site to a tiered component in the DR environment, and initiating a DR environment allocation of DR virtual resources based on the virtual resource mapping.

Abstract: A dual solution candidate searcher receives an input of information about a constraint coefficient matrix and a cost vector, determines a dual problem of a linear programming problem being a primal problem and all active sets representing combinations of active formulas in constraints of the dual problem, finds, for each of the active sets, a feasible dual solution candidate meeting constraints, and stores the dual solution candidate into a storage in a manner associated with a corresponding one of the active sets. An optimal solution calculation device receives an input of a constraint vector as, selects an optimal one of the active sets as an optimal active set based on an inner product of the constraint vector and the dual solution candidate stored in the storage, and finds and outputs a basic feasible solution corresponding to the selected active set as an optimal solution.

Abstract: Media content may be recommended based on feedback weightings. Input data describing the presentation of a media content items in association with content management accounts may be represented as data points. Each data point may identify feedback data for a media content item previously recommended for presentation in association with a content management account. The feedback data may identify a viewer reaction to the recommended media content item. A weighting factor based on the feedback data for the media content item presented in association with the content management account may be applied to produce a recommendation.

Abstract: A time-series data processing apparatus according to the present invention includes: a generating unit configured to generate a generator having learned so as to generate state information representing a state of time-series data having a predetermined time width in accordance with a label given to the time-series data; a state information generating unit configured to generate, by using the generator, state information representing a state of division time-series data obtained by dividing the time-series data by a shorter time width than the predetermined time width; and a classifying unit configured to classify a plurality of division time-series data based on state information of the plurality of division time-series data.

Abstract: A fluorescence microscopy method includes a trained deep neural network. At least one 2D fluorescence microscopy image of a sample is input to the trained deep neural network, wherein the input image(s) is appended with a digital propagation matrix (DPM) that represents, pixel-by-pixel, an axial distance of a user-defined or automatically generated surface within the sample from a plane of the input image. The trained deep neural network outputs fluorescence output image(s) of the sample that is digitally propagated or refocused to the user-defined surface or automatically generated. The method and system cross-connects different imaging modalities, permitting 3D propagation of wide-field fluorescence image(s) to match confocal microscopy images at different sample planes. The method may be used to output a time sequence of images (e.g., time-lapse video) of a 2D or 3D surface within a sample.

Abstract: According to some aspects, disclosed methods and systems may include having a user input one or more speech commands into an input device of a user device. The user device may communicate with one or more components or devices at a local office or headend. The local office or the user device may transcribe the speech commands into language transcriptions. The local office or the user device may determine a mood for the user based on whether any of the speech commands may have been repeated. The local office or the user device may determine, based on the mood of the user, which content asset or content service to make available to the user device.

Abstract: Implementations disclose determining connections between sound recordings and compositions. A method includes maintaining, by a server device of a content sharing platform, a serving structure to represent resources comprising sound recording resources and composition resources and explicit connections between the resources, generating a hypothesis structure from the serving structure, the hypothesis structure comprising the resources and the explicit connections of the serving structure and inferred connections between the resources, the inferred connections identified based on data comprised in the resources of the serving structure, and upon receiving a user approval of at least one of the inferred connections, adding, by the server device, the at least one of the inferred connections to the serving structure.

Abstract: Systems are provided for facilitating the building and use of models used to make data preparation recommendations. The systems identify ground truth from a plurality of notebooks and utilizes the ground truth to generate the corresponding data preparation recommendation models. The data preparation recommendation models are used to predict accurate (e.g., useful and relevant) data preparations steps based on user input and user notebook data. The data preparation computing system generates a recommendation prompt based on output from the data preparation recommendation model that can be viewed and/or selected by the user to be applied to the user's notebook data.

Abstract: A method for processing a static pattern in an image and a circuit system are provided. In the method, each of frames of a video is divided into multiple areas. Several algorithms are used to calculate a static pattern index of every area. The static pattern index is used as a reference for determining if the area covers part or entire of a static pattern. An index threshold can be used to check the areas that are determined as the static pattern initially. A time threshold is then used to confirm the areas with the static pattern in every frame. Image data of the areas which are determined as the static patterns can then be adjusted, such as having brightness of the area that is determined as part or entire of the static patterns decreased, for preventing the display panel from negative effects of the static pattern.

Abstract: Embodiments are directed to managing data correlation over a network. Role success models that correspond to roles and to success criteria may be provided. A student profile that includes skill vectors may be provided based on student information. Role success models may be employed to determine intermediate scores based on the skill vectors and the success criteria. A predictive score for the student that corresponds with a predicted performance of the student in the roles may be generated based on the one or more intermediate scores. Actions for the student may be determined based on a mismatch of the skill vectors and role skill vectors that correspond to the roles. In response to the student performing the actions: updating the one or more skill vectors based on a completion of the actions; and updating the predictive score based on the role success models and the updated skill vectors.

Abstract: Software, firmware, and systems are described herein that migrate functionality of a source physical computing device to a destination physical computing device. A non-production copy of data associated with a source physical computing device is created. A configuration of the source physical computing device is determined. A configuration for a destination physical computing device is determined based at least in part on the configuration of the source physical computing device. The destination physical computing device is provided access to data and metadata associated with the source physical computing device using the non-production copy of data associated with the source physical computing device.

Abstract: Methods, apparatus, systems, computing devices, computing entities, and/or the like for a hybrid question-answering (QA) application are provided. The hybrid QA application includes both a retrieval QA application and a deep QA application, and uses feedback from the retrieval QA application to improve the deep QA application.

Abstract: Provided is a brain-machine interface based intention determination device using a virtual environment, including: an input unit that receives a control mode; a training output unit that outputs training information; a collection unit that collects a brain signal; a first preprocessing unit that extracts time-frequency information; a second preprocessing unit that generates physical information; a learning unit that learns control information according to a pattern of physical information; a determination unit that determines control information; and a command output unit that outputs a control command matching the control information.

Abstract: According to some embodiments, a system, method, and non-transitory computer readable medium are provided comprising a plurality of real-time monitoring nodes to receive streams of monitoring node signal values over time that represent a current operation of the cyber physical system; and a threat detection computer platform, coupled to the plurality of real-time monitoring nodes, to: receive the monitoring node signal values; compute an anomaly score; compare the anomaly score with an adaptive threshold; and detect that one of a particular monitoring node and a system is outside a decision boundary based on the comparison, and classify that particular monitoring node or system as anomalous. Numerous other aspects are provided.

Abstract: A method for personalizing media content for a user is provided. The method includes, at an electronic device, streaming a first media item from a first set of media items, the first set of media items compiled using a first recommendation hypothesis. The method further includes, while streaming the first media item, in response to a first user request, selecting, without user intervention, a second set of media items, distinct from the first set of media items, including determining a presentation order of a plurality of sets of media items using a heuristic applied to the plurality of sets of media items. The second set of media items is compiled using a second recommendation hypothesis, wherein the second recommendation hypothesis is distinct from the first recommendation hypothesis. The method includes streaming a second media item from the second set of media items.

Abstract: Techniques and apparatus presented herein are directed toward monitoring an electric power delivery system to detect and locate a power generation event. A power generation event may include a tripped generator, a loss of a transmission line, or other loss of power generation. To detect the event, an analysis engine may receive and monitor input data. A detection signal may be generated based on the input data. Upon detecting the event, the analysis engine may determine a source and propagation of the event through the delivery system. Based on the source and propagation of the event, the analysis engine may determine the location of the event. The analysis engine may generate an overlay with the input data to provide the location and other information about the event to a user such that remedial action can be taken to resolve the event and restore the lost power generation.

Abstract: Methods and systems are presented for generating a machine learning model using enhanced gradient boosting techniques. The machine learning model is configured to receive inputs corresponding to a set of features and to produce an output based on the inputs. The machine learning model includes multiple layers, wherein each layer includes multiple models. To generate the machine learning model, multiple models are built and trained in parallel for each layer of the machine learning model. The multiple models use different subsets of features to produce corresponding output values. After a layer in built and trained, a collective error may be determined for the layer based on the output values from the different models in the layer. An additional layer of models may be added to the machine learning model to reduce the collective error of a previous layer.

Abstract: A big data analysis system may include a big data repository communicatively coupled to a data accumulation server and a predictive graph processing system. The data accumulation server may be configured to receive information from a plurality of data sources, the information corresponding to user interaction with one or more computing devices associated with an organization via a networked computing system, store the information received from the plurality of sources in the big data repository; and monitor the plurality of data sources to update the data stored in the big data repository. The predictive graph processing system is configured to receive information stored in the big data repository, transform the information received from the big data repository into a predictive graph data set based on a predictive model, and store the predictive graph data set to a visualization data repository.

Abstract: In various embodiments, a communication fusion application enables other software application(s) to interpret spoken user input. In operation, a communication fusion application determines that a prediction is relevant to a text input derived from a spoken input received from a user. Subsequently, the communication fusion application generates a predicted context based on the prediction. The communication fusion application then transmits the predicted context and the text input to the other software application(s). The other software application(s) perform additional action(s) based on the text input and the predicted context. Advantageously, by providing additional, relevant information to the software application(s), the communication fusion application increases the level of understanding during interactions with the user and the overall user experience is improved.

Abstract: A hypotheses generation and event recognition system that enables event recognition by analyzing documents to construct one or more qualitative metrics (e.g., frequency of keywords, changes in sentiment, occurrence of ontological terms, evolution of topics, etc.), establishing a baseline for the qualitative metric(s), and outputting changes to that baseline for display. In aggregate, those qualitative metrics comprise temporal and/or spatial signals that, when combined, define signatures of events of interest. Accordingly, the user and/or the system may identify an event of interest based on the change in baseline. The system may further provide functionality to generate hypotheses by coding data according to an ontology, populating an ontology space, and using an optimization algorithm to rank points or neighborhoods in the coded ontology space.

Abstract: A method, apparatus, and device for erasing a Solid State Disk (SSD), and a storage medium are provided. The method includes: receiving, from a Virtual Flash Translation Layer (VFTL) running in a preset virtual environment, an erasing request for erasing a target block; selecting a target spare block from preset spare blocks and feeding back the target spare block to the VFTL; collecting running information of the VFTL within a preset period of time, and determining whether a function of the VFTL is normal according to the running information; and in a case of determining that the function of the VFTL is normal, performing an erasing operation on the target block. The method may fully and comprehensively detect the abnormal condition of the VFTL, and may guarantee the integrity of data information in the solid state disk, thereby improving the reliability of the solid state disk.

Abstract: Embodiments are disclosed for performing fact correction of natural language sentences using data tables. In particular, in one or more embodiments, the disclosed systems and methods comprise receiving an input sentence, tokenizing elements of the input sentence, and identifying, by a first machine learning model, a data table associated with the input sentence. The systems and methods further comprise a second machine learning model identifying a tokenized element of the input sentence that renders the input sentence false based on the data table and masking the tokenized element of the tokenized input sentence that renders the input sentence false. The systems and method further includes a third machine learning model predicting a new value for the masked tokenized element based on the input sentence with the masked tokenized element and the identified data table and providing an output including a modified input sentence with the new value.

Abstract: Provided is a storage system including a plurality of controllers. The storage system adopts a write-once data storage system and can implement high Input/Output (I/O) processing performance while ensuring data consistency when a failure occurs. Before metadata duplication, recovery data including information necessary for performing roll forward or roll back is stored in each controller, and then the metadata duplication is performed. A recovery data storage processing and the metadata duplication are offloaded to a hardware accelerator.

Abstract: A system for generating an input signal for a video game includes: a receiving unit configured to receive a video game output signal output by a video game playing device during the playing of a video game; a reasoning model configured to determine an action to take within the video game based on the received video game output signal, the reasoning model being configured to generate a representation of the determined action that is independent of the video game; and an encoding unit configured to translate the representation into an input signal and to provide the input signal to a video game playing device at which the video game is being played, thereby causing the determined action to be performed within the video game.

Abstract: The present disclosure relates to a system for detecting an anomaly and a method for setting a threshold thereof. The method includes setting a minimum true positive rate (TPR) and a differentiation slope based on a user input, creating a receiver operating characteristic (ROC) curve using anomaly data and normal data, estimating a true negative rate (TNR) from the ROC curve based on the minimum TPR and the differentiation slope, and estimating and setting the threshold based on the estimated TNR.

Abstract: Methods, systems, and non-transitory machine-readable mediums for ranking on an absolute scale include displaying, on an electronic display, a first handle, a second handle, and an interactor, determining a value of the first handle and a value of the second handle based on their respective positions on the interactor, in response to a user dragging the first and second handles on the interactor, and determining a rank of the first and second handles based on the values of the first and second handles, in response to the user dragging the first and second handles on the interactor.

Abstract: A predicting device includes: a recognizer configured to recognize a state of a moving body; a predictor configured to predict a position of the moving body in the future on the basis of the recognition result of the recognizer; and a processor configured to determine whether to set a first risk region or a second risk region larger than the first risk region for the moving body on the deviation between a position of the moving body which has been recognized by the recognizer and a position of the moving body in the future which has been predicted previously by the predictor.

Abstract: A method for extracting a kansei adjective of a product based on principal component analysis and explanation (PCA-E) includes constructing a product kansei evaluation vector matrix through original kansei adjectives; performing dimensionality reduction through PCA; and determining, based on principal component load factors, kansei adjectives representing principal components. In this way, the kansei adjectives extracted are explanatory to help users understand the selected kansei adjectives and make accurate evaluation.

Abstract: Embodiments of a wearable cardioverter defibrillator (WCD) system include a support structure for wearing by an ambulatory patient and at least one processor. When worn, the support structure maintains electrodes on the patient's body, and using the patient's ECG received via the electrodes, the processor determines widths of the QRS complexes, consistency of the QRS complexes, and/or heart rate and uses these determinations to make no-shock, delay-shock, and shock decisions. Shock decisions can be made for heart rates lower than a VF threshold.

Abstract: Embodiments may provide a document system that receives a responsiveness call from a user through the task/queue framework regarding a machine call document. Theses responsiveness calls may be used to refining the scoring algorithm used by the document system of to generate a desired confidence score for the document system.

Abstract: Methods, systems, and computer programs are presented for reconciling a transaction against data in a database to identify the transaction parameters based on text descriptors provided for the transaction. One method includes an operation for identifying features for reconciling transactions of a first entity by a machine-learning program. The features include, at least, a description of the transaction, a name of a second entity in the transaction, a location of the second entity, and an account for the transaction. The machine-learning program is trained with training data that includes values of the features for previously reconciled transactions. A received first transaction includes a description, a date, and an amount. The first transaction is input for the machine-learning program, which generates one or more suggestions for reconciling the first transaction. Each suggestion includes the name of the second entity in the first transaction and an account.

Abstract: Examples herein include systems and methods for providing capacity forecasting for high-usage periods of a computing infrastructure. An example method can include segmenting a first portion of a data stream and generating a first core set for a forecasting model that predicts future usage of computing resources. The example method can further include segmenting a second portion of the data stream, generating a second core set, and using both core sets to forecast usage. The first core set can then be phased out after a predetermined time period has elapsed such that forecasting is based only on the second core set. The example method can further include defining at least two clusters of data and performing predictive analysis on that specific cluster. Cluster-specific results can be displayed on a GUI, which can also provide a user with options for increase or decrease computing resources based on the predictions.

Abstract: Embodiments of the disclosure provide methods, systems, computer program products for manipulating a table with an XML column in database. According to the method, a SQL statement for a table is received first, wherein the table comprises an XML column which is a representation of a plurality of logical columns of the table, there is at least one XML element in at least one row and in the XML column of the table, and each of the at least one XML element corresponds to a non-null value in a specific row and in a logical column of the plurality of logical columns of the table. Then the SQL statement is parsed. And then the SQL statement is transformed into a hybrid statement being able to process the XML column in response to the parsing result indicating that the SQL statement relates to at least one logical column of the plurality of logical columns. At last the hybrid statement is executed.

Abstract: A system and method for creating mobile device or vehicle profiles (mobile object) using location, time, event, and unrelated data.

Abstract: A computer system develops models and generates decision logic based on the developed models. The decision logic is distributed to end user devices, and the end user devices are able to implement the decision logic to detect events, determine event sequences, and correlate the determined event sequences to predicted outcomes.

Abstract: Disclosed herein are system, method, and computer program product embodiments for recommending a service provider. An embodiment operates by receiving a search request specifying a service request by a user and location for the requested service, retrieving prior financial transactions of the user to purchase the service from a database, and determining a preference of the user based on the retrieved prior financial transactions. The embodiment further operates by applying the preference to a model trained to output a recommended service provider and selecting the recommended service provider for display on a user's device. The model more likely recommends service providers in the location having a higher ratio of tip-to-price relative to other providers for the service in the location based on a history of prior financial transactions of a plurality of users.

Abstract: A method of computing a join pattern agnostic aggregate function includes loading source code and parsing the source code to identify different database query operations. In response to the identification of an aggregate function in one of the different database query operations in which records in multiple tables are joined together in a table join, a multiplicity of primary keys are retrieved, each corresponding to a different one of the multiple tables and also an object of one of the tables referenced by the identified aggregate function. An existence of a fan out condition associated with the table join is then computed. On condition that an existence of the fan out condition is not computed a non-fan out sensitive implementation of the aggregate function is invoked with respect to the object. But, otherwise, a different, fan-out sensitive implementation of the aggregate function is invoked with respect to the object.

Abstract: A method of identifying causal relationships includes receiving data comprising a set of values corresponding to one or more variables, and generating a list of candidate causal models of relationships between or within the variables. The list is ranked based on a likelihood of each candidate causal model, wherein the likelihood includes at least a correlation value. The method further includes receiving feedback identifying a candidate causal model and a change in rank of the candidate causal model, re-ranking the list based on the feedback, and displaying the re-ranked list. The method generates an intervention comprising a suggested modification corresponding to a variable of a selected causal model among the candidate causal models in the re-ranked list, receives additional data corresponding to the variable of the suggested modification and evaluates the additional data to determine whether the likelihood of the selected causal model has changed.

Abstract: Improved mechanisms for collecting information from a diverse suite of sensors and systems, calculating the current precipitation, atmospheric water vapor, atmospheric liquid water content, or precipitable water and other atmospheric-based phenomena, for example presence and intensity of fog, based upon these sensor readings, predicting future precipitation and atmospheric-based phenomena, and estimating effects of the atmospheric-based phenomena on visibility, for example by calculating runway visible range (RVR) estimates and forecasts based on the atmospheric-based phenomena.

Abstract: A system coordinates services between users and providers. The system trains a computer model to predict a user state of a user using data about past services. The prediction is based on data associated with a request submitted by a user. Request data can include current data about the user's behavior and information about the service that is independent of the particular user behavior or characteristics. The user behavior may be compared against the user's prior behavior to determine differences in the user behavior for this request and normal behavior of prior requests. The system can alter the parameters of a service based on the prediction about the state of the user requesting the service.

Abstract: A system is provided for streaming media content in a vehicle. The system includes a personal media streaming appliance system configured to connect to a media delivery system and receive media content from the media delivery system at least via a cellular network. The personal media streaming appliance system operates to transmit a media signal representative to the received media content to a vehicle media playback system so that the vehicle media playback system operates to play the media content in the vehicle. Customized voice communications are generated based on receiving input, such as a user query and/or a media track change indication.

Abstract: In general, embodiments described herein relate to methods and systems for automating the configuration of network devices. More specifically, embodiments of the invention relate to using configuration commands that specify protocol-specified relationships in order to generate granular (or specific) filtering rules (also referred to as rules). The rules are subsequently programmed into the network device.

Abstract: In one aspect, the present disclosure is directed to a method for improving the speed of determining a health risk profile associated with a patient. The method may include the step of retrieving patient medical information about the patient, wherein the patient medical information is an uncoded natural language expression in a first language. The method may also include comparing the patient medical information with records in a first database. If the patient medical information matches the preselected medical information, the method includes performing a first data conversion procedure. If the patient medical information fails to match any record in the first database, the method includes performing a second data conversion procedure, wherein the first data conversion procedure is performed faster than the second data conversion procedure.

Abstract: Techniques for detecting and correcting anomalies in computer-based reasoning systems are provided herein. The techniques can include obtaining current context data and determining a contextually-determined action based on the obtained context data and a reasoning model. The reasoning model may have been determined based on one or more sets of training data. The techniques may cause performance of the contextually-determined action and, potentially, receiving an indication that performing the contextually-determined action in the current context resulted in an anomaly. The techniques include determining a portion of the reasoning model that caused the determination of the contextually-determined action based on the obtained context data and causing removal of the portion of the model that caused the determination of the contextually-determined action, to produce a corrected reasoning model.

Abstract: Example implementations described herein are directed to a novel Automated Machine Learning (AutoML) framework that is generated on an AutoML library so as to facilitate functionality to incorporate multiple machine learning model libraries within the same framework through a solution configuration file. The example implementations further involve a solution generator that identifies solution candidates and parameters for machine learning models to be applied to a dataset specified by the solution configuration file.

Abstract: Systems and methods relating to multitask transfer learning. Neural networks are used to accomplish a number of tasks and the results of these tasks are used to determine parameters common to these and other tasks. These parameters can then be used to accomplish other related tasks. In the description, data fitting as well as image related tasks are used. Task conditioning as well as the use of a KL regularizer have greatly improved results when testing the methods of the invention.

Abstract: Methods, apparatus, systems and articles of manufacture to detect phishing websites are disclosed. An example apparatus includes a plurality of website analyzers to analyze a requested website for evidence of a phishing attack, the plurality of website analyzers including a first website analyzer and a second website analyzer. An analysis selector is to select the first website analyzer for execution, the analysis selector to, in response to determining that an additional analyzer is to be executed, select the second website analyzer to analyze the requested website. A website classifier is to, in response to a website analyzer indicating a classification that exceeds a confidence threshold, classify the requested website as a benign site or presenting a phishing attack.

Abstract: Technologies are provided for dynamic delivery of travel planning information based at least on a travel condition. The travel planning information can be crowdsourced from multiple supplier accounts and can be retained in a distributed storage platform. Some embodiments of the technologies include a computing system that can receive data identifying travel conditions of a subject from a mobile user equipment (UE). The travel conditions can include a health condition and/or biometric data. The mobile UE can be a portable computing device or a vehicle. The computing system can determine that a rule to provide travel planning information is satisfied, and then can acquire specific travel planning information using at least one of the travel conditions. The computing system can provide the first travel planning information to the mobile UE.