Abstract: Methods, systems, and apparatus for receiving a request that includes an entity identifier of an entity referenced by a search query and one or more context terms that are included in the search query; determining that the one or more context terms describe a relationship connecting the entity referenced by the search query with a plurality of other entities; and in response to determining that the one or more context terms describe the relationship, generating user interface elements that provide facts related to the entity referenced. Generating the user interface may include identifying relationship knowledge elements that include facts relating to at least some of the other entities connected to the entity by the relationship, and identifying one or more additional knowledge for the entity referenced by the search query. The ranking the relationship knowledge elements may be highest ranked knowledge elements in the user interface elements.

Abstract: At least one unique collection of storage artifacts can be specified to indicate that the storage artifact is a member of the unique collection. Each storage artifact can be a discrete object comprising digitally encoded content that is stored as a node within a tree structure of a tangible storage medium. The collection can be referenced by a set of different storage artifacts to form a collection of related storage artifacts. Each storage artifact can correspond to different collections, wherein membership within a collection is independent of a storage path within the tree structure. A file management action relating to the storage artifact can be performed. The file management action can be dependent upon the storage artifact being a member of the unique collection.

Abstract: Intelligent merchant onboarding is described. A service provider can determine first directive(s) that were presented to one or more first potential merchants that completed an onboarding process with the service provider and second directive(s) that were presented to one or more second potential merchants that did not complete the onboarding process. The service provider can train a data model based at least partly on the first directive(s) or the second directive(s), Based at least partly on receiving an indication that a potential merchant initiates the onboarding process, the service provider can determine information associated with the potential merchant. Based at least partly on the data model and the information, the service provider can intelligently determine an order that one or more directives are to be presented to the potential merchant such that a first directive of the first directive(s) is presented prior to a second directive of the second directive(s).

Abstract: In one embodiment, a method involves accessing training data, where the training data contains an ordered sequence of data associated with a plurality of entities, training one or more deep learning models to determine, from the ordered sequence of data, a first set of embeddings for each entity of the plurality of entities, where each entity has a plurality of entity attributes, determining, for each of the plurality of entity attributes, a corresponding initial embedding, training the one or more deep-learning models to refine the initial embeddings according to one or more criterion, generating one or more updated embeddings for each of the plurality of entities based on the refined initial embeddings of the plurality of entity attributes, and modifying the first set of embeddings based on the one or more updated embeddings.

Abstract: Apparatuses, methods, and systems are provided for making sequential recommendations using transition regularized non-negative matrix factorization. A non-application specific collaborative filtering based personalized recommender system can recommend a next logical item from a series of related items to a user. The recommender system can recommend a next desirable or series of next desirable new items to the user based on the historical sequence of all user-item preferences and a user's most recent interaction with an item. An asymmetric item-to-item transition matrix can capture aggregate sequential user-item interactions to design a loss function for matrix factorization that incorporates the transition information during decomposition into low-rank factor matrices.

Abstract: Novel techniques are described for generation, distribution, and management of route connectivity optimization (RCO) mapping. For example, as mobile devices traverse travel routes serviced by one or more mobile networks, they can experience periods of different levels of connectivity with the mobile network(s). Embodiments can collect route segment connectivity data as experienced by consumer devices during traversal through mobile networks (e.g., indicating, for each route segment, which carriers are providing service to a mobile device, the level of service being provided, etc.). A RCO can be computed and stored for the set of route segments as a function of the route segment connectivity data. The RCO can be requested by consumers, and a corresponding link can be generated. Selecting the link can provide the consumers with remote access to the RCO, with which the consumers can generate connectivity-optimized route guidance maps.

Abstract: The present disclosure provides a generalization processing method, apparatus, device and computer storage medium, and relates to technical field of artificial intelligence and specifically to a deep learning technique. A specific implementation solution is: determining a set of candidate queries in a query library that are similar to a requested query in at least one of a literal matching manner, a semantic matching manner and a query rewriting manner; determining a generalized query corresponding to the requested query from the set of candidate queries by using a pre-trained query matching model; wherein the query matching model is obtained by pre-training based on a cross attention model. The generalization for the requested query can be achieved according to the present disclosure.

Abstract: System(s) and method(s) that analyze image data associated with a computing screen operated by a user, and learns the image data (e.g., using pattern recognition, historical information analysis, user implicit and explicit training data, optical character recognition (OCR), video information, 360°/panoramic recordings, and so on) to concurrently glean information regarding multiple states of user interaction (e.g., analyzing data associated with multiple applications open on a desktop, mobile phone or tablet). A machine learning model is trained on analysis of graphical image data associated with screen display to determine or infer user intent. An input component receives image data regarding a screen display associated with user interaction with a computing device. An analysis component employs the model to determine or infer user intent based on the image data analysis; and an action component provisions services to the user as a function of the determined or inferred user intent.

Abstract: A method, system, and computer program product for contextual generation of an ephemeral network are provided. The method detects an initiating event for network generation associated with a user of a first computing device. The method determines a duration of the initiating event. A set of network members is determined based on the initiating event and the duration. The method establishes an ephemeral network, connecting at least a portion of computing devices associated with the set of network members. The ephemeral network is terminated in response to detecting a completion event. In response to terminating the ephemeral network, member information for the set of network members is removed from the portion of the computing devices.

Abstract: Described is a system for detecting adversarial activities based on detection of activity patterns in a multiplex network. The system detects one or more subnetworks that are matches to a template network of template nodes. The subnetworks are detected by filtering multiplex network nodes according to a filtering criteria that utilizes monotone function properties in the multiplex network. Nodes that do not meet the filtering criteria are eliminated, resulting in a list of candidate nodes in the multiplex network. The one or more subnetworks are formed from the list of candidate nodes. An activity pattern corresponding to a pattern of adversarial activity is identified in the one or more subnetworks. Based on the identified activity pattern, an alert of adversarial activity is generated and transmitted.

Abstract: A storage device includes a nonvolatile memory device; and a controller configured to, sequentially receive first read commands and a first write command, the first write command being associated with first write data, slice the first write command to generate a plurality of sub-commands, slice the first write data to generate a plurality of sub-data elements, and alternately transmit, to the nonvolatile memory device, at least one read command of the first read commands, and one sub-command of the plurality of sub-commands and one sub-data element of the plurality of sub-data elements.

Abstract: A caption of a multimodal message (e.g., social media post) can be identified as a named entity using an entity recognition system. The entity recognition system can use a visual attention based mechanism to generate a visual context representation from an image and caption. The system can use the visual context representation to identify one or more terms of the caption as a named entity.

Abstract: According to one aspect of the present invention, a learning device is configured to perform learning of decision trees by gradient boosting. The learning device includes an initializer and a learning unit. The initializer is configured to perform initialization processing on an address memory that stores an address in a data memory of learning data used for learning of a node in a decision tree. The learning unit is configured to perform learning of the decision tree by using the learning data stored in the data memory. The initializer is configured to perform, after a certain decision tree is learned by the learning unit, the initialization processing for a next decision tree of the certain decision tree in parallel with processing for the certain decision tree.

Abstract: Aspects of the present disclosure involve a system comprising a computer-readable storage medium storing a program and method for determining conditions for a set of webpages. The program and method provide for receiving a first set of URLs, each URL of the first set corresponding to a respective webpage of a web site, the first set corresponding to a subset of webpages that comprise the website; identifying repeating patterns of field values in the first set; determining positive conditions that characterize the first set; determining, from the web site webpages other than those in the first set, a second set of URLs that satisfy the positive conditions; identifying repeating patterns of field values in the second set; determining positive conditions that characterize the second set; and computing negative conditions for the first set, by subtracting the positive conditions characterizing the second set from the positive conditions characterizing the first set.

Abstract: In certain embodiments, reference identifiers may be generated and embedded into a website's content. A request for a reference identifier to be embedded into first content on a first website may be obtained based on a user accessing the first website. Based on the request, interaction data related to the first website may be retrieved indicating that a prior user interacted with text on the first website and subsequently accessed a second website. A reference identifier related to the second website may be caused to be embedded into the text on the first website based on: (i) the second website comprising second content related to the text, (ii) the prior user interacting with the text on the first website, and (iii) the prior user accessing the second website after interacting with the first content on the first website.

Abstract: A computer-implemented interaction control method includes determining, based on a first requirement, a first category from one or more categories estimated from each of a plurality of pieces of information input by a user, selecting, based on the determined first category, a first conversation topic for interaction with the user from conversation topics, executing, by using the first conversation topic, the interaction with the user via a user interface, determining, when detecting that the first conversation topic is inappropriate in accordance with a result of the interaction executed by using the first conversation topic, based on a second requirement, a second category from the one or more categories, selecting, based on the determined second category, a second conversation topic for the interaction with the user from the conversation topics, executing, by using the second conversation topic, the interaction with the user via the user interface.

Abstract: The present invention relates to an artificial intelligence (AI), which emulates functions of a human brain, such as recognition and reasoning, by utilizing a machine learning algorithm such as deep learning, and relates to context awareness based artificial intelligence application technology for obtaining interest information of a user from an image displayed to the user.

Abstract: A computer-implemented method for optimizing a search index of a search service may be provided. The method may comprise storing continuously search queries of the search service, receiving a content item to be indexed, adding the content item temporarily to the a main index of the search service, determining a set of document relevancy factor (DRF) values for the content item for a selected set of stored search queries, determining a combined document relevancy factor value out of the determined set of document relevancy factor values for the content item, comparing the combined document relevancy factor value for the content item to a document relevancy factor threshold value, and adding, in response to the comparing, the content item permanently to an index based on the results of the comparing.

Abstract: A method and apparatus for reserving a usable storage space on a storage device is provided. The method includes collecting context data representing an environment surrounding the storage device; selecting at least one file from among files stored in the storage device by using at least one of the context data and user profile data; and processing the selected file and reserving a usable storage space on the storage device. The method reserves the usable storage space by using the context data or user profile data, thereby allowing efficient reserving of usable storage space without a user's manual intervention and preventing waste of unnecessary resources.

Abstract: Embodiments of the systems and methods disclosed herein provide a prescriptive analytics platform, a polarity analysis engine, and a semantic analysis engine in which a user can identify a target objective and use the system to find out whether the user's objectives are being met, what predictive factors are positively or negatively affecting the targeted objectives, as well as what recommended changes the user can make to better meet the objectives. The systems and methods may include a polarity analysis engine configured to determine the polarity of terms in free-text input in view of the target objective and the predictive factors and use the polarity to generate the recommended changes. The systems and methods may also include a semantic analysis engine to extend the results of the polarity analysis engine for improved determination of predictive factors and improved recommendations.

Abstract: Techniques are presented herein for improved search based on group relevance. The techniques include determining account node groups based on the interactions of accounts with content. When a search query is received from a particular account, the group(s) associated with that account are determined, and the content items determined to be relevant to other accounts in the group(s) are given higher relevance in the search results. In some embodiments, groups are determined using partitioning and/or queries can be rewritten based on the group(s).

Abstract: Technology is described herein for ranking candidate result items in at least two stages. In a first stage, the technology uses a first attention-based neural network to determine an extent of attention that each token of an input query should pay to the tokens of each candidate result item. In a second stage, the technology uses a ranking subsystem to perform listwise inference on output results provided by the first stage, to generate a plurality of ranking scores that establish an order of relevance of the candidate results items. The ranking subsystem may use a second attention-based neural network to perform the listwise inference. According to some implementations, the technology is configured to process queries and candidate result items having different kinds and combinations of features. For instance, one kind of input query may include text-based features, structure-based features, and geographic-based features.

Abstract: A resource-efficient technique is described for producing and utilizing a set of trained embeddings. With respect to its training phase, the technique receives a group of sparsely-expressed training examples of high dimensionality. The technique processes the training examples using a distributed training framework of computing devices. With respect to its inference stage, the technique draws on the embeddings produced by the training framework. But in one implementation, the inference-stage processing applies a different prediction function than that used by the training framework. One implementation of interference-stage processing involves determining a distance between a query embedding and a candidate item embedding, where each such embedding is obtained or derived from the trained embeddings produced by the training framework. Another manifestation of inference-stage processing involves adjusting click counts based on identified relations among items embeddings.

Abstract: Methods of and systems for searching a catalog include parsing the items of the catalog into tokens, determining the frequency with which each token appears in the catalog, and storing the frequencies in a cache. Queries to the catalog are likewise parsed into tokens, and the tokens of the query string are compared to frequency values in the cache to identify a smaller search space within the catalog.

Abstract: Provided is a method, performed by an information processing device, of processing information by using an Internet of things (IoT) device, the method including: receiving, from a user, a web search query; fetching context information of at least one IoT device related to the web search query; automatically generating a synthetic web search query including the web search query and the context information of the at least one IoT device; and determining a control to be applied to the at least one IoT device by using a search result regarding the synthetic web search query.

Abstract: Embodiments relate to capturing referenced information in a report to resolve a computer problem. A report for a problem is determined to contain at least one referenced document, the report being generated based on at least one log. Relevant content is determined in the at least one referenced document based at least in part on hint information associated with the at least one referenced document and a term in the report. An enhanced report is generated for the problem of the computer equipment by inserting the relevant content into the report. A problem experienced by computer equipment is resolved by causing execution of at least one operation identified in the enhanced report.

Abstract: Systems and methods for identifying recommended topics are disclosed. An example method may be performed by one or more processors of a system and include identifying one or more attributes of a system user, identifying a subset of topics relevant to the system user based on analyzing the one or more attributes of the system user using an analysis model trained with a machine learning process to identify relevant topics for system users based on historical user attributes, generating, for each respective topic of the subset of topics, using the trained analysis model, a relevance score for the respective topic based at least in part on a most recent system page previously accessed by the system user and a current system page accessed by the system user, and generating one or more recommended topics for the system user based on the relevance scores.

Abstract: Aspects include receiving a search request, comparing at least one term of the search request to a plurality of metadata items, wherein each metadata item of the plurality of metadata items is associated with a respective asset of a plurality of assets, generating a plurality of scores, wherein each score of the plurality of scores is associated with a respective asset included in the plurality of assets, and wherein a value of each score of the plurality of scores is based on an identification of a user, selecting at least one asset of the plurality of assets in accordance with the plurality of scores and in accordance with at least one threshold, providing the at least one asset to a communications device that is associated with the user, and providing at least a portion of a metadata item included in the plurality of metadata items to the communications device.

Abstract: In the field of government engagement management, for users of an employee desktop web client, it is now possible, within the web client application, to search and read articles and/or knowledge content that has been authored to external locations. Due to this integration to external, third-party applications, content and/or articles can be displayed to an agent on the employee desktop web client graphical user interface. Agents can enter free text into a specific search field and review the results in summary form, and then select an article in HTML format to progress the current interaction with the client. An additional feature extending from this capability is to add an amount of coding to external knowledge content websites that are owned and/or operated by the owner of the system such that when the website is viewed through the third-party integration module, a button or icon appears within the website that when selected takes the agent to an appropriate form.

Abstract: An agent control device capable of executing a plurality of agents, the agent control device including a processor. The processor is configured to output information from an agent that is being executed to a reporting section, and change a reporting style of the reporting section for each of the agents.

Abstract: A framework for enhanced search is provided to enable the user to contextually search documents, based on specified events. A management framework provides systems for obtaining a user's contextual data and the user's interaction data for which a Context Graph and an Interaction Graph are created, respectively. The user's context data include the various events associated with the user. The user's interaction data include times and locations the user interacted with a particular document, which device was used, the type of interaction, etc. The data from the two graphs are correlated and joined into an event graph. Upon receiving a natural language query from a user, one or more contextual query elements are identified, and the event graph is spanned to identify matching results which are then provided to the requesting user device.

Abstract: A computing system generates, for each of a plurality of sessions, a session representation indicating the order through which a user navigated through a plurality of webpages during the session, and generates for each pair of session representations, a score indicating how similar the session representations in each pair are to each other. The computing system further generates a plurality of clusters by clustering the session representations based on the score for each pair of session representations, determines an optimal number of clusters based on key performance indicators corresponding to each cluster of the plurality of clusters, and generates a natural language description for each cluster of the optimal number of clusters based on the key performance indicators associated with each cluster.

Abstract: Systems and methods are described herein for disambiguating a voice search query that contains a command keyword by determining whether the user spoke a quotation from a content item and whether the user mimicked or approximated the way the quotation is spoken in the content item. The voice search query is transcribed into a string, and an audio signature of the voice search query is identified. Metadata of a quotation matching the string is retrieved from a database that includes audio signature information for the string as spoken within the content item. The audio signature of the voice search query is compared with the audio signature information in the metadata to determine whether the audio signature matches the audio signature information in the quotation metadata. If a match is detected, then a search result comprising an identifier of the content item from which the quotation comes is generated.

Abstract: A data aggregation system is described, wherein the data aggregation system may include: a plurality of sensors distributed throughout an environment; a tile database comprising a memory for storing a hierarchy of tiled layers, wherein each layer in the hierarchy of tiled layers comprises a plurality of tiles; a tiling server, the tiling server configured to: receive sensor data from one or more sensors in the plurality of sensors; assign the sensor data to a base tile in a first layer in the hierarchy of tiled layers based on one or more properties of the one or more sensors; retrieve one or more aggregate tiles from the tile database based on an identity of the base tile in the first layer, the one or more aggregate tiles each taken from one or more further layers in the hierarchy of tiled layers; determine aggregate sensor data for each of the retrieved one or more aggregate tiles based on the sensor data stored on the base layer tile; assign the determined aggregate sensor data to the corresponding one or

Abstract: A method is provided, including: storing comments generated in response to a content item served over a network; analyzing the comments to determine features associated with each of the comments; using a scoring model to score each comment based on the comment's corresponding features; receiving a request to serve a subset of the comments; responsive to the request, selecting a ranking of the comments that is one permutation from possible rankings of the comments, wherein selecting the ranking is in accordance with a probability distribution of the possible rankings that is based on the scores of the comments; serving comments identified by the selected ranking over the network to a client device; determining a dwell time on the served comments; applying the dwell time to update the scoring model.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer-readable storage media, for incorporating opinion information with semantic graph data. In some implementations, semantic graph data indicating objects relationships among the objects is stored. Opinion data indicating user sentiment with respect to the objects is generated. A particular object of the objects indicted by the semantic graph is identified. An opinion measure for the particular object is determined based on the generated opinion data. Output data is provided to a client device for presentation, where the output data is generated based on the opinion measure for the particular object.

Abstract: Systems and methods for providing any point in time functionality. With a storage system such as a VSAN, any point in time protection is achieved by combining a metadata stream with snapshots of the storage system. This allows snapshots to be generated in hindsight such that any point in time functionality is provided.

Abstract: A storage manager control application facilitates interactions with, and configurations of, a storage manager that manages one or more client computing devices of an information management system. The storage manager control application may be implemented using a variety of modules that allow a user to interact with the storage manager in various ways. Using the storage manager control application, a user may view a backup status of the information management system, and view the backup status of specific client computing devices that are managed by the storage manager. The user may also search for various entities and/or objects within the information management system using the storage manager control application. In addition, a user may view the status of one or more backup jobs configured within the information management system, as well as backup plans that have been established for one or more of the client computing devices.

Abstract: A system and method for de novo drug discovery using machine learning algorithms. In a preferred embodiment, de novo drug discovery is performed via data enrichment and interpolation/perturbation of molecule models within the latent space, wherein molecules with certain characteristics can be generated and tested in relation to one or more targeted receptors. Filtering methods may be used to determine active novel molecules by filtering out non-active molecules and contain activity predictors to better navigate the molecule-receptor domain. The system may comprise neural networks trained to reconstruct known ligand-receptors pairs and from the reconstruction model interpolate and perturb the model such that novel and unique molecules are discovered. A second preferred embodiment trains a variational autoencoder coupled with a bioactivity model to predict molecules exhibiting a range of desired properties.

Abstract: A computer system is provided that includes a matching engine and a freezing process. The matching engine freezes one side of a two-sided data structure when an order is determined to matchable. The freezing process starts a timer based on the matching determination. Orders that are handled by the matching engine while the side is frozen are added to a queue. When the timer ends, the orders in the queue are processed against those orders that are now resting within the data structure.

Abstract: A framework for reducing the number of textual items reviewed to determine the source of or reason for an anomaly in a time series that is used to track metrics in textual data is provided. According the framework, textual items in a time window corresponding to the anomaly are ranked according to the cross-entropy as determined by applying a language model to the relevant textual items and ranking textual items that most likely triggered an anomaly in time series data based on the cross-entropy value. In an aspect, a predetermined number of textual items having the highest cross-entropy are provided or all textual items having cross-entropy value higher than predetermine threshold are provided.

Abstract: Systems and methods are presented to facilitate caching of programming from broadcast and/or content streaming services to a user device for local playback. In addition, metadata can be provided to guide the user in selecting, caching, or playing the cachable content. In some exemplary embodiments, a receiver or other user device can automatically select the programming to be cached based on a user's historical topic interest and listening history. In some exemplary embodiments, the metadata can also include DRM information that can limit the ability of a user to playback the content, so as to comply, if necessary, with operative legal requirements related to usage of the content.

Abstract: Techniques for template generation from image content includes extracting information associated with an input image. The information comprises: 1) layout information indicating positions of content corresponding to a content type of a plurality of content types within the input image; and 2) text attributes indicating at least a font of text included in the input image. A user-editable template having the characteristics of the input image is generated based on the layout information and the text attributes.

Abstract: The present application describes a system and method for converting a natural language query to a standard query using a sequence-to-sequence neural network. As described herein, when a natural language query is receive, the natural language query is converted to a standard query using a sequence-to-sequence model. In some cases, the sequence-to-sequence model is associated with an attention layer. A search using the standard query is performed and various documents may be returned. The documents that result from the search are scored based, at least in part, on a determined conditional entropy of the document. The conditional entropy is determined using the natural language query and the document.

Abstract: A system for providing advertisements with search results in response to a search query comprises a front end and an advertisement server. The front end is configured: to receive a search query; to send a first search request to a search server and sending a first advertisement request to an advertisement server, wherein the first search request includes the search query or information based on the search query, and wherein the first advertisement request includes the search query or information based on the search query and an indication that an advertisement response is not to be provided; to receive search results from the search server; and to send at least some of the search results to the advertisement server in a second advertisement request, wherein the second advertisement request includes an indication that an advertisement response is to be provided.

Abstract: An information processing method to be carried out by a terminal that is configured to communicate with an electronic device, is provided. The electronic device may be configured to manage information related to a vehicle. The information processing method may include: receiving vehicle usage fee information related to a vehicle usage fee for the vehicle from the electronic device, through a communication interface of the terminal, the vehicle usage fee being based on an advertisement that is displayed on a display device of the vehicle; and displaying a first display that is based on the vehicle usage fee information, in a display area of the terminal.

Abstract: A respective preferred installation height above ground level is determined for each of a plurality of locations for a subscriber module for receiving a radio link from an access point in a wireless network, the access point having a given height above ground level and a specified location, and the subscriber module being situated within a given geographical area including the location of the access point. The method comprises accessing elevation data for the given geographical area, processing the elevation data to generate a preferred height data file representing a preferred height for a subscriber module to be wirelessly visible by the access point at each of the plurality of locations and processing the required height data file to provide output data indicating the preferred height of the subscriber module as a function of location.

Abstract: A device, system and method controlling operation of a client device via an intermediation server are provided. The server maintains indications of predefined data descriptions, associated with respective provider systems, comprising a subset of given predefined data descriptions to be provided at a client device during implementation of a computing-process flow in conjunction with communicating with the respective provider systems. The server: provides, to the respective provider systems, a communication from the client device received in conjunction with implementing a given step of the computing-process flow; and receives, from a given provider system, a response to the communication.

Abstract: Systems, apparatus, methods, and articles of manufacture provide for evaluation of and prediction for data queries (e.g., SQL statements). In one example implementation, a controller device trains a data model for classifying input SQL queries. The controller classifies each input SQL query by testing it against the data model, without requiring actual execution of the SQL query.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for analyzing images for generating query responses. One of the methods includes determining, using a textual query, an image category for images responsive to the textual query, and an output type that identifies a type of requested content; selecting, using data that associates a plurality of images with a corresponding category, a subset of the images that each belong to the image category, each image in the plurality of images belonging to one of the two or more categories; analyzing, using the textual query, data for the images in the subset of the images to determine images responsive to the textual query; determining a response to the textual query using the images responsive to the textual query; and providing, using the output type, the response to the textual query for presentation.