Abstract: Systems and methods are disclosed for suggesting expressive graphical content to a user to share within a messaging system. Expressive graphical content can include images, video, animated images, any of which may include sound, that a message system user is likely to share because it is related to a conversation between one or more users of the message system. Expressive graphical content may be content that is recent, humorous, and/or popular or trending in popularity, and can be shared without a copyright. A transcript of a message session between two or more users of a message system can be used to generate keywords for identifying and prioritizing expressive graphical content. The keywords can be prioritized and used as search terms to proactively suggest to the user to search for expressive graphical content. A search can be initiated automatically upon selecting one or more of the keywords.

Abstract: An analytics platform configured to create clusters of media content consumers and to automatically generate human-readable, text labels that describe each of the created clusters is disclosed. An analytics platform may use any of a number of clustering algorithms to group a population of media content consumers into a set of user clusters based on viewing history data, user profile data, and/or any other data associated with the consumers. Furthermore, the analytics platform may be configured to create, based on the viewing history data and/or other data associated with a selected sample of users from each cluster, a text label for each cluster. For example, if viewing history data for a population of users includes channel and genre identifiers associated with media content items accessed by users of the population, a generated text label may provide an indication of channels and/or genres popular with users of each particular cluster.

Abstract: To address the issue of latency in relational database systems, a computerized database management system comprising a database server, a processor, and associated random access memory is provided. The processor may configure an in-memory database in the random access memory to have tables, each containing a plurality of cells associated memory addresses, establish a direct table association between an origin cell and a target cell within the tables, receive a request to perform a join database operation of a first table including the origin cell on a second table including the target cell, dereference the direct table association for the origin cell to obtain a value of the target cell stored at the memory address for the target cell, execute the join database operation to produce a joined table including the value, and store the joined table at a new location in the random access memory.

Abstract: A knowledge management system and method may include receiving a first set of keywords used to describe a problem from a first user, storing in a coincidence database a domain sample related to the domain of the problem and a context sample from the set of keywords, searching a knowledge management repository containing potential solutions to the problem, determining that a solution to the problem does not exist in the knowledge management repository, locating a second user that has previously described the problem, determining that the first user and the second user are a collaborative match, and connecting the first user to the second user.

Abstract: A system optimizes performance of parallel queries by dividing a query into a plurality of sub queries. The query is divided according to a plurality of data ranges associated with the query. The system creates an available queue, and a defer queue. The system assigns the plurality of sub queries into the available queue. The system then processes, in parallel, the plurality of sub queries in the available queue. During the processing, if the system determines that a data range associated with a sub query comprises unavailable data, the system assigns the sub query to the defer queue as a defer range sub query. After completing the processing of the plurality of sub queries in the available queue, the system processes a plurality of defer range sub queries in the defer queue.

Abstract: Systems and methods for a data synchronization architecture to synchronize provide a forward synchronization of data from a helper instance to a customer facing instance and reverse synchronization of customer configuration data changes to the helper instance are disclosed. Helper instances may be configured to provide an upgrade or maintenance path for a customer facing instance. Helper instances may be a different version and have a different schema than the primary customer facing instance. Database triggers and business rules may be used to manipulate data calculated or discovered by a helper instance prior to synchronization to the primary instance. Database triggers and business rules may fire in the primary instance after each synchronization to perform necessary recalculations or adjust data associations based on data provided from the helper instance. Helper instances may queue data over time and then perform a synchronization of data from the queue.

Abstract: Disclosed herein is method and system for providing context based adaptive response to user interactions. A primary context of the user interactions is determined based on intents and associated named entities extracted from the user interactions. Further, secondary context of the user interactions is determined by detecting enquiry intent in user responses for queries provided for the primary context of the user interactions. Information related to primary and the secondary contexts are stored as Key Context Information (KCI) and is dynamically updated during the user interactions. Finally, context based adaptive responses are generated based on the updated KCI upon determining non-enquiry intent in subsequent user responses. The method of present disclosure maintains track of the user interactions and automatically detects changes in the context of the user interactions.

Abstract: Techniques related to instance recovery using Bloom filters are disclosed. A multi-node node database management system (DBMS) includes a first database server instance and a second database server instance. A recovery set includes a set of data blocks that have been modified by a first database server instance and not persisted. A Bloom filter is generated to indicate whether data blocks are excluded from the recovery set. The Bloom filter is sent to the second database server instance, which determines whether the Bloom filter indicates that a particular data block is excluded from the recovery set. Based on determining that the Bloom filter indicates that the particular data block is excluded from the recovery set, access to the particular data block is granted.

Abstract: A method for determining reach of a content item that is displayed on one or more client devices associated with at least one unresolved identifier. An unresolved identifier defines a context in which a client device accesses one or more online systems, the context not determined to be associated with a specific user. The method comprises identifying a set of unresolved identifiers, and identifying information describing one or more access events associated with each unresolved identifier. For each pair of unresolved identifiers, a similarity score for the pair is determined based on the identified information. Responsive to the similarity score exceeding a threshold similarity score, the pair of unresolved identifiers is clustered, the clustering indicating a prediction that the pair of unresolved identifiers are associated with a common user. Finally, for the reach of the displayed content item is determined based on the clustering of the set of unresolved identifiers.

Abstract: Techniques for detection of document similarity are provided. The computer-implemented method can comprise identifying, by an electronic device operatively coupled to a processing unit, a first pragmatic association of a first segment in a first document portion, the first pragmatic association indicating meaning of the first segment specific to a context of the first segment in the first document portion. The computer-implemented method can also comprise generating a first intermediate document portion from the first document portion by using the first pragmatic association to replace the first segment. The computer-implemented method can further comprise determining a similarity degree between the first document portion and a second document portion by comparing the first intermediate document portion with the second document portion.

Abstract: A backup method is provided for a file system for managing hierarchical storage including one or more storage media on a computer system including one or more computers. Each of the one or more storage media includes a data part which is separated from an index and on which updated information is appended and kept. Each of a plurality of pieces of data within the file system is managed to be migrated to at least any of the one or more storage media. The backup method includes backing up metadata of the plurality of pieces of data migrated within the file system; and backing up the index of each of the one or more storage media to which the plurality of pieces of data have been migrated, substantially at the same time point as the metadata.

Abstract: A method for managing a document search index. The method includes obtaining indexing terms for documents in a document repository to generate search index fragments, storing the search index fragments in a document search index, and constructing a hierarchical structure from a first set of stored search index fragments. The selected first set of search index fragments is arranged by size, in the hierarchical structure. The method further includes selecting, based on a minimal size, a second set of stored search index fragments from the hierarchical structure, merging the second set of stored search index fragments to obtain a larger search index fragment, storing the larger search index fragment in the document search index, and serving at least one search request using the larger search index fragment.

Abstract: Among other things, techniques and systems are disclosed for providing sync server process. Processing data sync sessions includes processing a request to initiate a sync session from one or more clients to update one or more data records that include one or more data items. The one or more clients are selectively identified as one of a trusted client and an untrusted client based on one or more properties of the sync session, wherein the trusted client is configured to support all dataclasses defined by the server. The one or more clients are detected as being configured to support field level differencing that sends changed data items only for the one or data records or record level differencing that sends all of the data items included in the one or more data records.

Abstract: An information processing device according to the present technology includes an action recognition unit that recognizes an operation action of a user based on sensor information, and an action representation generation unit that analyzes operation action data showing the operation action of the user recognized by the action recognition unit to generate an action segment represented by a meaning and content of the operation action from the operation action data.

Abstract: Aspects of the embodiments are directed to synchronizing at least a portion of a link-state database. A network element can lose an adjacency. The network element can transmit a request to a neighboring network element for synchronization of a link-state database. The request can include a version number of a last synchronized link-state database from the neighboring network element. The neighboring network element can determine whether the version of the link-state database is greater than or less than a copy of the link-state database stored by the neighboring network element. If the requested version number is less than the neighboring network element's link-state database version number, then the neighboring network element can send changes to the link-state database since the requested link-state database version number.

Abstract: A file deduplication processing system is provided. The system deduplicates raw files to generate deduplicated vault files and a descriptor indicating a storage location of each data chunk in the vault files corresponding to the raw files. When receiving a writing request of a write data, the system finds at least one data chunk including old data corresponding to the write data according to the descriptor, loads and recovers the data chunk whose boundary is not overlapped with a boundary of the write data in the vault file comprising the old data corresponding to the write data so as to generate an update data by incorporating the recovered data chunk and the write data, deduplicates the update data to generate a new vault file and stores the same in the chunk store, and updates a content corresponding to each data chunk in the descriptor.

Abstract: Techniques have been provided for storing data in a de-duplicated manner on a data storage system in a manner that allows for real-time reference to an index that is too large to fit within memory. This may be accomplished by segmenting the index into smaller segments, stored on disk. Only a subset of the segments may be loaded into memory at a given time. A predictive filter is stored in memory for each segment, allowing a de-duplication driver to quickly predict whether any given new block is likely to be indexed by each segment. Since identical blocks are often stored in long identical sequences (e.g., upon copying a disk image to a disk for a virtual machine), once a segment stored on disk is referenced many times in a short period, it is loaded into memory to allow the remainder of the long sequence to be de-duplicated.

Abstract: Embodiments are directed to managing data in a file system. A pre-fetch engine may receive requests from a client the file system that includes a pre-fetch storage tier and a file storage tier of storage devices. The pre-fetch engine determines a pre-fetch policy based on the requests such that the pre-fetch policy determines which blocks to copy to the pre-fetch storage tier. And, the pre-fetch policy may be associated with a score model that includes score rules where one of the rules may be associated with a client score. The pre-fetch engine may obtain scores associated with the score rules such that the scores are based on previous requests made by the client. In response to scores exceeding a threshold value, the pre-fetch engine may copy the blocks to the pre-fetch storage tier. The pre-fetch engine may update the scores based on the performance of the pre-fetch policy.

Abstract: An online system stores objects that may be accessed by users. The online system also stores indexes of terms related to different entity types of objects. When a user provides a search query, the online system compares the search terms with terms stored in the indexes. Based on the comparisons, the online system determines term features for entity types associated with an index. The online system provides the term features as inputs to a machine learning model. The machine learning model outputs a score for each entity type indicating a likelihood that the search query is for an object associated with the entity type. The machine learning model output is used by the online system to select one or more entity types that the user is likely searching for. The online system offers objects of the likely entity types to the user as results of the search query.

Abstract: Techniques are disclosed for identifying legitimate files using a hash-based cloud reputation using parts of a file to generate a hash value for reputation score lookup. A reputation service receives a request for a reputation score associated with a file. The request specifies a hash value for the file. The hash value is generated based on one or more parts of the file. The service identifies one of a plurality of file clusters that includes one or more files that matches to the specified hash value. The service determines a reputation score for the file based on the identified file cluster. The reputation score indicates a rating of the file based on a distribution of the file in a user base. The service returns the reputation score in response to the request.