Abstract: The subject technology determines that a first visible version of a key is not one of a N number of newest versions of the key based on searching for the key in a regular space to locate the first visible version of the key. The subject technology performs a search of an undo space to locate a second visible version of the key. The subject technology determines whether the first visible version or the second visible version of the key is newer. The subject technology provides a newer version of the key between the first visible version and the second visible version of the key.

Abstract: A method for synchronizing a file system (FS) and a remote file system (RFS) includes monitoring the FS for FS events, generating FS event records, receiving RFS event records of RFS events, generating file system operations (FSOs) based on the FS and RFS event records, and communicating the FSOs to the FS and RFS to synchronize them. A method for generating the FSOs includes accessing a plurality of FS and/or RFS event records, processing the accessed records to generate processed event records, generating the FSOs based on the processed event records, and outputting the FSOs to cause synchronization of the FS and RFS. Systems are also described. The invention facilitates event-based, steady-state synchronization of local and remote file systems.

Abstract: Methods, systems, and computer programs are presented for scheduling and executing request plans using an opportunistic approach. An opportunistic scheduler generates a request plan for a request on a cloud data platform, the request plan comprising a plurality of operations and identifies a plurality of contingent operations from the plurality of operations of the request plan. The opportunistic scheduler schedules the plurality of contingent operations for execution and sets the scheduled plurality of contingent operations to execute at a specific position in the request plan. The opportunistic scheduler sets remaining operations for execution by any available thread as threads that are processing the request plan become available and processes the request plan according to the scheduled plurality of contingent operations.

Abstract: A data classification system obtains a data set including first data associated with a first characteristic and second data associated with a second characteristic. In response to obtaining the data set, the system uses a classification model to generate a classification by classifying the first data into majority data and the second data into minority data. From the classification and using the classification model, the system determines a loss. The system modifies the classification model based on this loss to generate an updated classification model. The system uses the updated classification model to re-classify a subset of the first data into the minority data and a subset of the second data into the majority data.

Abstract: A system generates a topic graph based on the SERP data for high-ranking keywords in a search engine. Clustering may be based on (for example) degrees of intersection between links in search results of keywords from the SERP data, or keyword embeddings on the SERP data. The topic graph loosely clusters the keywords, such that the keywords have at least a threshold degree of similarity to their clusters, but not necessarily to all the other keywords in the cluster. As a consequence of the loose clustering, a given topic contains keywords that represent different aspects of the same concept, such that a content viewer would likely be interested in a piece of content that addresses the different aspects, and a search engine would be more likely to highly rank the content within its search results for one of the keywords.

Abstract: A method of preconditioning time-series data for improved efficiency. The method includes repeatedly recognizing new time series entries for each of a plurality of tracked entities. For each of the plurality of tracked entities, a new time-series data entry for the tracked entity is promoted responsive to the new time-series data entry including a changed value in a normalized field. However, a previous time-series data entry for the tracked entity is maintained responsive to the new time-series data entry including an unchanged value in the normalized field. The method further incudes generating a synthetic time-series data entry for every tracked entity without a time-series data entry in a fixed time period.

Abstract: A data processing system implements identifying one-way correctness issues in datasets of large computing systems including a first computing system and a second computing system. The second computing system is associated with a dependent dataset that includes references to data in a second datastore associated with the first computing system. These references updated in response to changes to the data referred to by these references. However, errors can cause the two datasets to become out of sync. The system herein implements automatically identifying unattached items in a dependent dataset that references items in a reference dataset that is no longer present in the reference dataset by comparing the dependent dataset with the reference dataset, and automatically causing the second computing system to delete the unattached items from the dependent dataset.

Abstract: A computer program product is provided wherein a batch process operates on structured data in multiple tables, including a first and a second source table, and a first and a second processing table. An additional key is added within each natural key to at least the second source and processing table for identifying a sequence of rows for updating.

Abstract: A method includes receiving a request to write a data block to a volume resident on a multi-tenant storage array, wherein the request is associated with a first tenant of the multi-tenant storage array, and determining whether the data block matches an existing data block on the multi-tenant storage array, wherein the existing block corresponds to a second tenant. In response to determining that the decrypted data block matches the existing data block: encrypting the existing data block with a shared volume encryption key; encrypting the shared volume encryption key with a first tenant encryption key and providing the shared volume encryption key encrypted with the first tenant encryption key to the first tenant; and encrypting the shared volume encryption key with a second tenant encryption key and providing the shared volume encryption key encrypted with the second tenant encryption key to the second tenant.

Abstract: Computer-implemented methods and systems are provided which are suitable for implementation in transaction validation nodes of a blockchain network. Modified blockchain node structures, network architectures, and protocols for handling large numbers of transactions and large transaction blocks are described. The invention is particularly suited, but not limited, to use with the Bitcoin blockchain. A computer-implemented method is provided which includes: (i) receiving transactions from the blockchain network; (ii) validating transactions received from the blockchain network; (iii) maintaining a distributed, decentralized storage of validated transactions with other transaction validation nodes in the blockchain network; and (iv) distributing data corresponding to said validated transactions to the blockchain network for mining.

Abstract: An apparatus and method for using a data mesh to structure unstructured data is provided.

Abstract: This disclosure relates generally to multi-class multi-label classification and more particularly to contradiction avoided learning for multi-class multi-label classification. Conventional classification methods do not consider contradictory outcomes in multi-label classification tasks wherein contradictory outcomes have significant negative impact in the classification problem solution. The present disclosure provides a contradiction avoided learning multi-class multi-label classification. The disclosed method utilizes a binary contradiction matrix constructed using domain knowledge. Based on the binary contradiction matrix the training dataset is divided into two parts, one comprising contradictions and the second without contradictions. The classification model is trained using the divided datasets using a contradiction loss and a binary cross entropy loss to avoid contradictions during learning of the classification model.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for determining and recommending the types of digital components that content providers can generate and provide for distribution to client devices. In one aspect, a method can determine whether a content provider has not previously provided a first digital component of a first media type. A first set of user interaction data can be obtained and input into a machine learning model. The model can can output result data for expected affirmative user actions related to the first digital component of the first media type. Based on the result data, a recommendation specifying whether the content provider should provide the first digital component of the first media type can be generated and provided to the content provider.

Abstract: Various embodiments of the present disclosure provide methods, apparatus, systems, computing devices, computing entities, and/or the like for classifying unstructured data by: (i) generating probability scores of natural language classification labels for classifying unstructured data elements using an NLP-based model, (ii) generating probability scores of structured data classification labels for classifying the unstructured data elements using a classification-based model, and (iii) assigning classifications labels based on: a) the probability scores of the natural language classification labels if a distance measure difference associated with the natural language classification labels is greater than a predetermined distance, or b) a determination using an ensemble model if the distance measure difference is less than a predetermined distance.

Abstract: Semantic type-ahead suggestions are provided for natural language database searches. The system maintains an index of previous natural language database searches and a sorted prefix map based on words used in previous natural language database searches. The system receives user input for a new search, creates a search list based on the user input and user input synonyms, searches the index using the search list, and creates a candidate match list with matching previous searches in the index. If the system determines that the user input includes a database object reference, it filters out searches in the candidate match list that are associated with a different database object. The system ranks the remaining searches in the candidate match list and displays the top n-ranked previous searches in the candidate match list as type-ahead suggestions. The system repeats the steps for each incremental input character and updates the type-ahead suggestions accordingly.

Abstract: A method for managing big metadata using columnar techniques includes receiving a query request requesting data blocks from a data table that match query parameters. The data table is associated with system tables that each includes metadata for a corresponding data block of the data table. The method includes generating, based on the query request, a system query to return a subset of rows that correspond to the data blocks that match the query parameters. The method further includes generating, based on the query request and the system query, a final query to return a subset of data blocks from the data table corresponding to the subset of rows. The method also includes determining whether any of the data blocks in the subset of data blocks match the query parameters, and returning the matching data blocks when one or more data blocks match the query parameters.

Abstract: A system and method for data discovery. A method includes performing a scan of a plurality of snapshots, each snapshot corresponding to a respective disk of a plurality of disks; identifying a plurality of data store files in the plurality of disks based on file metadata found during the scan; and detecting at least one data store based on the identified plurality of data store files, wherein each of the at least one data store is in a disk of the plurality of disks including one of the plurality of data store files.

Abstract: According to one aspect, data collected from various modules on an autonomous vehicle is synchronized, and snapshots of the synchronized data, referred to herein as ‘data snapshots’, are created. Contextual information associated with the data snapshots, e.g., indexes and/or metadata, is collected and/or derived, and the contextual information is stored along with the snapshots. The data snapshots may then be substantially searched, using the contextual information, to identify particular scenes for analysis. In addition to searching for particular scenes, series of scenes may be searched using the contextual information such that scenarios encountered by the autonomous vehicle over time may be identified.

Abstract: A list of command entries is displayed in a search interface, each of the command entries representing one or more commands of a plurality of commands of a search query. The list of command entries are displayed in a sequence corresponding to the plurality of commands of the search query. Based on a user interaction with a designated command entry in the displayed list of command entries, the displayed list of command entries is modified with respect to the designated command. Furthermore, the search query is automatically modified with respect to the corresponding one or more commands represented by the designated command entry. The modification can include causing the designated command entry to be removed from or reordered in the displayed list of command entries and the automatic modification cam include causing the corresponding one or more commands to be removed from or reordered in the search query.

Abstract: A computing device, method, and system are provided of improving data quality to conserve computational resources. The computing device receives a raw dataset. One or more data quality metric goals corresponding to the received raw dataset are received. A schema of the dataset is determined. An initial set of validation nodes is identified based on the schema of the dataset. The initial set of validation nodes are executed. A next set of validation nodes are iteratively expanded and executed based on the schema of the dataset until a termination criterion is reached. A corrected dataset of the raw dataset is provided based on the iterative execution of the initial and next set of validation nodes.