Abstract: A method begins by a computing device of a dispersed storage network (DSN) selectively bringing online and taking offline storage units of a set of storage units of the DSN. When bringing a first storage unit of the set of storage units online in accordance with the selectively bringing online and taking offline storage units, the method continues with the computing device determining a rebuilding approach of the first storage unit. The method continues with the computing device bringing the first storage unit online in accordance with the rebuilding approach.

Abstract: This disclosure relates to systems, methods, and computer readable media for performing filtering of computer vision generated tags in a media file for the individual user in a multi-format, multi-protocol communication system. One or more media files may be received at a user client. The one or more media files may be automatically analyzed using computer vision models, and computer vision generated tags may be generated in response to analyzing the media file. The tags may then be filtered using Natural Language Processing (NLP) models, and information obtained during NLP tag filtering may be used to train and/or fine-tune one or more of the computer vision models and the NLP models.

Abstract: Methods, apparatus, and processor-readable storage media for automatically detecting data offloading methods using data bucketing and machine learning techniques are provided herein. An example computer-implemented method includes obtaining operations data and configuration data for one or more storage objects in a database; determining one or more times at which data offloading is to be carried out for at least one of the storage objects in the database, wherein determining the one or more times includes processing at least a portion of the operations data using one or more machine learning techniques; generating at least one data offloading protocol, comprising one or more data offloading methods, by processing at least a portion of the configuration data; and automatically executing, in accordance with the one or more determined times, the at least one generated data offloading protocol for at least a portion of the one or more storage objects in the database.

Abstract: This disclosure relates generally to a system and method for creating unified personalized content for an individual user receiving communications from a given website across separate platforms and channels.

Abstract: Generating an improved/more accurate geolocation database is provided. Given a dataset of reverse DNS hostnames for IP addresses, ground truth information, and a hierarchical geographical database, a machine learning classifier can be trained to extract and disambiguate location information from the reverse DNS hostnames of IP addresses and to apply machine learning algorithms to determine location candidates and to select a most probable candidate for a reverse DNS hostname based on a confidence score. The classifier can be used to generate an accurate geolocation database, or to provide accurate geolocation information as a service.

Abstract: In some examples, a system computes respective values for corresponding data value indicators added to and removed from a deduplication data store in which duplicated data values have been eliminated, where each respective data value indicator of the data value indicators represents presence of a unique data value in the deduplication data store. The system updates an estimator based on the respective values, to reflect an addition of a first data value indicator to the deduplication data store and a removal of a second data value indicator from the deduplication data store. The system computes, using the updated estimator, a parameter relating to data deduplication at the deduplication data store.

Abstract: The present application discloses an information push method and apparatus, a device, and a storage medium, which relate to the field of intelligent search technology. A specific implementation of the method in the present application lies in: acquiring a user feature and an item feature, where both the user feature and the item feature include click through rate information; determining a degree of similarity between the user feature and the item feature; determining, according to the degree of similarity, item information that is recalled; and pushing the item information to a user terminal. According to the present application, item information having a high click through rate may be acquired through a one-channel recall approach, so that information push efficiency is improved and the pushed item information better conforms to user needs.

Abstract: A graph query is executed against a graph index that connects actors with objects through edges. A graph ranking model is obtained and results of the graph query are ranked, using the graph ranking model, based upon edge data available from edges in the graph that match the query.

Abstract: A method and a system for garbage collection on a system. The method includes initiating a garbage collection process on a system by a garbage collector. The garbage collector includes one or more garbage collector threads. The method also includes marking a plurality of referenced objects using the garbage collector threads and one or more application threads during a preemption point. The method includes replicating the referenced objects using the garbage collector threads and marking for replication any newly discovered referenced objects found by scanning the application thread stack from a low-water mark. The method also includes replicating the newly discovered referenced objects and overwriting any reference to the old memory location.

Abstract: A graph query is executed against a graph index that connects actors with objects through edges. A graph ranking model is obtained and results of the graph query are ranked, using the graph ranking model, based upon edge data available from edges in the graph that match the query.

Abstract: In general, the invention relates to a method for managing data, the method includes obtaining data from a host, applying an erasure coding procedure to the data to obtain a plurality of data chunks and at least one parity chunk, deduplicating the plurality of data chunks to obtain a plurality of deduplicated data chunks, generating storage metadata associated with the plurality of deduplicated data chunks and the at least one parity chunk, storing the storage metadata in an accelerator pool, storing, across a plurality of fault domains, the plurality of deduplicated data chunks and the at least one parity chunk, and initiating storage metadata distribution on the storage metadata across the plurality of fault domains.

Abstract: Example object storage systems, bookkeeping engines, and methods provide quota usage monitoring for control entities, such as accounts, users, and buckets. An object data store is configured to enable control entities to access data objects associated with each control entity. Data objects are mapped to the control entities and the data objects are processed to identify object usage values corresponding to each combination of data object and control entity. Total usage values are calculated for each control entity and used to determine a data object access response for a target data object and associated control entities.

Abstract: Methods for file archiving using machine learning are disclosed herein. An exemplary method comprises archiving a first file of a plurality of files from a storage server to a tiered storage system, training a machine learning module based on file access operations for the plurality of files, determining one or more rules for predicting access to the archived files using the machine learning module, determining a prediction of access of the archived file based on the one or more rules and retrieving the archived file from the tiered storage system into a file cache in the storage server based on the prediction of access.

Abstract: A system for real-time search, including: a set of partitions, each including a set of segments, each segment corresponding to a time slice of messages posted to the messaging platform, and a real-time search engine configured to receive a search term in parallel with other partitions in set the set of partitions, and search at least one of the set of segments in reverse chronological order of the corresponding time slice to identify document identifiers of messages containing the search term; and a search fanout module configured to: receive a search query including the search term; send the search term to each of the set of partitions for parallel searching; and return, in response to the search query, at least one of the identified document identifiers of messages containing the search term.

Abstract: A system and method of performing a cross-datasource query to provide a cross-datasource query result. A composite shape comprises a composite shape query and a plurality of shape identifiers identifying a plurality of shapes. Each of the plurality of shapes comprises a datasource identifier identifying a datasource and a shape query defining a query to be performed on the datasource. The composite shape is used to perform the composite shape query across the datasources identified in the plurality of shapes to provide the cross-datasource query result.

Abstract: A method of joining a first dataset configured to store a set of data entries each identified by a respective key of a first type and a second dataset configured to store a second set of data entries identified by a respective key of a second type, the method comprising: selecting an intermediate mapping entity from a set of possible intermediate mapping entities, each mapping entity storing association between keys of the first type and keys of the second type; providing the selected intermediate mapping entity for use in joining the first data set with the second data set; wherein the step of selecting the intermediate mapping entity is based on the intersection weight between the first and second data sets via each of the intermediate mapping entities, wherein the intersection weight is the proportion of overlapping data entries between the first and second datasets.

Abstract: An example operation may include one or more of identifying a new entry, receiving one or more new entry signatures to approve the new entry to, such that the one or more new entry signatures each include a signature of the new entry and hashes of previously signed entries for a common block, verifying the one or more new entry signatures for commit by performing a signature verification of the new entry, determining the common block has more than one entry with a hash signed by a same member, and responsive to the signature check being performed on the new entry and determining the common block has more than one entry with a hash signed by a same member, designating one or more previously received entries as verified for commit without performing a signature verification of the one or more previously received entries.

Abstract: Performing snapshot conscious internal file modification for network-attached storage is presented herein. A file system can comprise a first component configured to modify, during a service request, storage for a subset of data blocks of a file—the service request not being recognized by an external entity as a change of content of the file. Further, the file system can comprise a second component configured to prevent, based on the service request, a copy of the storage from being created for servicing of a snapshot—the snapshot comprising a point-in-time copy of the file system.

Abstract: System and methods are disclosed for redacting analyzing unstructured data in a request for data associated with a data subject to determine whether the unstructured data is relevant to the request. The relevancy of pieces of the unstructured data may be determined by determining a categorization for each such piece of unstructured data and comparing them to known personal data associated with the data subject having the same categorization. Pieces of the unstructured data that do not match known personal data having the same categorization are redacted from the request before the request is processed.

Abstract: One or more techniques and/or systems are provided for implementing a pipeline used to generate, train, test, and implement a document scoring model for assigning document scores to documents. Features from various sources are combined to create a joined page level feature set, a joined domain level feature set, and a host level feature set. Numerical features and content features are extracted from ground truth documents and random documents. The numerical features are joined with the joined feature sets to create a set of joined features. The document scoring model is trained using the set of joined features and a training technique. A document is scored with a document score using the document scoring model based upon the content features and the set of joined features with document scores obtained during training.