Abstract: The present disclosure provides systems and methods that perform stochastic generative hashing. According to one example aspect, a machine-learned hashing model that generates a binary hash for an input can be trained in conjunction with a machine-learned generative model that reconstructs the input from the binary hash. The present disclosure provides a novel generative approach to learn hash functions through Minimum Description Length principle such that the learned hash codes maximally compress the dataset. According to another example aspect, the present disclosure provides an efficient learning algorithm based on the stochastic distributional gradient, which avoids the notorious difficulty caused by binary output constraints, to jointly optimize the parameters of the hashing model and the associated generative model.

Abstract: Methods, systems, and computer program products for systematic browsing of automated conversation exchange program knowledge bases are provided herein. A computer-implemented method includes generating one or more initial questions to be submitted to an automated conversation exchange program; generating one or more natural language variants of the initial questions; submitting the one or more natural language variants of the initial questions to the automated conversation exchange program; identifying one or more valid responses returned by the automated conversation exchange program in response to the submitted natural language variants of the initial questions; deriving one or more items of data from the identified valid responses; storing the derived items of data in an automated conversation exchange program knowledge base; and systematically browsing the automated conversation exchange program knowledge base in connection with one or more application tasks.

Abstract: This disclosure provides various techniques that may allow for key-level lock requests in a database to be managed by code that manages a memory buffer for the database. The transaction-level locks and other higher level locks may be managed by a separate lock manager. In an embodiment, the code inserts key-level lock requests in the record chain even if there is a conflict, and the key-level lock requests remain in the chain after the locks are subsequently granted. Preemption may also be supported in the record chain, in some embodiments.

Abstract: Since storage devices are activated in the issuance order of I/O commands, the time required for processing of a plurality of I/O commands based on one or more I/O requests may become long. When issuance destinations of two or more I/O commands are two or more storage devices in a non-activated state from among a plurality of storage devices, a computer system determines an activation order of the two or more storage devices on the basis of two or more I/O quantities which respectively correspond to the two or more storage devices. Each of the I/O quantities of the two or more storage devices is an I/O quantity in accordance with one or more I/O commands for the relevant storage device from among the two or more I/O commands. The computer system activates the two or more storage devices in the determined activation order.

Abstract: Queries received at a query engine may be executed for structured data and not-structured data. A query execution plan may be generated for the query that includes stateless operations to apply the query to the not-structured data at remote query processing engines. The remote query processing engines may perform the stateless operations and return results to the query engine. The query engine may generate a result for the query based on the results received from the remote query engine as well as results determined as part of applying the query to structured data. The result to the query may be returned.

Abstract: This disclosure provides various techniques that may allow for accessing values stored in a data structure that stores multiple values corresponding to database transactions using a skip list. A key may be used to traverse the skip list to access data associated with the key. The skip list maintains on ordering of multiple keys, each associated with a particular record in the data structure, using indirect links between data records in the data structure that reference buckets included in hash table. Each bucket includes pointers to one or more records in the skip list.

Abstract: A computer-implemented automated review method for structured query language (SQL) source code files seeking to access a data store comprises the steps of calculating an SQL complexity score; setting an issue flag if the SQL complexity score is greater than a threshold; checking the SQL source code for SQL-92 compliance; setting the issue flag if any of the SQL source code files is not SQL-92 compliant; comparing a type of data in the SQL source code files to be input to the data store with a type of a field in the data store set to receive the input data; setting the issue flag if the type of any input data does not match the type of a corresponding data store field; checking the SQL source code for SQL triggers set to activate when selected SQL commands are executed; and setting the issue flag if there are any SQL triggers.

Abstract: A computer-implemented technique is described herein for providing a digital content item using a generator component. The generator component corresponds to a sequence-to-sequence neural network that is trained using an adversarial generative network (GAN) system. In one approach, the technique involves: receiving a query from a user computing device over a computer network; generating random information; generating a key term using the generator component based on the query and the random information; selecting at least one content item based on the key term; and sending the content item(s) over the computer network to the user computing device.

Abstract: A data processing system backs up snapshots of storage objects to cloud storage by identifying to a cloud gateway a set of snapshots to be archived, and making the set available to the cloud gateway. Upon request by the cloud gateway, snapshot metadata is provided that includes offsets and lengths of data extents of the set of snapshots. On further request of the cloud gateway using the metadata, the data extents of the set of snapshots are provided to the cloud gateway, which uses the extents to create corresponding objects and stores the objects in the cloud storage, each object including the data extents and the snapshot metadata for a respective snapshot. Backup and restore are initiated by the gateway or by the data storage system in different embodiments.

Abstract: The present application relates to a computer network, a computer-implemented method, a computer system and computer program products for adaptively adjusting network responses to client requests received from a client device. A parsing module is introduced to generate a request data object based on data relevant to a client request. A matching module is introduced to: apply a learning method based on a final result generated in relation to calculating a matching score of potential service data objects, where the matching score is based on applying a matching algorithm; and generate a learning result of the learning method. A learning engine operable to adapt the matching algorithm of the matching module based on the learning result is introduced.

Abstract: A data generation stack executing on data processing circuitry may generate substitute data for provision in place of data sample from sensor circuitry. A configuration layer of the data generation stack may store a configuration file. A data ingestion and generation layer of the data generation stack may access the configuration file to determine static relationships among different sensors within the system. The data ingestion and generation layer may further determine dynamic relationships among the sensors. Based on a hybrid relationship that accounts for the static and dynamic relationships, the data ingestion and generation layer may generate substitute data for a sensor based on sampled data from another sensor. A data export layer of the data generation stack may access the generated substitute data for output.

Abstract: Method to provide data protection through light fidelity network is disclosed herein. Method starts with a data protection server (DPS) that is enabled with light fidelity (Li-Fi) receiving a backup request from a first client device through a Li-Fi network. The first client device is enabled with Li-Fi. DPS registers the first client device through the Li-Fi network and triggers to start the backup of the first client device to a storage included in DPS. DPS receives a registration request from a proxy server for the DPS via a network, and registers the proxy server. DPS receives via the proxy server a backup request from a second client device for backup to the DPS. The second client device and the proxy server are enabled with Li-Fi DPS triggers to start the backup of the second client device to the storage included in the DPS. Other embodiments are described.

Abstract: A data processing system comprises a loading server and a query server. The loading server is configured to extract updated data from at least two table databases and insert the updated data into the query server. The query server is configured to create an inverted index for querying the updated data, and the inverted index is indexed to an index table of the query server.

Abstract: To improve the responsiveness of a user device during a synchronization of a local calendar with a master calendar, the master calendar will provide the most-relevant events, relative to the current time, first to the user device. The user device is thus enabled to display those most-relevant events to the user.

Abstract: A technique for repairing damage to a file system stores compression headers in association with respective extents of in-line compressed data. A repair utility reads contents from the compression headers and applies the contents to repair damaged metadata structures that the file system uses to access the associated extents. In this fashion, the repair utility restores access to compressed data, which would otherwise be lost.

Abstract: Systems and methods are provided for tracking System Management Facility (SMF) record types using a small array. An index entry can be modified by adding an extension that tracks what record types are missing, rather than what record types are included in a particular history file. As log data comprising a plurality of history files is received, a flag bit in the header of each history file indicates that the associated history file comprises extended record types. An extension is added to each of the index entries associated with the history files that contain extended record types. Extensions may indicate record type gaps in the associated history file. Upon receiving a query for a particular record type of data stored in the data store, the extension can be utilized to determine which history files do not have the particular record type and can be skipped in the search.

Abstract: Methods, systems and computer program products for clustering pages into headline clusters are provided by collecting web data, identifying pages from the web data, tokenizing unique words in each page, recognizing unique entities in each page, detecting media links in each page, and constructing a plurality of vector representations of each page. A first dimension of each vector representation includes the unique words tokenized in each page, a second dimension of each vector representation includes the unique entities recognized in each page, and a third dimension of each vector representation includes the media links detected in each page. The vector representations are, in turn, clustered.

Abstract: Described herein is a framework for collecting event related tweets. In accordance with one aspect of the framework, an initial set of keywords is constructed from a reference source. Tweets are collected from a messaging stream using the initial set of keywords for a first time window. The collected tweets are filtered to generate a candidate keywords set. The selected tweets of the candidate keywords set are grouped into a plurality of clusters. The clusters are classified into event related and non-event related clusters. The initial set of keywords is updated to obtain a new set of keywords.

Abstract: A tuning apparatus and a tuning method are provided, which make it possible to provide higher-accuracy searching tuning, while still reducing work load for the tuning executed every query language used to operate the relational database, and provide optimum indexes considering effect of the performance of commands not being processed for the tuning. Commands issued using query languages such as SQL are aggregated as acquisition patterns each consisting of search condition information and search method information which are for acquiring search results of data from a relational database. Predetermined execution information provided when the issued commands are executed is made to correspond to the acquisition patterns, thereby providing tuning information to each of the acquisition patterns.

Abstract: The disclosed embodiments illustrate methods and systems for detecting personal life events of users. The method includes training classifiers based on a set of features extracted from each of an annotated first set of social media data. The first set of social media is associated with one or more first categories. Further, the first set of social media data are annotated by one or more crowdworkers based on one or more second categories. The method further includes extracting a second set of social media data of one or more users, associated with the one or more first categories, from the one or more social media platforms. The method further includes categorizing the extracted second set of social media data into the one or more second categories by use of the trained classifiers. The categorization is further utilized to detect the personal life events of the one or more users.