Abstract: Systems, methods, and apparatus, including computer programs encoded on a computer storage medium for managing network traffic to and from a server configured to: (i) receive, from a client device, a query in a natural language, and (ii) generate a response to the query in the natural language. In one aspect, a method includes: receiving, from the client device via a network connection, a network message including a new query for the server; processing the new query, using a text encoder, to generate an embedding vector of the new query; identifying, from amongst multiple entries of a vector database, a particular entry based on a similarity metric between: (i) the embedding vector of the new query, and (ii) an embedding vector of a particular query stored in the particular entry; and determining whether the similarity metric is greater than a threshold similarity value.

Abstract: Systems as described herein may include detecting live browser information that a user navigates to a first website displayed in a first open browser tab and a second website displayed in a second open browser tab. A data sharing server may provide the live browser information to a machine learning model as input. Based on feedback from the machine learning model, one or more similar products displayed in the first website and the second website may be determined. The data sharing server may detect an update on the one or more similar products, and cause a user device to display an alert indicating the update on the similar products.

Abstract: A method is disclosed. In the method, a data generation process can continuously generate data in real time. The data generation process can store the data into discrete data blocks. An analyzer process can run analytical queries on the data from the data blocks. After the analytics is complete for different data blocks, data can be removed from the respective data blocks. The empty data blocks can be returned back to the generation process for reuse. The data blocks can be shared resources between the generation and the analyzer processes. The data can be stored in a directly queryable format. Though at any given time a given analytical query can run on a single data block, the analyzer process can preserve certain important records from that data block to be used while analyzing subsequent data blocks at a later time.

Abstract: An embodiment for encoding permutation-invariant representations of linearized tabular data. The embodiment may receive input including tabular data and linearize a column or row within the received tabular data. The embodiment may automatically assign an increasing sequence of position identifiers to each non-delimiting tokenized cell in the linearized column or row until a header delimiter is reached. The embodiment may, in response to reaching the header delimiter, automatically assign a monotonically increasing sequence of position identifiers for each non-delimiting tokenized cell positioned after the header delimiter, restarting from an integer corresponding to 1 greater than the position identifier assigned to the header delimiter for each non-delimiting tokenized cell positioned after cell delimiters.

Abstract: The present disclosure discloses a method and system for refreshing current page information. The method may include: obtaining currently displayed information data in response to receiving an information refresh request from a user; determining possibly-interested information of the user based on the information refresh request and the currently-displayed information data; displaying the possibly-interested information of the user. The present disclosure may determine the possibly-interested information of the user based on the currently-displayed information data, and recommend data content of a different type from the currently-displayed information data to the user, which may broaden a browsing horizon of the user, and improve user experience.

Abstract: This disclosure describes a natural language question (NLQ) query service within a service provider network that provides row level security (RLS) for autocomplete during entry of NLQs and fuzzy matching in NLQ answering. The rules take the form of per-user predicates such as Tim can only see rows with region=US. In configurations a complex extraction and preprocessing pipeline to extract distinct combinations of values against RLS predicate “rule keys” is used. Those distinct values are indexed along with grouped rule keys to enable pushing down predicates at auto-complete time. This enables pushing part of RLS rule handling to ingestion time of a dataset rather than handling all RLS rule handling at query time, enabling meeting of latency goals. In configurations, a single logical document of unique cell values is split into multiple documents with a subset of rule keys to handle scalability limits.

Abstract: Disclosed are some implementations of systems, apparatus, methods and computer program products for implementing a scalable computing system. The scalable computing system includes an intermediate system that facilitates communications between a core server system and a third-party system. The core server system processes a client request for a third-party service in association with a web page having a corresponding web address. The intermediate system communicates with the core server system to obtain a session token, and transmits the session token and web address to the third-party system. The third-party system may then access the web page via the web address using the session token.

Abstract: A graph data processing method includes: acquiring target graph data to be processed; compiling statistics on the target graph data according to a first preset rule, so as to divide the target graph data into a plurality of graph data blocks and determine a boundary value and weight of each of the plurality of graph data blocks; and storing the boundary value and weight of each of the plurality of graph data blocks in a corresponding memory according to a second preset rule, so as to schedule the target graph data during a graph calculation process by use of the boundary values and the weights.

Abstract: A system to facilitate quality checks on a plurality of attributes is disclosed. The system includes a processing subsystem including a screening module to filter at least one document from documents. The processing subsystem includes a data extraction module to extract attributes from the at least one document. The processing subsystem includes a quality checking module to highlight the attributes extracted in a user interface associated with quality checking levels. The quality checking module is to modify at least one of the attributes associated with a first level based on a first level user response. The quality checking module is to assign at least one tag and at least one colour to each of the attributes. The quality checking module is to display the attributes in the user interface associated with a second level to validate the plurality of attributes based on a second level user response.

Abstract: System and method for behavioral and contextual data analytics are disclosed. An example computer system to process observational data received from a wireless device includes a memory including machine readable instructions and a processor to execute the instructions to: process the observational data to identify temporally adjacent applications to generate usage metric data, the observational data including application usage data; build a behavior model based on the identified temporally adjacent applications, the behavior model to describe user behavior associated with the wireless device; and apply the behavior model to predict a usage duration of a second application in response to usage of a first application.

Abstract: Disclosed examples access data from a database, the data stored across multiple registers of the database; determine (a) a maximum rank for each of the multiple registers and (b) a maximum rank count for each of the multiple registers; determine a frequency distribution based on the maximum ranks and the maximum rank counts; and generate a report including at least one of the frequency distribution, the maximum ranks, or the maximum rank counts.

Abstract: Methods and systems for managing the operation of data processing systems are disclosed. The data processing systems may provide computer implemented services to any type and number of other devices and/or users of the data processing systems. To improve the likelihood of the data processing systems being able to provide the computer implemented services, a system may proactively attempt to identify and remediate attempts to limit access to data stored in the data processing systems. To do so, a security framework may be implemented by the system. The security framework may include both proactive identified of and remediation of maliciously locked data structures based on user data access patterns.

Abstract: A method can include receiving requests to perform actions in a file system environment. The method can include populating a first bucket with first tokens. The first bucket can be associated with actions in the file system environment. The method can include populating second buckets, which can correspond to different tenants, with corresponding second tokens based on priorities of the tenants. The second tokens may correspond to allowable actions on behalf of the tenants. Each token of the first tokens and the second tokens may be in one-to-one correspondence with a single action. The method can include prioritizing the second buckets. The method can include generating an execution list for executing the requests. The method can include executing the execution list based on the first tokens and the second tokens.

Abstract: A system, method, and memory for handling real-time transactional events is disclosed. The exemplary system a processor to detect and add an event to a queue and identify an associated event type. Event types can require downstream processing by at least one provider. The processor decomposes the event into tasks by comparison to event types which associate downstream provider requirements to tasks and routes. One or more routes are assigned to the tasks, each route defined by at least one processor of a plurality of processors. One or more routes are defined by transmitting a request for downstream processing to the at least one provider. The processor, upon detecting incomplete performance of the one or more routes, either updates the tasks associated with the incomplete route, or marks the event associated with the route as incomplete in the queue.

Abstract: A method converts a request received from a client application to a query by a serverless function operating on a virtual machine instance provisioned responsive to the request. The query is transmitted to an application database corresponding to the client application. A result is received from the application database. The result is converted into a response and the response is transmitted to the client application in response to the request. Display of the client application is updated based on response with the result.

Abstract: Methods and systems of managing database failure based on transaction request time is disclosed. A method includes copying a first dataset stored in a primary deployment to a secondary deployment to generate a second dataset. The method includes determining a first arrival time of a first request to perform a first transaction. The method includes determining a second arrival time of a second request to perform a second transaction. The method includes executing the first transaction on the first dataset at the primary deployment in response to determining the first arrival time of the first request to perform the first transaction. The method includes executing the second transaction on the second dataset at the secondary deployment in response to determining the second arrival time of the second request to perform the second transaction.

Abstract: Techniques are disclosed for managing dependencies in an orchestration service. A computer-implemented method can include operations performed by a declarative infrastructure provisioner (DIP). In some embodiments, the DIP parses configuration data associated with a computing system and generates a directed acyclic graph (DAG) for booting a first resource. The DAG may specify a dependency of the first resource on a capability of a second resource. The DIP may traverse the DAG and determine, based at least in part on the traversal, that the dependency has been reached. The DIP may publish, to a scheduling process, an indication that the first resource is awaiting availability of the capability of the second resource. In some embodiments, the DIP receives a subsequent indication that the capability is available, regenerates the DAG, and recommences traversal of the DAG. Additional operations for booting the first resource may be performed in accordance with the recommenced traversal.

Abstract: The integration of news about members of a group into symmetrical group communications is described. Group members are delivered news items and information regarding members of the group, and are able to communicate with other members of the group in response to the news items. One embodiment dynamically generates aggregated RSS news feeds for a group of people or entities and correlates that aggregated data with biographical information, audio, and video media to provide rich information about the relevant parties to the news story. The embodiment also provides the capability of transmitting email or other electronic messages to the parties in the news story to facilitate two-way communication while maintaining the group members' privacy and anonymity.

Abstract: Certain aspects of the present disclosure provide techniques for managing a search engine based on search performance metrics. An example method generally includes dividing a set of search history data into a first subset of search history data and a second subset of search history data. The first subset of data is associated with interaction with search results, and the second subset of data is associated with non-interaction with search results. A first quality score is generated for searches in the first subset of data. A second quality score is generated for searches in the second subset of data based on different search intents identified for each temporally related group in the second subset of data. An overall quality score is generated for a search engine, and one or more actions with respect to the search engine are taken based on the overall quality score.

Abstract: Disclosed herein are embodiments of systems and methods for synchronizing file-catalog table with a file stage. In an embodiment, a data platform receives a notification of a modification to one or more files in a file stage. The file stage includes data storage having a storage location. The data platform updates, based on receiving the notification of the modification, a first file-catalog table for the file stage based on the modification. The first file-catalog table includes a row for each file in the file stage, as well as a column for each of one or more metadata properties of the one or more files in the file stage.