Abstract: A data query method includes: providing a data query interface; obtaining an aggregation query task from the data query interface, where the aggregation query task is used to query for a total amount of data that meets a query condition; and displaying, in the data query interface, a first estimated query result of the aggregation query task, where the first estimated query result includes a first total amount of data that meets the query condition and that is obtained through approximate estimation based on a first partial exact query result.

Abstract: An interface management method includes obtaining interface call records of a plurality of microservices in an application, where the interface call records include an interface call provided inside the application by the microservice and/or an interface call provided outside the application by the microservice; drawing an interface call relationship diagram of the application in a running state based on the interface call records of the plurality of microservices; drawing a boundary of the application in the interface call relationship diagram based on microservice registration information of the application; and when there is an interface call that breaks through the boundary of the application in the interface call records, displaying, to a user, an interface call that has a security risk, for example, the interface call that breaks through the boundary of the application.

Abstract: A computing acceleration method based on a cloud computing technology and a cloud management platform are provided. In the method, the cloud management platform may obtain a job task entered by a tenant, allocate, to at least one programmable compute device for processing, processing logic that is in job processing logic carried in the job task and that is applicable to a programmable compute device, and allocate, to at least one server for processing, other processing logic that is in the job processing logic and that is applicable to a server. This solution can be used to provide surplus cloud resources for the tenant to use, to obtain higher revenue and improve cloud resource utilization.

Abstract: There is provided a method for compressing an Artificial Intelligence (AI) model comprising a neural network. The method involves running the model using calibration data, followed by evaluating the activation sparsity and the importance of each model layer. Based on this evaluation, each layer is more or less compressed, until a target compression ratio is achieved. In some embodiments, the type of compression (quantization, or low rank decomposition) is based on the rank deficiency of a layer. In some embodiments, the compression may be adjusted dynamically at inference time.

Abstract: A data analysis method includes receiving input content of a user, performing intent classification based on the input content, and determining a task type indicated by the input content. The method further includes, when the task type indicated by the input content is an instruction-based task, determining a subtask type of the instruction-based task through instruction keyword matching, and obtaining historical dialog–associated content of the input content. The method further includes processing the historical dialog–associated content based on the subtask type to obtain an adjusted data analysis result.

Abstract: A data processing method includes: obtaining a custom row-level expression, where the custom row-level expression is for specifying a condition for determining a field of a target attribute of each piece of row data that is in a first data table as cold data; obtaining the field of the target attribute of each piece of row data in the first data table using the custom row-level expression, where all pieces of row data in the first data table are stored in a first storage medium; determining cold data, where the cold data is row data whose field of the target attribute meets a specified field; and compressing or migrating the cold data to a second storage medium.

Abstract: An embodiment of this disclosure provides an internet of things technology-based message transmission configuration method applied to the field of internet of things technologies. The method is applied to an internet of things platform that is run on infrastructure. The internet of things platform obtains message transmission configuration information including a channel selection policy. The internet of things platform constructs a plurality of message transmission channels based on a channel identifier and a message transmission condition that are of each message transmission channel of the plurality of message transmission channels and that are included in the channel selection policy.

Abstract: A data sharding method includes obtaining a plurality of pieces of stored data, and inputting the plurality of pieces of data into a ranking model to obtain a score corresponding to each piece of data, where input information of the ranking model includes the plurality of pieces of data, output information of the ranking model includes the score corresponding to each piece of data, the ranking model is determined based on the plurality of pieces of data and at least one query command, and the query command is used to query target data from the plurality of pieces of data. The method further includes determining, based on the score corresponding to each piece of data, file blocks respectively corresponding to the plurality of pieces of data.

Abstract: A data processing method includes a data processing system that receives a first write request, where the data processing system includes a first compute cluster, a second compute cluster, and a shared storage, the shared storage is configured to store data of a first data table, the first write request is used to write first data into the first data table, and the first data table belongs to the first compute cluster. The second compute cluster writes, into the shared storage, the first data that belongs to the first data table. The second compute cluster sends, to the first compute cluster, first metadata corresponding to the first data. The first compute cluster stores the first metadata corresponding to the first data.

Abstract: Embodiments of this application disclose a neural network training method. A first module in a federated neural network is deployed in each of a plurality of first devices. The first module includes a feature extraction module. A plurality of second modules in the federated neural network and early exit modules connected to the respective second modules are deployed in a second device. It can be learned that, in a federated learning process, the second device may include a plurality of early exit nodes, and each early exit node corresponds to one second module and a corresponding early exit module connected to the second module. In this way, target networks of different structures may be flexibly deployed in different first devices. For example, flexible scheduling may be performed based on resource statuses of different first devices, such that each first device can implement efficient data processing.

Abstract: A method for providing a cloud service includes receiving a service request sent by a terminal, where the service request includes information about a total resource quantity required by a user; and providing one or more cloud service solutions for the terminal, where at least one of the one or more cloud service solutions includes a plurality of virtual machines, and a total resource quantity of the virtual machines in the one or more cloud service solutions meets the information about the total resource quantity in the service request.

Abstract: The present disclosure provides a data processing method and related products, the method includes: consuming a first event in a data stream based on a first data structure corresponding to a first function, where the first function is used for finding a target value for a target attribute in the data stream, and a size of a resource occupied by the first data structure in a memory is determined according to an accuracy of a first estimation and with a sublinear cost, and the first estimation is an estimation of the target value; and obtaining the first estimation as an output of the first function based on the consumption.

Abstract: This disclosure provides a method and an apparatus for programming. The method includes: identifying a first target profile from a set of profiles according to information related to a target user. The set of profiles is obtained according to information related to users, and each profile in the set of profiles corresponds to a category of the information related to the users. The method further includes determining one or more candidate models from a model database according to the first target profile. The models in the model database are trained models obtained through subsets of a training dataset. The models are used for programming tasks. The subsets are obtained by dividing the training dataset based on the set of profiles. The method further includes obtaining a target action from one or more candidate actions generated by applying the one or more candidate models to context.

Abstract: The present disclosure relates to access control methods, access control ends, and clusters. The One example method includes receiving a target data packet, where the target data packet includes a to-be-matched value, obtaining target code of at least one rule from an access control list (ACL), where the target code includes at least one matching logic instruction and at least one matching item corresponding to the at least one rule, and each matching item is used to determine whether the to-be-matched value matches a rule corresponding to the matching item, executing the at least one matching logic instruction based on the at least one matching item and the to-be-matched value to obtain a matching result between the to-be-matched value and the at least one rule, and based on the matching result, allowing the target data packet to pass through or preventing the target data packet from passing through.

Abstract: A code generation method includes: after receiving a code generation request that carries a first question and is sent by the terminal device, the cloud computing platform decomposes the first question to obtain a step of resolving the first question; then, the cloud computing platform searches a code knowledge base based on each step in the step of resolving the first question, to obtain code knowledge matching each step; and finally, the cloud computing platform inputs each step in the step of resolving the first question and the code knowledge matching each step into a code generation model, to obtain code generated by the code generation model, and sends the code to the terminal device, for the terminal device to present the code.

Abstract: A data processing method includes setting a connection processing module to obtain a connection request of a user for a database system, and a thread group bound to the user is created for the user based on the connection request, such that each created thread group processes only a data operation request of an user, thereby ensuring independence of processing resources between different users and avoiding mutual interference between different users. In addition, a load detection module of the database system can detect a load status of a created thread group, so as to delete some thread groups that have not loaded for a long time.

Abstract: A policy preview method includes receiving, by a cloud management platform, through an editing interface, a first access control policy and assignment information that are input by a tenant, assigning, based on the first value, a value to the first parameter in the first access control policy, to obtain a second parameter with the value assigned and a second access control policy including the second parameter, and displaying the second access control policy to the tenant through a preview interface, where the second access control policy indicates whether the authentication performed by the cloud management platform on the tenant succeeds or fails.

Abstract: A method for data processing includes obtaining data from a user; obtaining a probabilistic knowledge graph (PKG), where facts in the PKG are assigned with probabilities, and domain of the data is the same as domain of the PKG; and obtaining a first output result based on the data, the PKG and a language model, where the language model is configured to perform a natural language processing (NLP) task.

Abstract: A data labeling method and apparatus, a computing device, and a storage medium are provided, which belong to the field of artificial intelligence technologies. The method includes: an AI platform determines a plurality of hard examples in an unlabeled image set and hard example attributes of the plurality of hard examples, where the hard example attribute includes a hard example coefficient; the AI platform displays at least one hard example in the plurality of hard examples and a corresponding hard example attribute to a user by using a display interface; and the AI platform obtains a labeling result obtained after the user confirms the at least one hard example in the display interface based on hard example coefficients of the plurality of hard examples. According to this application, data labeling efficiency is improved, and AI model optimization efficiency is improved.

Abstract: Support for distributed database backtracking, in which the database is reverted or iteratively reverted to a previous state, is provided. Log sequence number ranges and storage offset ranges for use in reconstructing the previous state are determined and used in the reconstruction. Clock sampling records can be generated indicating a feasible backtracking window with time points indicating allowable times to which backtracking can be performed. Appropriate log records, snapshots and checkpoints for database states are identified and organized with backtracking information record and timeline. A minimal and necessary set of log records, snapshots and checkpoints enabling database reconstruction within a time window is maintained. Measures are taken to maintain the necessary data to support backtracking even after prior backtrackings have occurred in the backtracking time window.