26 results about "Task sharing" patented technology

The invention provides a task execution method and device based on a large model, an intelligent agent and electronic equipment, and relates to the technical field of artificial intelligence, in particular to the technical fields of voice interaction, deep learning, large models and the like. The method comprises the following steps: acquiring a demand feature representing a demand intention; a target processing task is executed according to the demand features by utilizing a large model, a reply text matched with the demand intention is obtained, and target reply words in the reply text are determined on the basis of the following modes: respective query features of a plurality of attention sub-tasks in the target processing task are determined on the basis of associated reply word features, the associated reply character features are determined based on associated reply characters in the reply text; based on the demand feature read from the storage unit as a value feature and a key feature shared by the plurality of attention sub-tasks, the plurality of attention sub-tasks are executed by the computing unit according to the plurality of query features, the value feature and the key feature.

The invention provides a scheduling method and device of a multi-task shared processor and a cluster. The method comprises the steps of obtaining a first instruction of a first task in a plurality of tasks in a first time slice of a plurality of time slices; confirming a second task in the plurality of tasks corresponding to the first time slice; when the first sum of the required computing power of the current execution instruction of the processor in the first time slice and the required computing power of the first instruction is smaller than or equal to the maximum computing power of the processor, the first instruction is sent to the processor in the first time slice, so that the processor executes the first instruction in the first time slice; wherein the current execution instruction comprises an instruction of the second task. The method can improve the computing power utilization rate of the processor.

The invention discloses an AIGC model resource scheduling system and method based on task sharing. The AIGC model resource scheduling system based on task sharing comprises an information acquisition module, a cluster resource management module and a model scheduling module. The information acquisition module is used for acquiring model resource demand information; the cluster resource management module is used for acquiring server resource information of equipment in a cluster; the model scheduling module is used for generating a scheduling equipment sequence queue based on the model resource demand information and the server resource information, and matching and deploying a plurality of AIGC models to corresponding computing cards; the objective is to use up computing power and video memory resources of equipment as much as possible. Therefore, the AIGC model resource scheduling system based on task sharing can reasonably allocate the models with different resource demands to the same computing card, and realizes efficient utilization of computing power and video memory resources while ensuring that the tasks meet the service level target.

The embodiment of the invention discloses a task scheduling method and device for a computing power network and electronic equipment, belongs to the technical field of computer networks, and can improve the flexibility of task scheduling in the computing power network. Comprising the steps of obtaining task data of a plurality of tasks related to computing power network scheduling, inputting a trained task allocation model, outputting loss values of the tasks, and matching a forgetting gate, an input gate and an output hidden layer feature when target equipment is matched; updating the forgetting door through a parameter updating module of the task allocation model to obtain an updated forgetting door; updating the input gate to obtain an updated input gate; according to the historical cell state, updating a forgetting gate, updating an input gate and a loss value, outputting a new cell state of the task, and determining a meta-strategy parameter of each task based on the new cell state; and inputting the meta-strategy parameter of each task into the task allocation model, outputting a new strategy parameter shared by the task, and performing task scheduling on all the tasks based on the new strategy parameter.

This application provides a method, device, and storage medium for dimension table joins in Flink, relating to the field of big data technology. The method identifies a target subtask among multiple subtasks of a unified dimension table join operator in the current task manager. After controlling the target subtask to send a query request to an external database and obtain the dimension table data sent by the external database, the method controls the target subtask to load the dimension table data into a static dimension table object set in the metadata space of the task manager. This avoids putting significant operational pressure on the external database in a short period. The method controls the target subtask and multiple other subtasks to perform dimension table join operations based on the obtained raw data and the dimension table data in the static dimension table object, allowing all subtasks in the same task manager to share the same dimension table data, thereby saving physical memory resources.

This application provides an interactive method, apparatus, electronic device, and storage medium based on task sharing. The method includes: acquiring a first task corresponding to a first virtual character in a game; sharing the first task with at least one second virtual character based on the task sharing rules of the first virtual character; and determining the interactive relationship between the first virtual character and the second virtual character based on the execution result of the second virtual character on the first task. This approach solves the technical problems of low social efficiency and high cost in existing game social modes.

The invention relates to the field of embedded systems, in particular to a multi-task shared resource negotiation using method under a single-core embedded processor platform. According to the method, a resource permission management module, a resource permission application module and a resource permission release module are designed, the resource permission management module is used for initializing a shared resource negotiation function, the resource permission application module is used for enabling a task to obtain shared resource use permission, and the resource permission release module is used for enabling the task to release the shared resource use permission. A section of DDR address space is occupied to record shared resource management information, and a timer hardware peripheral of a processor system is called to deal with the overtime abnormal problem. According to the method, management of negotiation conflict resources among the tasks is realized by calling function codes in the tasks, and exception processing is performed by using the timer hardware peripheral of the processor system, so that the sequence of accessing the same resource by a plurality of tasks is ensured, and behavior conflicts among the tasks are avoided.

The embodiment of the invention provides a storage resource allocation method for a chip, the chip, equipment and a program product, the chip is provided with a storage space shared by multiple tasks, and the method comprises the steps that in response to a received to-be-processed target task, storage demand information of the target task is acquired; comparing the storage demand information with the space configuration information of the storage space to obtain a comparison result; the space configuration information comprises a current distributable space corresponding to at least one task type; under the condition that the comparison result represents that the current distributable space corresponding to the target task type of the target task meets the data storage demand corresponding to the storage demand information, determining the current distributable space corresponding to the target task type as a target distributable space; and distributing a data storage area for each task thread corresponding to the target task in the target distributable space. In this way, the parallelism degree of task processing can be improved, and therefore the overall running performance of the chip is improved.

The invention provides a distributed computing system, method and device based on a hybrid cluster and a storage medium, the distributed computing system based on the hybrid cluster comprises a cluster manager used for distributing a stream computing topology configuration file to a task manager; the cluster coordinator is used for collecting the states of the task process and the task manager, sending the state of the task process to the task manager and sending the state of the task manager to the cluster manager; the task manager is used for creating a task process based on the flow calculation topology configuration file, establishing a flow calculation topology based on the task process and managing the task process based on the state of the task process; wherein the task process serves as a node of the stream computing topology, and the task process is used for executing distributed stream computing and runs in an independent process mode. According to the method, the task process is limited to run in an independent process mode, resource scrambling and logic pollution caused by multi-service task sharing processes are avoided, and the stability of the whole computing cluster is improved.

The embodiment of the invention provides a multi-device task processing method and device, equipment and a medium, and the method comprises the steps: obtaining a shared task of a user group, and the shared task comprises a task submitted by at least one user in the user group; determining at least one target user in the user group; determining equipment of at least one target user and acquiring the processing capability of the equipment; according to the method, the task is assigned to the equipment of at least one target user in the shared task according to the processing capability of the equipment, so that efficient organization, accurate execution, reasonable utilization of equipment resources and intelligent allocation of task processing are realized, the cooperation requirements of the users can be better met in the environment of the Internet of Things, and the efficiency and quality of task processing are improved.

The invention provides a data coding method and system, electronic equipment and a computer program product, and relates to the technical field of data processing. The method comprises the following steps: determining whether a coding terminal has a task sharing demand or not according to state parameters of the coding terminal; if it is judged that the coding terminal has the task sharing requirement, coding tasks of the coding terminal are classified into local tasks and remote tasks; performing coding processing based on the local task to obtain local data; sending the remote task to a cloud in communication connection with the coding terminal, and receiving remote data sent by the cloud based on the remote task; and carrying out assembly processing on the local data and the remote data to obtain coded data. Whether a task sharing requirement exists or not can be judged based on the actual state of the coding terminal, so that coding tasks are classified when the tasks need to be shared, the coding terminal executes local tasks, a cloud end executes remote tasks, data obtained through coding of equipment at the two ends are assembled, and final coding data are obtained.

A method and apparatus for information processing in a game, and an electronic device. The method comprises: generating a target component in response to a task creation operation, the target component comprising at least one task entry (S102); and in response to a sharing operation for the target component, sharing the target component to a designated system of a game, so as to add, by triggering the target component in the designated system, the task entry corresponding to the target component to a task system of the game (S104). In the method, a player can perform task creation according to requirements, and generate the target component comprising the task entry, thereby improving the efficiency of task creation. In addition, the player can share the target component to the designated system of the game, so that the player can import the target component into the task system of the game by one click. In the method, the target component is used for task sharing and circulation, which helps other players to view, import, and perform execution, thereby improving the game experience of players.

The embodiment of the invention discloses a transaction processing method, electronic equipment, a storage medium and a program product, and relates to the technical field of financial science and technology. The method comprises the following steps: receiving a batch processing state of an accounting batch processing task sent by a batch detection thread by utilizing a transaction processing thread, determining processable time of an online transaction task from a daily switching time interval based on the batch processing state, and sharing an accounting table by the accounting batch processing task and the online transaction task; dequeuing the online transaction tasks in the to-be-processed asynchronous task queue within the processable time by using the transaction processing thread to obtain a to-be-processed transaction task, and executing the to-be-processed transaction task, the to-be-processed asynchronous task queue comprising at least one online transaction task received within the daily switching time interval; the accounting table is updated based on the execution result of the transaction task to be processed by using the data synchronization thread, so that a bank core business system can provide 7 * 24-hour uninterrupted service for a customer, and the user experience is improved.

This disclosure provides a task execution method, apparatus, intelligent agent, and electronic device based on a large model, relating to the field of artificial intelligence technology, particularly to the fields of voice interaction, deep learning, and large models. The method includes: acquiring demand features representing demand intent; using a large model to execute a target processing task based on the demand features to obtain a response text matching the demand intent, wherein the target response word in the response text is determined based on: determining query features for each of the multiple attention subtasks in the target processing task based on associated response word features, and the associated response word features are determined based on associated response words in the response text; using the demand features read from a storage unit as value features and key features shared by the multiple attention subtasks, and using a computing unit to execute multiple attention subtasks based on the multiple query features, value features, and key features.

To reduce the computing load of edge servers and improve the system response speed, a task sharing and offloading method based on deep reinforcement learning is proposed. It includes: (1) decomposing in-vehicle tasks into relatively independent task units and assigning a complete task unit ID field; (2) caching all offloaded computing task units in the ID pool of the VEC server, and the ID pool updates cached tasks according to the effective time of task units and the least recently used principle; (3) establishing a matching mechanism between newly created task units and cached task units, and only task units that match successfully can share task calculation results; (4) establishing a task offloading utility function and its optimization model for computing latency, energy consumption, and computing cost, and equivalent the overall optimal task offloading scheme to a Markov process; (5) completing the selection of the optimal offloading scheme through deep reinforcement learning methods based on the task unit offloading state space and policy space.

This invention discloses an intelligent control method for collaborative autonomous linkage inspection of unmanned equipment in a dedicated power grid. The method includes: opening a task sharing channel for each inspection party, through which the inspection parties share shared inspection tasks in real time; marking the shared inspection tasks in a power grid visualization information model; sharing the power grid visualization information model with each inspection party; identifying the inspection tasks dispatched by the inspection parties and marking them accordingly in the power grid visualization information model; determining the timeliness requirements of each inspection task and the shared inspection task in the power grid visualization information model; marking the inspection tasks and the shared inspection tasks with corresponding classification tags according to the timeliness requirements; performing real-time collaborative inspection analysis of the unmanned inspection equipment based on the classification tags to obtain the inspection control mode, and sharing task completion data in real time with each unmanned inspection equipment of the inspection parties during the inspection process.

The application provides a model initialization method based on cross-task shared expert inheritance, constructs a multi-task mixed expert ancestor model and trains; each task data is input into the ancestor model, and the selection probability of each expert is counted according to layers; the importance score is constructed based on the cross-task average selection probability and the entropy of the expert, and the shared expert is selected according to layers to form a learning gene; each layer of the shared expert is used as a basic expert, a derived expert set is generated through linear expansion to initialize the mixed expert module of the corresponding layer of the offspring model; in the training or fine-tuning of the offspring model, Top-K sparse routing is used to activate the expert and output; when the multi-task training is performed, the task-expert mutual information loss and the task loss are introduced to form a total loss and update the model parameters, and the convergence speed and the generalization performance of the application in unknown downstream tasks can be improved.

The invention relates to a multi-granularity task adaptive sharing multi-task collaborative adaptation method, which belongs to the field of multi-task learning, and comprises the following steps: establishing an information combination tree to establish multi-granularity task sharing feature information, and designing a task attention module to extract feature information beneficial to a specific task from the tree; the problems of difficult optimization and unstable performance caused by large difference between tasks in existing multi-task learning are solved, and the negative transfer phenomenon is effectively relieved. The multi-task collaborative adaptation method comprises the following steps: S1, constructing a multi-granularity information combination tree shared by visual tasks; s2, extracting combination task feature information on different granularity levels of the multi-granularity information combination tree, and performing intra-layer fusion and inter-layer fusion to obtain intra-layer soft prompts and cross-layer soft prompts; and S3, obtaining a final soft prompt according to the in-layer soft prompt and the cross-layer soft prompt, splicing the final soft prompt to the input of the visual pre-training model, and obtaining a prediction result on the corresponding visual task.

An object of a system according to an embodiment is to propose optimal task sharing on the basis of a skill set and task content of an employee and to realize efficient task execution.SOLUTION: A system includes a reception unit, a proposal unit, a monitoring unit, and a readjustment unit. The reception part inputs the work contents and skill set of the employee. The proposal unit analyzes the information input by the reception unit and proposes task sharing. The monitoring unit monitors the progress status of the task in real time based on the task sharing proposed by the proposal unit. The readjustment unit readjusts the task sharing based on the progress status monitored by the monitoring unit.SELECTED DRAWING: Figure 1

The embodiment of the application discloses a data shunting processing system, method and device, and belongs to the technical field of data processing. In the embodiment of the application, a first code reader with weak hardware performance can shunt image data to be processed of the first code reader to a second code reader with strong hardware performance for processing. In this way, a data channel between the first code reader with weak hardware performance and the second code reader with strong hardware performance is opened. In this case, when the processing capacity of the first code reader with weak hardware performance is insufficient, the second code reader with strong hardware performance can be used for task sharing. Especially when the first code reader with weak hardware performance has a processing fault, as long as the communication function of the first code reader is normal, the data can be shunted to the second code reader with strong hardware performance for processing, so that the data is not discarded.

The invention provides a multi-task scene-oriented model reasoning method and device, equipment, a medium and a product. The method comprises the following steps: acquiring multi-field data; and inputting the multi-field data into the multi-task learning model to obtain a multi-dimensional model reasoning result output by the multi-task learning model. According to the multi-task scene-oriented model reasoning method provided by the invention, explicit separation is carried out from a feature level into two branches of task sharing feature extraction and task specific feature extraction, the extraction of task sharing features by task specific features is effectively isolated, the accuracy of feature extraction is effectively improved by the decoupling mechanism, and in addition, the extraction efficiency is improved. The task specific feature extraction is realized by reading the task sharing feature from the cache, so that the task sharing feature can be efficiently reused among multiple tasks, the condition of repeated calculation is greatly reduced, the calculation cost is reduced, the reasoning speed is accelerated, and therefore, the calculation efficiency is improved. The improved multi-task learning model can realize efficient and high-precision multi-task collaborative reasoning on multi-field data.

The invention discloses a multi-task scheduling system and a multi-task scheduling method for a heterogeneous GPU (Graphic Processing Unit) computing system. The system comprises a user front end, an energy perception task scheduling strategy system and a task sharing system. An energy consumption model is accurately described by considering a degradation effect, dynamic power consumption, idle power consumption and actual execution time brought by concurrent task execution. By optimizing task allocation, a high-load task is prevented from running on low-power equipment for a long time, or invalid energy consumption of a low-load task on a high-performance GPU is avoided, so that the performance is ensured, the energy consumption is remarkably reduced, and task execution and energy efficiency are balanced.

The invention discloses a data processing method and device and electronic equipment. The method comprises the steps of displaying a to-be-selected first task associated with a target category; tasks in the to-be-selected first tasks can be selected to be executed by a user, and rights and interests corresponding to the tasks are triggered to be issued to the user after the tasks are selected to be executed; obtaining first task execution data of a user for a target task in the to-be-selected first tasks; displaying task sharing information associated with the target task in response to the situation that the first task execution data meets a successful completion condition of the target task; after a user triggers sharing of the task sharing information, a right and interest issuing request is sent to a server side so that target rights and interests corresponding to the target task can be issued to the user, and the target rights and interests can be cancelled after verification in a transaction link of a category object associated with the target task in the target category. By adopting the method, the problem that the interestingness and the propagation degree of the existing rights and interests obtaining mode for the categories are not high is solved.

The application relates to a content sharing method and device, computer equipment and a storage medium. The method comprises the following steps: in a first instant messaging application corresponding to a first user identifier, task completion content for a target task sent by a terminal corresponding to a task user identifier is displayed; content evaluation information corresponding to the task completion content is displayed; when the content evaluation information meets a sharing condition, in response to a content sharing operation on the task completion content, the task completion content and the content evaluation information are shared to a terminal corresponding to a second user identifier which has a task association relationship with the task user identifier, and the second user identifier is a user identifier corresponding to a second instant messaging application. According to the above scheme, the task completion content is displayed in the first instant messaging application, and the shared user can view the task completion content and the content evaluation information in the second instant messaging application, so that the efficiency of task display and task sharing is improved.

The invention relates to a task processing method and device, a base station and computer equipment, and belongs to the technical field of communication. The method is applied to a base station of a multi-core processor, the base station is provided with at least one task sharing queue and a plurality of task exclusive queues, the task sharing queues are used for storing parallel tasks of all cells, one task exclusive queue is used for storing serial tasks of one cell, and for each processor, the task sharing queues and the task exclusive queues are distributed in parallel. Traversing at least one task sharing queue and a plurality of task exclusive queues; under the condition that the currently traversed task queue is the task exclusive queue, locking the currently traversed target task exclusive queue, executing the serial tasks in the target task exclusive queue, and unlocking the target task exclusive queue after the serial tasks in the target task exclusive queue are executed; and under the condition that the currently traversed task queue is the task sharing queue, executing the parallel tasks in the task sharing queue. According to the invention, the processing performance of the CPU can be improved.

The invention discloses an intelligent control method for cooperative autonomous linkage inspection of unmanned equipment of a power private network, and the method comprises the steps: opening a task sharing channel for each inspection party, and enabling the inspection parties to share a shared inspection task for inspection in real time through the task sharing channel; marking in the power grid visual information model according to the shared inspection task; sharing the power grid visual information model to each inspection party; identifying an inspection task dispatched by an inspection party, and correspondingly marking the inspection task in the power grid visual information model; determining timeliness requirements of each inspection task and the shared inspection task in the power grid visual information model; marking corresponding classification labels for the inspection tasks and the shared inspection tasks according to timeliness requirements; and performing real-time collaborative inspection analysis on the unmanned inspection equipment according to the classification labels to obtain an inspection control mode, and performing real-time task completion data sharing with each unmanned inspection equipment of the inspection party in the inspection process.