CN111309821A

CN111309821A - Graph database-based task scheduling method and device and electronic equipment

Info

Publication number: CN111309821A
Application number: CN202010066812.1A
Authority: CN
Inventors: 王红宾; 吕昊; 李梦钰
Original assignee: Shanghai Yitu Network Science and Technology Co Ltd
Current assignee: Shanghai Yitu Network Science and Technology Co Ltd
Priority date: 2020-01-20
Filing date: 2020-01-20
Publication date: 2020-06-19
Anticipated expiration: 2040-01-20
Also published as: CN111309821B

Abstract

The application provides a graph database-based task scheduling method, a graph database-based task scheduling device and electronic equipment, wherein the method comprises the following steps: creating a task entity and a resource entity in a database; creating a relation entity influencing the scheduling of the task entity and the resource entity, and establishing a relation model based on the task entity, the resource entity and the relation entity; setting weight values of the task entity, the resource entity and the relationship entity; calculating the shortest weight path between the task entity and the resource entity according to the weight value to obtain a scheduling decision; and implementing scheduling decisions and updating the database. According to the graph database-based massive task scheduling method, when the annotation tasks are distributed, the task priority, the proficiency level of annotation personnel, the group annotation type, the characteristics of the annotation personnel, the characteristics of task requirements, the confidence coefficient of the annotation results and the annotation results can be considered, the annotation tasks which are most matched with the annotation personnel are distributed for the annotation personnel, and the distribution efficiency and the accuracy rate of the massive annotation tasks can be obviously improved.

Description

Graph database-based task scheduling method and device and electronic equipment

Technical Field

The present application relates to the field of scheduling of a large number of tasks, and in particular, to a graph database-based task scheduling method, device, electronic device, and computer-readable storage medium.

Background

In the prior art, a task scheduling method based on a relational database is usually adopted, and task scheduling and resource allocation are completed through SQL transactions in combination with a read-write lock mechanism or message middleware. The scheduling logic in the relational model can be represented to a certain extent by the SQL transaction and read-write lock based mechanism scheduling method, but the concurrency capability and the real-time performance are poor, factors influencing the scheduling logic are increased under the complex relational model, the calculated amount is increased, and the processing efficiency of the method is reduced along with the increase of the complexity of the relation. The task scheduling method based on the message middleware can process larger concurrency, but the scheduling logic in the complex relation model is difficult to represent, and the processing efficiency under the complex relation model is low. Such as assignment of data annotation tasks to annotating personnel (resources). Data annotation is used as the basis of artificial intelligence industry and is the starting point of the machine perception real world, data annotation is to label data, the work of data annotation is usually outsourced to a data crowdsourcing company or a data annotation platform at present, and the data crowdsourcing company or the data annotation platform arranges annotation personnel to label the received data to be annotated. The accuracy of the label marked for the data to be marked by the marking personnel has an important influence on the final identification accuracy of the machine, so that the marking accuracy of the marking personnel needs to be strictly controlled.

At present, the control of the labeling accuracy of a labeling person is usually to judge whether labeled data is labeled accurately or not after the labeling person completes a labeling task, but to distribute the data to be labeled to the labeling person averagely when the labeling person distributes the labeling task, and factors which objectively influence the accuracy of a labeling result and exist on the labeling person are not considered.

Disclosure of Invention

In view of this, the present application provides a graph database-based task scheduling method, device, electronic device, and computer-readable storage medium, which can allocate a large number of tasks to the most suitable resources according to the best matching, improve objectivity of allocation of labeling tasks, and facilitate improvement of accuracy of labeling results.

In order to solve the technical problem, the following technical scheme is adopted in the application:

in a first aspect, an embodiment of the present application provides a task scheduling method based on a database, including the following steps:

and creating a task entity and a resource entity in the database, wherein the task entity can comprise a priority attribute of the task and the like, and the resource entity can comprise a proficiency level attribute of the annotating personnel and the like.

And establishing a relation entity influencing the task entity and the resource entity scheduling, and establishing a relation model based on the task entity, the resource entity and the relation entity, wherein the relation entity influencing the task entity and the resource entity scheduling comprises a labeling group entity, a label entity, a task description file entity and a labeling result entity.

And setting weight attributes of the task entity, the resource entity and the relationship entity and weight values of relationships among the task entity, the resource entity and the relationship entity, wherein the weight attribute factors comprise task priority, skill level of a marking person, group marking type, characteristics of the marking person, task requirement characteristics, marking result confidence and efficiency.

And calculating the shortest weight path between the task entity and the resource entity according to the weight attribute and the weight value to obtain a scheduling decision.

Implementing the scheduling decision and updating the database.

According to the graph database-based task scheduling method, when the tasks are allocated to the resources, the logic factors influencing the scheduling tasks, such as task priorities, the proficiency levels of annotators, group annotation types, the characteristics of the annotators, the characteristics of task requirements, the confidence degrees of annotation results, the annotation results and other influencing factors, can be considered, the tasks which are most matched with the annotators are allocated to the resources, the annotation tasks which are most matched with the annotation tasks are allocated to the annotators, a large number of tasks can be allocated to the most suitable resources according to the best matching, and the allocation efficiency and the accuracy of the tasks are obviously improved.

As an embodiment of the first aspect of the present application, the database is a graph database, where the graph database has very efficient query performance, each object of the graph database is a node, and the relationship between the nodes is an edge, and the traversal of the unique algorithm graph of the graph data structure can be designed by using the natural extension characteristic of the graph structure, that is, from a node, its neighboring nodes can be quickly and conveniently found according to the relationship of its connection.

As an embodiment of the first aspect of the present application, in the graph database, the relationship entity may include: and one or more of a label group entity, a label entity, a task description file entity and a label result entity, wherein the relationship entity is connected with the task entity and the resource entity by at least one edge.

As an embodiment of the first aspect of the present application, the weight value of the relationship is set according to a priority degree of the relationship entity affecting the scheduling of the task entity and the resource entity.

As an embodiment of the first aspect of the present application, the weight attribute includes: the task priority, the proficiency level of the annotating personnel, the group annotation type, the characteristics of the annotating personnel, the characteristics of the task requirement, the confidence degree of the annotation result, the efficiency of the annotation result and the like. For example, the higher the task priority, the larger the weight value is set, and when assigning a task, the task priority is a first consideration element, and a task with a high task priority is preferentially assigned, and when assigning a task, the higher the level of proficiency of a annotating person, the higher the weight value is set, and when assigning a task, a tagging task with a high tagging result requirement is preferentially assigned to a tagging person with a high level of proficiency.

As an embodiment of the first aspect of the present application, the relationship model is a DAG relationship model, wherein DAG is a directed acyclic graph, whereby there is one and only one path from one node of the graph database to another, i.e. the path between the task entity and the resource entity is computed to be consistent.

In a second aspect, an embodiment of the present application provides a graph database-based task scheduling device, including:

the entity management module is used for creating a task entity and a resource entity in the database;

the relation management module is used for creating a relation entity which influences the task entity and the resource entity scheduling and establishing a relation model based on the task entity, the resource entity and the relation entity;

the weight management module is used for setting weight attributes of the task entity, the resource entity and the relationship entity and weight values of relationships among the task entity, the resource entity and the relationship entity;

the processing module is used for calculating the shortest weight path between the task entity and the resource entity according to the weight value and the weight attribute so as to obtain a scheduling decision;

the processing module implements the scheduling decision and updates the database.

According to the graph database-based massive task scheduling method, when the annotation tasks are distributed, the influence factors such as task priorities, annotation personnel proficiency levels, group annotation types, annotation personnel characteristics, task requirement characteristics, annotation result confidence degrees and annotation results can be considered, the annotation tasks which are most matched with the annotation tasks are distributed to the annotation personnel, the massive annotation tasks can be distributed to the most suitable annotation personnel according to the best matching, and the distribution efficiency and accuracy of the massive annotation tasks are obviously improved.

As an embodiment of the second aspect of the present application, the database is a graph database.

As an embodiment of the second aspect of the present application, in the graph database, the relationship entities include: one or more of a label group entity, a tag entity, a task description file entity, and a label result entity.

As an embodiment of the second aspect of the present application, the weight management module is specifically configured to:

and the weight value of the relationship is set according to the priority degree of the relationship entity influencing the scheduling of the task entity and the resource entity.

As an embodiment of the second aspect of the present application, the weight attribute includes: the task priority, the proficiency level of the annotating personnel, the group annotation type, the characteristics of the annotating personnel, the characteristics of the task requirement, the confidence degree of the annotation result and the efficiency of the annotation result.

As an embodiment of the second aspect of the present application, the relationship model is a DAG relationship model.

In a third aspect, an embodiment of the present application provides an electronic device, including: a processor; and a memory having computer program instructions stored therein,

wherein the computer program instructions, when executed by the processor, cause the processor to perform the graph database based bulk task scheduling method described above.

In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, where a computer program is stored in the computer-readable storage medium, and when the computer program is executed by a processor, the computer program causes the processor to execute the method for scheduling a large number of tasks based on a graph database.

The technical scheme of the application has at least one of the following beneficial effects:

according to the graph database-based massive task scheduling method, the logic factors influencing the scheduling tasks can be considered when the tasks are allocated to the resources, the tasks which are most matched are allocated to the resources, the massive tasks can be allocated to the most suitable resources according to the best matching, and the allocation efficiency and accuracy of the massive tasks are obviously improved.

Drawings

FIG. 1 is a flow chart of a database-based task scheduling method according to an embodiment of the present application;

FIG. 2 is a diagram of a database-based task scheduler according to an embodiment of the present application;

FIG. 3 is a diagram of a graph database model of a annotation platform according to an embodiment of the present application;

fig. 4 is a schematic structural diagram of an electronic device according to an embodiment of the present application.

Detailed Description

In order that the content of the present application may be more clearly understood, specific embodiments of the present application will be described in further detail below with reference to the accompanying drawings and examples. The following examples are intended to illustrate the present application but are not intended to limit the scope of the present application.

The graph database-based task scheduling method according to the embodiment of the present application will be described with reference to the accompanying drawings, where fig. 1 shows a flowchart of the graph database-based task scheduling method, and as shown in fig. 1, the monitoring method includes the following steps:

step S110, a task entity and a resource entity are created in the database.

Specifically, when a new task and/or a newly added annotating person is issued on the annotation platform, a corresponding task entity and/or resource entity is created in the database, and attributes, weight values and the like related to the task entity and/or resource entity are set, for example, a priority attribute and a priority weight value of the newly added task are set, which annotation group the newly issued task belongs to, and the resource entity is the attribute such as the proficiency level of the annotating person.

Step S120, a relation entity influencing the scheduling of the task entity and the resource entity is created, and a relation model based on the task entity, the resource entity and the relation entity is established.

After the newly added task entities and resource entities are created, the relationship entities affecting the scheduling of the task entities and the resource entities are associated, for example, the relationship entities affecting the scheduling of the task, such as a label group entity, are used to determine which label type the task belongs to, such as text label, picture label, video label or voice label, and the relationship entities affecting the scheduling of the task also include a task description file entity used to describe the requirements and characteristics of the task. The relation entity influencing the resource entity, such as the labeling result entity, is used for recording the confidence coefficient and the labeling efficiency of the labeling result, and the relation entity influencing the resource entity also comprises a tag entity for recording the characteristics of the labeling personnel. And establishing a relation model based on the task entity, the resource entity and the relation entity based on the real world relation model.

Step S130, setting weight attributes of the task entity, the resource entity, and the relationship entity, and weight values of relationships among the task entity, the resource entity, and the relationship entity.

After the above steps are performed, after the relationship model is defined, a weight attribute and a weight value are set for the relationship between the entity pieces according to the real world scheduling principle, where the weight attribute is used to represent the degree of influence on scheduling and logic allocation, that is, the weight attribute is set for each node and the weight value is set for the relationship (edge) between the entities, for example, the higher the task priority is, the higher the attribute value is, the higher the proficiency level of the annotating staff is, the larger the attribute value is, the larger the set weight value is, the higher the confidence level and efficiency of the annotation result are, and the larger the set attribute value is.

Step S140, calculating the shortest weight path between the task entity and the resource entity according to the weight attribute and the weight value to obtain a scheduling decision.

After determining a task entity, a resource entity and a relationship entity model influencing the scheduling of the task entity and the resource entity, calculating a shortest weight path between the task entity and the resource entity according to an attribute value of a weight attribute of the entity and a weight value of the relationship, for example, newly issuing a task, calculating to obtain a shortest weight path according to all related weight values of the priority of the task, the requirements and characteristics of labeling group tasks, and the like, allocating the task to a labeling person according to the corresponding resource entity which can be found by the obtained shortest weight path, namely the labeling person, namely completing the scheduling or allocation of the task, and scheduling according to the shortest weight path is optimal allocation, and the proficiency level of the labeling person, the confidence coefficient of a labeling result, the efficiency and the like are optimal scheduling resources of the task.

Step S150, implement scheduling decision, and update the database.

The scheduling decision is implemented, the task is distributed to the marking personnel according to the shortest weight path between the task entity and the resource entity obtained through calculation, meanwhile, the database is updated, the task priority queue is updated, the working state of the resource entity is modified, when the database is updated, the database needs to be occupied, the database is prevented from being deadlocked, if other database operations exist, the database needs to be released after the database updating operation is completed, and then the scheduling decision is carried out.

According to one embodiment of the present application, the database is a graph database. The graph database has very efficient query performance, each entity of the graph database is a node, for example, a task entity is a node, the relationship between the entities is an edge, the traversal of the unique algorithm graph of the graph data structure can be designed by using the natural extension characteristic of the graph structure, namely, from one node, the adjacent nodes can be quickly and conveniently found out according to the connection relationship.

According to one embodiment of the application, a method for scheduling a plurality of tasks based on a graph database, wherein in the graph database, a relationship entity comprises: one or more of a label group entity, a tag entity, a task description file entity, and a label result entity. The relationship entity is established through a real-world relationship model, for example, the factors affecting the resource allocation task in the real world include: marking the type of the group, the characteristics of the marking personnel, the requirements of the task and the like, and correspondingly establishing a relationship entity according to the real world relationship model.

According to one embodiment of the application, the method for scheduling a large number of tasks based on a graph database sets weight values of task entities, resource entities and relationship entities, and comprises the following steps:

and setting attributes of the task entity, the resource entity and the relationship entity which influence the scheduling, and setting a weight value according to the attributes.

According to one embodiment of the present application, the attributes may include: the task priority, the proficiency level of the annotating personnel, the group annotation type, the characteristics of the annotating personnel, the characteristics of the task requirement, the confidence degree of the annotation result and the efficiency of the annotation result. For example, the higher the task priority, the larger the weight value is set, and when assigning a task, the task priority is a first consideration element, and a task with a high task priority is preferentially assigned, and when assigning a task, the higher the level of proficiency of a annotating person, the higher the weight value is set, and when assigning a task, a tagging task with a high tagging result requirement is preferentially assigned to a tagging person with a high level of proficiency.

According to an embodiment of the present application, in the Graph database-based task scheduling method, the relationship model is a DAG (Directed Acyclic Graph) relationship model, where the DAG is a Directed Acyclic Graph, and thus there is only one path from one node of the Graph database to another node, that is, it is calculated that the paths between the task entities and the resource entities are consistent.

In the following, a database of the annotation platform is taken as an example to further describe the task scheduling method based on the database of the present application, and fig. 3 shows a diagram database model diagram of the annotation platform, as shown in fig. 3, in the diagram database,

the method comprises the nodes of tasks, labeling personnel, labeling groups, labels, task description files, labeling results and the like.

The weight attribute of the node may be set according to actual requirements, for example, the weight attribute of the task entity may be set as a high-low setting attribute value M of the task priority: the priority is 5 higher, the priority is 3 generally, and the priority is 1 lower.

The weight attribute of the annotating personnel entity can be set as a high-low setting attribute value N of the proficiency level: the proficiency level is 5 at the highest, the proficiency level is 3 at the middle, and the proficiency level is 1 at the lowest.

The weight attribute of the annotation group entity can be set as the same or different setting attribute value O of the annotation type: the label types are the same as 5, and the label types are different from 1.

The weight attribute of the tag entity can be set as attribute value P for the level of attentiveness of the annotator: care is taken up to 5, typically 3, carelessness 1.

The weight attribute of the task description file entity can be set as a task requirement Q: the task requirement is high 5, the task requirement is general 3, and the task requirement is low 1.

The weight attribute of the labeling result can be set as the attribute value R set according to the confidence level of the labeling result: the confidence of the labeling result is high by 5, the confidence of the labeling result is medium by 3, and the confidence of the labeling result is low by 1.

The relationship between the nodes is as follows:

user- [ Creation (CREATE) ] - > task

USER- [ belongs to (USER _ OF) ] - > Label group

User- [ HAS (HAS _ TAG) ] - > TAG

User- [ AnNOTAT _ IN) ] - > task

User- [ completion of annotation (ANNOTAT) ] - > annotation result

TASK- [ belongs to (TASK _ OF) ] - > Mark group

Task description FILE- [ belongs to (FILE _ OF) ] - > task

Annotating personnel- [ belong to (USER _ OF) ] - > annotating group

Annotating personnel- [ AnNOTAT _ IN) ] - > tasks

Annotating personnel- [ completion of annotation (ANNOTAT) ] - > annotating result

Annotation personnel- [ having (HAS _ TAG) ] - > TAG

Label- [ belongs to (TAG _ OF) ] - > Label group

Annotation result- [ belongs to (ANNOTAT _ OF) ] - > task description file

After the graph database model is built, the real world labeling tasks, the resources and the corresponding relation data of the tasks and the resources are imported into the graph database model, the weight attributes of the tasks and the resources are configured, the graph database model is maintained regularly, and the uniformity between the real data and the graph database model is kept.

After the graph database model is configured, in order to realize convenient and fast scheduling of a large number of tasks on the graph database model, the shortest weight path between the task nodes and the resource nodes is found, and different weight values are set for the relationship between the nodes according to the priority degree of each relationship entity influencing the task and resource scheduling.

When the weight value is set, the higher the priority degree which will affect the scheduling of the task entity and the resource entity, the higher the weight value.

For example, the relationship weight values between the following nodes are:

the weight value OF the relationship between task description FILEs- [ belonging to (FILE OF) ] - > tasks is 0.8.

TASK- [ belongs to (TASK OF) ] - > label the weight value OF the relation between groups is 1.

The weight value for the relationship between callout- [ having (HAS _ TAG) ] - > TAGs is 0.3.

The weight value of the relation between the annotation personnel- [ annotation completion (ANNOTAT) ] - > annotation results is 0.5.

And then, finding the shortest weight path between the task node and the resource node, and scheduling the task.

For example, a task is newly issued, the value OF the task weight attribute M is 5, where 5 represents the highest priority, and the attribute value M is multiplied by the weight value OF the relationship between task description FILEs- [ belonging to (FILE _ OF) ] - > tasks, i.e., M × 0.8.

According to the TASK with the highest priority, the nodes OF the label groups can be traversed, the label groups with the same label type O as the TASK are preferentially selected from high to low matching, and the label groups are multiplied by the weight value OF the relation between the TASK belonging to (TASK _ OF) and the label groups, namely O1.

And according to the task with the highest priority, further traversing the nodes of the annotators, matching from high to low, and preferentially selecting the annotators with high proficiency level N.

According to the task with the highest priority, label feature nodes traversing the relationship with the labeling personnel are further calculated, matching is carried out from high to low, the detail P of the labeling personnel is preferentially selected, and the detail P is multiplied by the weight value 0.3 of the relationship between the labels of the labeling personnel- [ having (HAS _ TAG) ] -, namely P is 0.3.

According to the task with the highest priority, the labeling result nodes are further traversed, matching is carried out from high to low, the labeling result confidence coefficient of the labeling personnel is preferably selected to be high R, and the high R is multiplied by the weight value of 0.5 of the relation between the labeling personnel- [ completion of labeling (ANNOTAT) ] - > the labeling results, namely R is 0.5.

According to the task with the highest priority, the final calculation and addition are carried out to obtain a result: and M0.8 + O1 + N + P0.3 + R0.5, and selecting the path with the largest calculated value as the shortest weight path. Namely, the task entity and the resource entity corresponding to the shortest weight path are the best scheduling decision.

The task is scheduled to the resource, the graph database is updated in real time, the scheduled task is deleted from the task priority list, and the attribute of the labeling personnel is modified into the labeling task, so that the real-time performance of task scheduling is guaranteed, and the goal of improving the labeling efficiency and quality is achieved.

Therefore, by the graph database-based massive task scheduling method, when the annotation task is distributed, the task priority, the proficiency level of the annotation personnel, the group annotation type, the characteristics of the annotation personnel, the characteristics of the task requirement, the confidence coefficient of the annotation result and the annotation result can be considered, the annotation task which is most matched with the annotation task is distributed to the annotation personnel, and the distribution efficiency and the accuracy of the massive annotation task can be obviously improved.

Based on the above description, the graph database-based bulk task scheduling device of the present application is described below with reference to specific embodiments, as shown in fig. 2 and 3, the graph database-based bulk task scheduling device of the present application includes:

an entity management module 1001 is used to create task entities and resource entities in a database.

The relationship management module 1002 is configured to create a relationship entity that affects scheduling of the task entity and the resource entity, and establish a relationship model based on the task entity, the resource entity, and the relationship entity, where the relationship entity includes one or more of a label group entity, a tag entity, a task description file entity, and a label result entity.

The weight management module 1003 is configured to set weight values of the task entity, the resource entity, and the relationship entity.

The processing module 1004 is configured to calculate a shortest weight path between the task entity and the resource entity according to the weight value, so as to obtain a scheduling decision.

The processing module 1004 implements the scheduling decisions and updates the database.

According to one embodiment of the present application, the database is a graph database.

According to one embodiment of the present application, in a graph database, relational entities include: one or more of a label group entity, a tag entity, a task description file entity, and a label result entity.

According to an embodiment of the present application, the weight management module is specifically configured to:

According to one embodiment of the application, the attributes include: the task priority, the proficiency level of the annotating personnel, the group annotation type, the characteristics of the annotating personnel, the characteristics of the task requirement, the confidence degree of the annotation result and the efficiency of the annotation result.

According to one embodiment of the application, the relationship model is a DAG relationship model.

It should be noted that, the specific working processes of each module of the graph database-based task scheduling device provided in the embodiment of the present application have been described in detail in the above method embodiment, and reference may be made to the above method embodiment specifically, and details are not repeated here.

Based on the same inventive concept as the above method, the present application also provides an electronic device, as shown in fig. 4, the electronic device includes: a processor 1401 and a memory 1402, in which memory 1402 computer program instructions are stored, wherein the computer program instructions, when executed by the processor, cause the processor 1401 to perform the steps of:

creating a task entity and a resource entity in a database;

creating a relation entity influencing the scheduling of the task entity and the resource entity, and establishing a relation model based on the task entity, the resource entity and the relation entity;

setting weight values of the task entity, the resource entity and the relationship entity;

calculating the shortest weight path between the task entity and the resource entity according to the weight value to obtain a scheduling decision;

implementing the scheduling decision and updating the database.

Further, as shown in fig. 4, the electronic apparatus further includes a network interface 1403, an input device 1404, a hard disk 1405, and a display device 1406.

The various interfaces and devices described above may be interconnected by a bus architecture. A bus architecture may be any architecture that may include any number of interconnected buses and bridges. Various circuits of one or more Central Processing Units (CPUs), represented in particular by processor 1401, and one or more memories, represented by memory 1402, are coupled together. The bus architecture may also connect various other circuits such as peripherals, voltage regulators, power management circuits, and the like. It will be appreciated that a bus architecture is used to enable communications among the components. The bus architecture includes a power bus, a control bus, and a status signal bus, in addition to a data bus, all of which are well known in the art and therefore will not be described in detail herein.

The network interface 1403 may be connected to a network (e.g., the internet, a local area network, etc.), obtain relevant data from the network, and store the relevant data in the hard disk 1405.

The input device 1404 may receive various instructions from an operator and send them to the processor 1401 for execution. The input device 1404 may include a keyboard or a pointing device (e.g., a mouse, trackball, touch pad, or touch screen, among others.

The display device 1406 may display a result obtained by the processor 1401 executing the instruction.

The memory 1402 is used for storing programs and data necessary for operating the operating system, and data such as intermediate results in the calculation process of the processor 1401.

It will be appreciated that the memory 1402 in embodiments of the invention may be either volatile memory or nonvolatile memory, or may include both volatile and nonvolatile memory. The nonvolatile memory may be a Read Only Memory (ROM), a Programmable Read Only Memory (PROM), an Erasable Programmable Read Only Memory (EPROM), an Electrically Erasable Programmable Read Only Memory (EEPROM), or a flash memory. Volatile memory can be Random Access Memory (RAM), which acts as external cache memory. The memory 1402 of the apparatus and methods described herein is intended to comprise, without being limited to, these and any other suitable types of memory.

In some embodiments, memory 1402 stores elements, executable modules or data structures, or a subset thereof, or an expanded set thereof as follows: an operating system 14021 and application programs 14014.

The operating system 14021 includes various system programs, such as a framework layer, a core library layer, a driver layer, and the like, for implementing various basic services and processing hardware-based tasks. The application 14014 includes various applications, such as a Browser (Browser), and the like, for implementing various application services. A program implementing a method according to an embodiment of the invention may be included in the application 14014.

The processor 1401 acquires the panoramic image when calling and executing the application program and data stored in the memory 1402, specifically, the application program or the instruction stored in the application 14014; preprocessing the panoramic image to obtain a subimage to be processed; inputting the sub-image to be processed into a multi-path convolution neural network to obtain a deep characteristic map of the sub-image to be processed; performing pooling treatment on the deep layer characteristic diagram; and inputting the deep characteristic map subjected to pooling into a full-connected model, and taking the output of the full-connected model as the position information after relocation.

The methods disclosed by the above-described embodiments of the present invention may be applied to the processor 1401, or may be implemented by the processor 1401. Processor 1401 may be an integrated circuit chip having signal processing capabilities. In implementation, the steps of the above method may be performed by instructions in the form of hardware integrated logic circuits or software in the processor 1401. The processor 1401 may be a general purpose processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), an off-the-shelf programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, and may implement or perform the methods, steps, and logic blocks disclosed in embodiments of the present invention. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. The steps of the method disclosed in connection with the embodiments of the present invention may be directly implemented by a hardware decoding processor, or implemented by a combination of hardware and software modules in the decoding processor. The software module may be located in ram, flash memory, rom, prom, or eprom, registers, etc. storage media as is well known in the art. The storage medium is located in a memory 1402, and a processor 1401 reads information in the memory 1402 and performs the steps of the above method in combination with hardware thereof.

It is to be understood that the embodiments described herein may be implemented in hardware, software, firmware, middleware, microcode, or any combination thereof. For a hardware implementation, the processing units may be implemented within one or more Application Specific Integrated Circuits (ASICs), Digital Signal Processors (DSPs), Digital Signal Processing Devices (DSPDs), Programmable Logic Devices (PLDs), Field Programmable Gate Arrays (FPGAs), general purpose processors, controllers, micro-controllers, microprocessors, other electronic units designed to perform the functions described herein, or a combination thereof.

For a software implementation, the techniques described herein may be implemented with modules (e.g., procedures, functions, and so on) that perform the functions described herein. The software codes may be stored in a memory and executed by a processor. The memory may be implemented within the processor or external to the processor.

In addition, an embodiment of the present invention further provides a computer-readable storage medium, where a computer program is stored, and when the computer program is executed by a processor, the processor is caused to execute the following steps:

creating a task entity and a resource entity in a database;

implementing the scheduling decision and updating the database. In the several embodiments provided in the present application, it should be understood that the disclosed method and apparatus may be implemented in other ways. For example, the above-described apparatus embodiments are merely illustrative, and for example, the division of the units is only one logical division, and other divisions may be realized in practice, for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may be in an electrical, mechanical or other form.

In addition, functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may be physically included alone, or two or more units may be integrated into one unit. The integrated unit can be realized in a form of hardware, or in a form of hardware plus a software functional unit.

The integrated unit implemented in the form of a software functional unit may be stored in a computer readable storage medium. The software functional unit is stored in a storage medium and includes several instructions to enable a computer device (which may be a personal computer, a server, or a network device) to execute some steps of the transceiving method according to various embodiments of the present invention. And the aforementioned storage medium includes: various media capable of storing program codes, such as a usb disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk, or an optical disk.

Therefore, according to the graph database-based massive task scheduling method, device and electronic equipment, task priority, the proficiency level of the annotating personnel, the group annotation type, the characteristics of the annotating personnel, the characteristics of task requirements, the confidence level of the annotation result and the annotation result can be considered when the annotation task is allocated, the annotation task which is most matched with the annotation task is allocated to the annotation personnel, real-time optimal allocation of massive tasks and resources is achieved, and the real-time response capability and the task processing efficiency of a system are improved.

Other structures and operations of a graph database based mass task scheduling method, apparatus, electronic device and computer readable storage medium according to embodiments of the present invention will be understood and readily implemented by those skilled in the art, and will not be described in detail.

The foregoing is a preferred embodiment of the present application, and it should be noted that, for those skilled in the art, several modifications and refinements can be made without departing from the principle described in the present application, and these modifications and refinements should be regarded as the protection scope of the present application.

Claims

1. A task scheduling method based on a database is characterized by comprising the following steps:

creating a task entity and a resource entity in a database;

setting weight attributes of the task entity, the resource entity and the relationship entity and weight values of relationships among the task entity, the resource entity and the relationship entity;

calculating the shortest weight path between the task entity and the resource entity according to the weight attribute and the weight value to obtain a scheduling decision;

implementing the scheduling decision and updating the database.

2. The database-based task scheduling method of claim 1, wherein the database is a graph database.

3. The database-based task scheduling method of claim 2, wherein the relational entities in the graph database comprise: one or more of a label group entity, a tag entity, a task description file entity, and a label result entity.

4. The database-based task scheduling method of claim 3, wherein the weight value of the relationship is set according to a priority degree of the relationship entity affecting the scheduling of the task entity and the resource entity.

5. The database-based task scheduling method of claim 4, wherein the weight attributes comprise: the task priority, the proficiency level of the annotating personnel, the group annotation type, the characteristics of the annotating personnel, the characteristics of the task requirement, the confidence degree of the annotation result and the efficiency of the annotation result.

6. A database-based task scheduler, comprising:

the processing module is used for calculating the shortest weight path between the task entity and the resource entity according to the weight attribute and the weight value so as to obtain a scheduling decision;

7. The database-based task scheduler of claim 6, wherein the database is a graph database.

8. The database-based task scheduler of claim 7, wherein the relational entity in the graph database comprises: one or more of a label group entity, a tag entity, a task description file entity, and a label result entity.

9. The database-based task scheduling device of claim 8, wherein the weight management module is specifically configured to:

10. The database-based task scheduler of claim 9, wherein the weight attributes comprise: the task priority, the proficiency level of the annotating personnel, the group annotation type, the characteristics of the annotating personnel, the characteristics of the task requirement, the confidence degree of the annotation result and the efficiency of the annotation result.