CN111712791B

CN111712791B - Method, device and machine equipment for newly adding blocks in artificial intelligence application building

Info

Publication number: CN111712791B
Application number: CN201880002689.8A
Authority: CN
Inventors: 薛俊恩
Original assignee: Individual
Current assignee: Individual
Priority date: 2018-12-29
Filing date: 2018-12-29
Publication date: 2024-03-19
Anticipated expiration: 2038-12-29
Also published as: WO2020133322A1; CN111712791A

Abstract

A method, a device and a machine device for newly adding blocks in artificial intelligence application building. The method comprises the following steps: receiving a new chunk selection instruction of a user for chunks configured in a graphical interface building area, wherein the new chunk selection instruction acts on the chunks (310) which are mutually linked; generating a dictionary according to the chunk correspondence acted by the newly added chunk selection instruction, wherein the dictionary comprises mathematical primitive identifications corresponding to the chunks and core parameters (330); the mathematical primitive identification and the core parameter in the dictionary are packaged into a newly added mathematical primitive through the mathematical language representation combination of graph theory (350); and updating the newly added mathematical primitive to the graphical interface and the server (360). The method freely realizes new addition of the blocks for constructing the artificial intelligence application.

Description

Method, device and machine equipment for newly adding blocks in artificial intelligence application building

Technical Field

The invention relates to the technical field of Internet application, in particular to a method, a device and machine equipment for newly adding a block in artificial intelligence application construction.

Background

With the development of artificial intelligence technology, more and more artificial intelligence applications make various decisions for the obtained data based on artificial intelligence technology in many scenarios. Users have different artificial intelligence application requirements in different scenarios, and thus, often need to seek different artificial intelligence applications for different scenarios.

However, in many cases, various artificial intelligence application needs of users exist in a personalized manner, and many artificial intelligence applications already published in the internet are not applicable.

At this time, with the rise of graphical programming, people gradually consider that graphical programming may be a path meeting the personalized requirements of the artificial intelligence application, and the graphical programming for performing simple application building is evolved into the implementation of the artificial intelligence application building.

That is, the algorithm logic related to the artificial intelligence application is also configured to the corresponding chunks, and people can build the artificial intelligence application by selecting the chunks and constructing the link relation between the selected chunks. The process is set up for the artificial intelligence application realized by reference to the graphical programming of the simple application.

The artificial intelligence application necessarily involves complex algorithms, each set of blocks corresponds to a certain algorithm logic, which is configured for graphical programming of the artificial intelligence application, and is loaded onto the graphical interface as the graphical programming is initiated. However, in view of the complex algorithm and the flexible and changeable link relationships between various algorithm logics, the fixed configuration of the blocks cannot necessarily meet the needs of the artificial intelligence application construction, and the blocks capable of being selected and linked are only limited.

Therefore, how to freely realize new block addition for the construction of the artificial intelligence application, thereby meeting the free construction of the artificial intelligence application, is a difficult problem to be solved in the current urgent need.

Disclosure of Invention

In order to solve the technical problem that the new block cannot be freely added for the construction of the artificial intelligence application in the related art, the invention provides a method, a device and machine equipment for adding the new block in the construction of the artificial intelligence application.

A method of adding chunks in artificial intelligence application building, the method comprising:

the method comprises the steps of setting up blocks configured in a region for a graphical interface, and receiving a new block selection instruction of a user, wherein the new block selection instruction acts on the blocks which are mutually linked;

generating a dictionary according to the block correspondence acted by the newly added block selection instruction, wherein the dictionary comprises mathematical primitive identifiers and core parameters corresponding to the blocks;

the mathematical primitive identification and the core parameters in the dictionary are combined and packaged into newly added mathematical primitives through mathematical language representation of graph theory;

and updating the newly added mathematical primitive to the graphical interface and the server.

Receiving the update of a user on the newly added mathematical primitive to obtain digital primitive definition information, wherein the mathematical primitive definition information comprises the newly added mathematical primitive identifier, at least one mathematical primitive identifier corresponding to the newly added mathematical primitive package under the newly added mathematical primitive identifier and core parameters;

and saving definition information of the mathematical primitives to enable the graphical interface to be initialized and configured with new blocks.

An apparatus for adding chunks in artificial intelligence application building, the apparatus comprising:

the new addition selection module is used for receiving a new addition block selection instruction of a user for the blocks configured in the graphical interface construction area, wherein the new addition block selection instruction acts on the blocks which are mutually linked;

the dictionary generating module is used for generating a dictionary according to the block correspondence acted by the newly added block selection instruction, wherein the dictionary comprises mathematical primitive identifiers and core parameters corresponding to the blocks;

the mathematical representation module is used for packaging the mathematical primitive identifications and the core parameters in the dictionary into newly added mathematical primitives through mathematical language representation combination of graph theory;

and the updating module is used for updating the newly-added mathematical primitive to the graphical interface and the server.

A machine apparatus, comprising:

a processor; and

a memory having stored thereon computer readable instructions which when executed by the processor implement a method as described above.

The technical scheme provided by the embodiment of the invention can comprise the following beneficial effects:

the method is characterized in that the method is applied to the graphical construction of artificial intelligence, objects which can be controlled in the graphical construction are subjected to the graphical construction, namely, the blocks are newly increased, firstly, the blocks configured in a graphical interface construction area receive a new block selection instruction of a user, the new block selection instruction acts on the blocks which are mutually linked, then, a dictionary is correspondingly generated according to the blocks acted by the new block selection instruction, the dictionary comprises mathematical primitive identifiers and core parameters corresponding to the blocks, the mathematical primitive identifiers and the core parameters in the dictionary are combined and packaged into new mathematical primitives through mathematical language representation of graph theory, and the new mathematical primitives can be updated to a graphical interface and a server, so that the new increase of the blocks is realized freely for the construction of the artificial intelligence.

It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the invention as claimed.

Drawings

The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and together with the description, serve to explain the principles of the invention.

FIG. 1 is a schematic illustration of an implementation environment in accordance with the present invention;

FIG. 2 is a block diagram of an apparatus according to an example embodiment

FIG. 3 is a flowchart illustrating a method of adding new chunks in artificial intelligence application build, according to an example embodiment;

FIG. 4 is a flow chart depicting a procedure 330 shown in accordance with the corresponding embodiment of FIG. 3;

FIG. 5 is a flow chart depicting step 370, shown in accordance with the corresponding embodiment of FIG. 3;

FIG. 6 is a flowchart illustrating a method of adding new chunks in artificial intelligent application building, according to an example embodiment;

FIG. 7 is a flowchart illustrating a method of adding new chunks in artificial intelligent application building, according to another example embodiment;

FIG. 8 is a flowchart depicting step 650, shown in accordance with the corresponding embodiment of FIG. 7;

FIG. 9 is a diagram illustrating a DAG structure corresponding to a simple mathematical primitive, according to an exemplary embodiment;

FIG. 10 is a graphical interface schematic diagram shown in accordance with an exemplary embodiment;

FIG. 11 is a schematic diagram of a convolutional neural network operation corresponding chunk on an operation interface, according to an example embodiment;

FIG. 12 is a block distribution and linking diagram of a three-layer convolutional neural network on an operator interface, according to the corresponding embodiment of FIG. 11;

FIG. 13 is a schematic diagram illustrating interactions between front-end and back-end in accordance with an exemplary embodiment of the present invention;

FIG. 14 illustrates a schematic implementation of the overall scheme of the present invention in an exemplary embodiment;

FIG. 15 is a block diagram illustrating an apparatus for adding chunks in artificial intelligence application building, according to an example embodiment;

figure 16 is a block diagram illustrating an apparatus for newly added chunks in artificial intelligence application build implemented at the server side, according to an example embodiment.

Detailed Description

Reference will now be made in detail to exemplary embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, the same numbers in different drawings refer to the same or similar elements, unless otherwise indicated. The implementations described in the following exemplary examples do not represent all implementations consistent with the invention. Rather, they are merely examples of apparatus and methods consistent with aspects of the invention as detailed in the accompanying claims.

Fig. 1 is a schematic diagram of an implementation environment in accordance with the present invention. In one exemplary embodiment, the implementation environment includes a user terminal 110 and a server 130 configured in the background. The user terminals 110 are not limited to a single number, that is, various types of users may interact with the server 130 through the held user terminals 110 to implement the artificial intelligence application building of the present invention and the operation of the built artificial intelligence application.

The server 130 is accessed by the user terminal 110 for artificial intelligence application setup and for the operation of the set up artificial intelligence application. The server 130 is oriented to the mass user terminals 110, and any user can freely build the artificial intelligence application and use the built artificial intelligence application as long as the user can access the server 130.

Under the action of the user terminal 110 and the server 130, the threshold of the artificial intelligence application is effectively reduced, and the user can build and run the artificial intelligence application in time according to the artificial intelligence application demand generated by the user terminal.

And under the construction of the realized artificial intelligence application, the invention also provides more blocks for the construction of the artificial intelligence application, namely correspondingly provides more available artificial intelligence algorithm logic for the construction of the artificial intelligence application, and for the introduction of the common complex algorithm logic in the construction of the artificial intelligence application, a user can not select and link a plurality of blocks in a complicated way, and the blocks are newly added into the blocks corresponding to the common complex algorithm logic.

The user only needs to select the newly added block, namely, the selection and the linking of the newly added block are equivalent to those of the original blocks, effective assistance is provided for artificial intelligent application building related to a complex algorithm, and the newly added block is accurately adapted to the executed artificial intelligent application building process.

It can be appreciated that by the implementation of the invention, a platform capable of building the required artificial intelligence application at any time and adding the blocks for the application at any time is provided for the user, but the invention is not limited to the platform, and no matter the enterprise user or the end user is required to independently develop the artificial intelligence application at great cost for own requirement.

Fig. 2 is a block diagram of an apparatus according to an example embodiment. For example, the apparatus 200 may be the user terminal 110 in the implementation environment shown in fig. 1. For example, the user terminal 110 is a terminal device, various cameras, etc., held by a user such as a smart phone, a tablet computer, etc.

Referring to fig. 2, the apparatus 200 includes at least the following components: a processing component 202, a memory 204, a power component 206, a multimedia component 208, an audio component 210, a sensor component 214, and a communication component 216.

The processing component 202 generally controls overall operation of the apparatus 200, such as operations associated with display, telephone calls, data communications, camera operations, and recording operations, among others. The processing assembly 202 includes at least one or more processors 218 to execute instructions to perform all or part of the steps of the methods described below. Further, the processing component 202 includes at least one or more modules that facilitate interactions between the processing component 202 and other components. For example, the processing component 202 may include a multimedia module to facilitate interaction between the multimedia component 208 and the processing component 202.

The memory 204 is configured to store various types of data to support operations at the apparatus 200. Examples of such data include instructions for any application or method operating on the apparatus 200. The Memory 204 is implemented by at least any type of volatile or non-volatile Memory device or combination thereof, such as static random access Memory (Static Random Access Memory, SRAM), electrically erasable Programmable Read-Only Memory (Electrically Erasable Programmable Read-Only Memory, EEPROM), erasable Programmable Read-Only Memory (Erasable Programmable Read Only Memory, EPROM), programmable Read-Only Memory (PROM), read-Only Memory (ROM), magnetic Memory, flash Memory, magnetic disk, or optical disk. Also stored in memory 204 are one or more modules configured to be executed by the one or more processors 218 to perform all or part of the steps of the methods shown in any of figures 3-8, described below.

The power supply component 206 provides power to the various components of the device 200. The power components 206 include at least a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power for the device 200.

The multimedia component 208 includes a screen between the device 200 and the user that provides an output interface. In some embodiments, the screen may include a liquid crystal display (Liquid Crystal Display, LCD for short) and a touch panel. If the screen includes a touch panel, the screen may be implemented as a touch screen to receive input signals from a user. The touch panel includes one or more touch sensors to sense touches, swipes, and gestures on the touch panel. The touch sensor may sense not only the boundary of a touch or sliding action, but also the duration and pressure associated with the touch or sliding operation. The screen also includes an organic electroluminescent display (Organic Light Emitting Display, OLED for short).

The audio component 210 is configured to output and/or input audio signals. For example, the audio component 210 includes a Microphone (MIC) configured to receive external audio signals when the device 200 is in an operational mode, such as a call mode, a recording mode, and a voice recognition mode. The received audio signals may be further stored in the memory 204 or transmitted via the communication component 216. In some embodiments, audio component 210 further includes a speaker for outputting audio signals.

The sensor assembly 214 includes one or more sensors for providing status assessment of various aspects of the apparatus 200. For example, the sensor assembly 214 detects the open/closed state of the device 200, the relative positioning of the assemblies, the sensor assembly 214 also detects a change in position of the device 200 or a component of the device 200, and a change in temperature of the device 200. In some embodiments, the sensor assembly 214 further includes a magnetic sensor, a pressure sensor, or a temperature sensor.

The communication component 216 is configured to facilitate communication between the apparatus 200 and other devices in a wired or wireless manner. The device 200 accesses a WIreless network based on a communication standard, such as WiFi (WIreless-Fidelity). In one exemplary embodiment, the communication component 216 receives broadcast signals or broadcast related information from an external broadcast management system via a broadcast channel. In one exemplary embodiment, the communication component 216 further includes a near field communication (Near Field Communication, NFC) module to facilitate short range communications. For example, the NFC module may be implemented based on radio frequency identification (Radio Frequency Identification, RFID) technology, infrared data association (Infrared Data Association, irDA) technology, ultra Wideband (UWB) technology, bluetooth technology, and other technologies.

In an exemplary embodiment, the apparatus 200 is implemented by one or more application specific integrated circuits (Application Specific Integrated Circuit, abbreviated ASIC), digital signal processors, digital signal processing devices, programmable logic devices, field programmable gate arrays, controllers, microcontrollers, microprocessors or other electronic components for executing the methods described below.

FIG. 3 is a flowchart illustrating a method of adding new chunks in artificial intelligence application building, according to an example embodiment. In an exemplary embodiment, the method for adding a block in the construction of the artificial intelligence application, as shown in fig. 3, at least includes the following steps.

In step 310, for the chunks configured in the graphical interface setup area, a new chunk selection instruction of the user is received, where the new chunk selection instruction acts on the chunks that are linked to each other.

The graphical interface is a user interface for freely building an artificial intelligence application, the building of the artificial intelligence application refers to the data operation, the configuration of the sequential execution of the algorithm logic can be understood, the mathematical operation of the configuration constitutes the algorithm realization of the artificial intelligence application, and the artificial intelligence application meeting the set requirement is further realized.

Optionally, the graphical interface includes an operation component selection bar, an operation interface, a toolbar, and an operation result display area. The selection command applied to the chunk is triggered by the operation of the component selection bar. The operating component selection field includes a plurality of available chunks corresponding to the model and/or data of the initialization configuration. The user can trigger the block to generate a selection instruction through the block triggering operation, so as to complete the selection of the corresponding block, and the selected block is configured on the operation interface.

For the operation component selection bar, on the one hand, the mathematical primitives involved in selecting for the currently built artificial intelligence application may be integrated models, for example, and on the other hand, the data used for iterative training may also be selected for the built artificial intelligence application, and whatever the selection, the selection will exist in the form of chunks on the operation component selection bar of the graphical interface.

The operation interface is a building area on the graphical interface and is used for placing selected blocks for building artificial intelligence application, and the blocks are placed under the control of a user, so that building is realized. After the chunks in the operation component selection bar are triggered, the chunks are placed in the operation interface. For example, the performed selection of the chunk may drag the chunk in the operation component selection bar to the operation interface through a drag operation applied to the chunk, and the chunk dragged to the operation interface may be used for performing the construction of the artificial intelligence application.

As noted in the foregoing description, a set of blocks corresponds to a set of models or data, which is a graphical representation of the set of models or data. The set model can be a single mathematical operation or a model which is realized by integrating more than two mathematical operations, and whether the single mathematical operation or the integration of more than two mathematical operations are all a complex formed by integrating input, output and mathematical operations, and the complex is an independent unit, which is also called a mathematical primitive. Mathematical primitives correspond to the chunks by which they are configured on the operator interface according to the algorithms that the built artificial intelligence application needs to involve.

It should be understood that mathematical operations at one level are defined corresponding to the mathematical primitives of the chunk. Of course, at a more refined level, the mathematical operations defined by the mathematical primitives will be refined, with the level being different, such that the chunks correspond to several mathematical operations. The mathematical primitive will define the mathematical operation that the chunk corresponds to from the input, output, and mathematical operations that are performed. In the construction of the artificial intelligence application, with the configuration of the blocks, the configuration of mathematical primitives related to the artificial intelligence application is performed, so as to deploy mathematical operations executed in the constructed artificial intelligence application, and control the input and output of each mathematical operation for the execution thereof.

The tool bar on the graphical interface is used for adjusting corresponding data and models for the configured blocks in the construction of the artificial intelligence application, for example, fine tuning the data and the models, etc.

And an operation result display area on the graphical interface is used for displaying the result after the constructed artificial intelligence application is operated. For example, after iterative training is performed on the currently built artificial intelligence application, the corresponding iterative training situation and the available classification effect are displayed in the operation result display area.

In summary, in the graphical interface displayed on the user terminal for the execution of the artificial intelligence application building, by performing the configuration of the building area in the initially provided block, the block configured to the building area is used for the execution of the artificial intelligence application building.

Therefore, the building blocks configured in the building area, namely the building blocks selected by the current built artificial intelligence application, are corresponding to the building blocks, mathematical primitives corresponding to the selected building blocks are used for forming the current built artificial intelligence application, the building of the artificial intelligence application obtained by the view is realized for the building of the artificial intelligence application, and the degree of freedom and the self-adaptability of the built artificial intelligence application can be enhanced.

The graphical interface is initialized and configured with blocks for building artificial intelligence applications, namely, optional blocks are built for the artificial intelligence applications which are currently performed, and besides, the drag operation, the selection operation and the like applied to the selected blocks can realize building areas, namely, the new addition of the selected blocks on the operation interface referred to as the above, namely, the blocks selected by the user are configured in the building areas along with the selection instruction of the blocks on the graphical interface.

The building area is added with at least one block under the control of a user, and the process of adding the block to the building area through the control of the user can be the building process of the artificial intelligence application or the new block adding process for the subsequent artificial intelligence application building.

In the construction of the artificial intelligence application or other conditions, the blocks can be newly added, so that the blocks of the graphical interface initialization configuration are newly added according to the user requirements, and further the construction of the artificial intelligence application which is carried out later is facilitated.

In the construction of artificial intelligence applications by users, there are situations where several tiles often need to be combined together, for which a new building block can be constructed. The mathematical primitives corresponding to the newly added chunks are combinations of the mathematical primitives corresponding to the chunks.

In the building area, as the user selects at least one chunk, a new chunk addition can be initiated to the selected at least one chunk, and then a new chunk selection instruction is generated. The newly added block selection instruction is applied to at least one block selected by the user in the setting-up area.

In step 330, a dictionary is generated according to the chunk correspondence acted by the newly added chunk selection instruction, where the dictionary includes mathematical primitive identifiers and core parameters corresponding to the chunks.

After receiving a new block selection instruction triggered and generated by a user, generating a dictionary for the blocks acted by the new block selection instruction. The dictionary generated correspondingly is used for indicating more than two mathematical primitives corresponding to the newly added chunk and the link relation between the two mathematical primitives, so that the dictionary comprises mathematical primitive identifiers and core parameters corresponding to the chunks acted by the newly added chunk selection instruction.

The core parameters correspond to the data primitive, and the core parameters comprise starting parameters, input dimensions and output dimensions used by the model in the mathematical primitive. And the configuration of the link relation between the mathematical primitives is realized under the action of the input dimension and the output dimension, so that the input and the output of the mathematical primitives are controlled, and the link relation between corresponding chunks is recorded.

Dictionary generation implemented according to the newly added chunk selection instruction includes: according to the mathematical primitive corresponding to the chunk acted by the newly added chunk selection instruction, obtaining a mathematical primitive identifier and a core parameter configured for the mathematical primitive, using the mathematical primitive identifier as an index item, constructing and obtaining dictionary data corresponding to the acted chunk by using the core parameter as an index value, and the like, and constructing the dictionary corresponding to all the chunks acted by the newly added chunk selection instruction.

It should be added here that the core parameters are identified for the mathematical primitive, i.e. the core parameters are configured for the corresponding mathematical primitive. The core parameters configured to correspond to mathematical primitives will be obtained by data and model fine-tuning performed on the toolbar by the corresponding components in the build area.

The implementation of the new chunk and the implementation of the artificial intelligence application all require the generation of a dictionary for this purpose. The dictionary generated for the newly added chunk is a description of the corresponding mathematical primitive identifications and core information for the combined chunks. For the newly added chunks, and corresponding newly added math primitives, the dictionary generated is described for this purpose by data.

At least one group of blocks which are combined together to form a new added block, in the dictionary generated for the new added block, the unique mark is carried out on the data primitive mark, and the core parameters corresponding to the blocks indicate the link relation among the blocks because the core parameters comprise the input dimension and the output dimension corresponding to the data operation.

Specifically, core parameters that specify the input dimension and the output dimension for a chunk, i.e., a mathematical primitive, will be used to determine the links between the chunks. The output dimension of the mathematical primitive corresponding to one chunk indicates that the two chunks are interlinked if taken as the input dimension of the mathematical primitive corresponding to the other chunk.

Therefore, through the generated dictionary, the block adding process can accurately acquire the blocks combined together and the link relation among the blocks, and excessive storage space and more calculation cost are not required to be consumed.

In step 350, mathematical primitive identifications and core parameters in the dictionary are packaged as newly added mathematical primitives by the mathematical language representation combination of graph theory.

The dictionary is used as the data description of the newly added mathematical primitive, and the data description performed by the dictionary is expressed as the mathematical language of the graph theory according to the mathematical primitive identification and the corresponding core parameters, so that the newly added mathematical primitive formed by combining and packaging a plurality of blocks is obtained.

It should be noted that the mathematical language representation of graph theory is to track the input values, the mathematical operations performed on the input values, and the output values in the basic mathematical primitive, so as to reconstruct and characterize the mathematical primitive. Whether the input value, the mathematical operation or the output value is reconstructed in the form of nodes, and a mathematical language representation of the mathematical primitive on the graph theory is formed. Through the mathematical language representation of graph theory, the newly added mathematical primitive can be accurately and quickly reconstructed.

It can be appreciated that for the built artificial intelligence application, after the server obtains the character strings and decodes and converts the character strings back to the dictionary, the server performs reconstruction of mathematical language representation of each mathematical primitive on the graph theory based on the dictionary, so as to quickly reconstruct mathematical operations performed by the artificial intelligence application and input and output dimensions of the performed mathematical operations.

The artificial intelligence application obtained by constructing a plurality of blocks is composed of a plurality of mathematical primitives, and the newly added blocks form an executable program and are similarly constructed by at least one block, so that the reconstruction of the mathematical primitives is also based on the reconstruction of each mathematical primitive expressed in a math language on a graph theory by a dictionary.

In step 370, the graphical interface and the server are updated with the newly added mathematical primitive.

After the mathematical primitive corresponding to the new chunk is obtained through the reconstruction performed in step 350, the update on the graphical interface and the server is required. On one hand, the graphical interface can provide a plurality of blocks selected by a user with newly added blocks so that the user can add the newly added blocks to the building area, and on the other hand, the server can configure the newly added blocks in the initialization of the blocks for the graphical interface, and data related to the newly added blocks are identified in the received dictionary, so that mathematical primitives corresponding to the newly added blocks can be operated at the server side.

The update of the newly added mathematical primitive comprises the update of the image information related to the block, and in addition, the unique identification of the newly added mathematical primitive is carried out, and the mathematical primitive identification and the core parameters are stored for all the mathematical primitives forming the newly added mathematical primitive, so that the update of the newly added mathematical primitive also comprises the update of mathematical primitive definition information such as the mathematical primitive identification, the core parameters and the like.

And updating the definition information of the related mathematical primitives at the user terminal and the server, and acquiring mathematical operations executed by the newly added mathematical primitives and other information related to the mathematical operations through the updated definition information of the mathematical primitives. The updated mathematical primitive definition information describes both the newly added mathematical primitive and the execution of mathematical operations in the newly added mathematical primitive.

According to the embodiment, the function of newly added blocks is provided for the artificial intelligence application construction, so that a user can construct the personalized configuration blocks according to own application construction requirements and habits, the blocks with the original configuration exist on a graphical interface, the blocks with the user-defined configuration exist on the graphical interface, and convenience and freedom of the artificial intelligence application construction are enhanced.

In this exemplary embodiment, chunks frequently combined together for artificial intelligence application building are packaged together to form new chunks. In other words, the originally thinned mathematical operation is combined together to form a new mathematical operation, so that a new operation is built for the artificial intelligent application by the user, and the execution performance of the artificial intelligent application building is enhanced.

Fig. 4 is a flow chart depicting a procedure 330 according to the corresponding embodiment of fig. 3. In one exemplary embodiment, as shown in FIG. 4, this step 330 includes at least:

in step 331, all chunks acted by the newly added chunk selection instruction respectively acquire corresponding mathematical primitive identifiers and core parameters.

Wherein, as mentioned above, the newly added chunk is implemented corresponding to at least one chunk selected by the user, at least one digital primitive corresponding to the chunk is combined together, and a new chunk is formed for this purpose. Once the user selects the newly added block, the newly added block is equivalent to at least one corresponding block, so that convenience and efficiency of artificial intelligent application construction can be enhanced, and the artificial intelligent application can be quickly constructed.

At least one block selected for block addition in the building area is connected with each other to form a network topology capable of completing at least one mathematical operation, at this time, since the blocks are mutually linked, input dimensions and output dimensions in core parameters corresponding to the blocks are configured according to the currently constructed link relation, corresponding mathematical primitive identifications and core parameter acquisition can be performed for each block so as to generate dictionary data for each block, wherein the dictionary data indicates the link relation of the block with respect to the linked other blocks.

In step 333, dictionary data is constructed for each chunk using the corresponding mathematical primitive identifier as an index and the core parameter as an index value to generate a dictionary corresponding to the newly added chunk.

Dictionary data is built for each group of blocks through a data structure of key value pairs, wherein the data structure of the key value pairs takes mathematical primitives as index items and core parameters as index values.

Dictionary data are respectively constructed for all the chunks combined to form the newly added chunk, and a dictionary corresponding to the newly added chunk is generated and obtained by all the dictionary data.

In one exemplary embodiment, step 370 includes: and adding a new added chunk to the graphical interface according to the configured chunk name for the new added mathematical primitive so as to update the added new added chunk in a plurality of chunks configured by the initialization of the graphical interface.

For the newly added math primitive, the corresponding newly added chunk is necessarily present on the graphical interface, that is, the newly added math primitive is necessarily represented in the form of the newly added chunk on the graphical interface. In a graphical interface, the chunks, also called tiles, that characterize the mathematical primitives are the targets that users manipulate to build artificial intelligence applications.

It should be appreciated that, for ease of user identification, the newly added chunk will be configured for the graphical interface corresponding to the newly added mathematical primitive, and the newly added chunk will be marked with the corresponding chunk name to distinguish from other chunks, thereby facilitating user selection.

Fig. 5 is a flow chart depicting step 370, according to a corresponding embodiment of fig. 3. In another exemplary embodiment, as shown in FIG. 5, this step 370 includes:

in step 371, a new mathematical primitive identification is generated for the new mathematical primitive.

The newly added mathematical primitive identifiers are similar to the mathematical primitive identifiers, and are used for uniquely identifying the corresponding mathematical primitive and the corresponding block.

In core parameter configuration for a newly added mathematical primitive by a newly added chunk, the configured core parameters correspond to the generated newly added mathematical primitive identification.

In step 373, mathematical primitive identification and core parameters, which are packaged with the newly added mathematical primitive, generate mathematical primitive definition information.

The new added block is formed by at least one block, and the new added mathematical primitive corresponding to the new added block is formed by packaging and packaging at least one mathematical primitive, so that at least one group of mathematical primitive identifiers and core parameters are corresponding to the new added mathematical identifier generated for the new added mathematical primitive.

And describing the newly added mathematical primitive together with other groups of mathematical primitive identifiers and core parameters, namely generating mathematical primitive definition information for the newly added mathematical primitive.

It should be appreciated that for the chunks present in the graphical interface available for the user to choose from, there is corresponding mathematical primitive definition information, so as to provide corresponding algorithmic implementations for the artificial intelligence application built by the user based thereon.

In step 375, the mathematical primitive definition information is updated to the server, and the update of the mathematical primitive definition information at the server enables the newly added mathematical primitive to be newly added with the corresponding chunk in the initialization of the graphical interface at the server.

The existence of the mathematical primitive definition information at the server side means that the corresponding newly added mathematical primitives and the corresponding building blocks are deployed in the subsequent artificial intelligence application building, and the subsequent artificial intelligence application building for a user through a graphical interface can be performed by using the newly added building blocks, so that the newly added mathematical primitives are configured in the built artificial intelligence application.

In addition, corresponding to the method for newly adding the chunk in the artificial intelligence application construction provided by the invention, the method for realizing the chunk in the server side is realized by the method for newly adding the chunk in the server in the implementation environment shown in fig. 1. FIG. 6 is a flowchart illustrating a method of adding new chunks in artificial intelligence application build, according to an example embodiment. In an exemplary embodiment, as shown in fig. 6, the method for adding a block in the construction of the artificial intelligence application at least includes the following steps.

In step 510, the server receives the update of the newly added mathematical primitive by the user, and obtains mathematical primitive definition information to update to the server, and the update of the mathematical primitive definition information at the server enables the newly added mathematical primitive to be newly added with the corresponding chunk in the initialization of the graphical interface by the server;

In step 530, the mathematical primitive definition information is saved for the newly added mathematical primitive, so that the graphical interface is initialized to configure the newly added chunk.

The server controlled by the user terminal must receive mathematical primitive definition information sent by the user terminal just like the new chunk addition of the graphical interface of the user terminal, and at this time, the mathematical primitive definition information is saved for the user so as to be called by the user in the subsequent artificial intelligence application construction.

Of course, the newly added chunks of the user may also be shared with other users for use in the construction of the artificial intelligence application of other users, which is not limited herein.

Figure 7 is a flowchart illustrating a method of adding new chunks in artificial intelligent application building, according to another example embodiment. In an exemplary embodiment, as shown in fig. 7, the method for adding a chunk to the artificial intelligence application building further includes the following steps after performing step 530.

In step 610, a user selection of the operation associated with the newly added chunk is received, and a string corresponding to the newly added chunk is obtained.

Wherein, it should be understood that the user performs artificial intelligence application construction on the blocks on the graphical interface, including triggering operation on the blocks used in construction to execute the corresponding algorithm logic process, so as to perform trial artificial intelligence application construction.

Therefore, as the artificial intelligence application building process proceeds, after receiving the operation selection related to the newly added block by the user, the dictionary corresponding to the newly added block is converted into a character string, so that the mathematical primitive identification carried by the dictionary and the transmission of the core parameters to the server side at the user terminal are facilitated.

In one exemplary embodiment, the core parameters of the index and mathematical primitive identifications in the dictionary are converted into JSON strings, and the dictionary is transmitted to the server in the form of JSON strings.

In step 630, the character string is decoded and converted into a dictionary, which contains the newly added math primitive identifications and core parameters corresponding to the newly added chunks.

After receiving the character string sent by the user terminal to the newly added block, the character string is decoded and converted back to the dictionary form, and then the executable text is obtained by decoding through the mathematical language representation of the graph theory.

The server receives the character string sent by the user terminal. The dictionary converted to this string describes and defines the corresponding mathematical primitive for the corresponding chunk by a piece of dictionary data.

It should be understood that the mathematical primitive is an operation unit for implementing an artificial intelligence application in user-oriented implementation of the artificial intelligence application, and the operation unit can be divided at different levels according to needs, thereby configuring the mathematical primitive and corresponding chunks for this purpose.

For example, the neural network operation, the dot multiplication operation, the matrix multiplication operation and the like can be configured correspondingly as operation units according to an artificial intelligence algorithm, and some neural network operations comprise the dot multiplication operation, so that the neural network operation, the dot multiplication operation and the matrix multiplication operation can be seen to be operation unit division on different layers, but the configuration of corresponding mathematical primitives and blocks is not affected.

Therefore, the subdivided operation units can be combined together to form a new operation unit on the upper layer, namely, a mathematical primitive and a block, which is the block adding process pointed by the invention.

In step 650, the mathematical operation is reconstructed from the newly added mathematical primitive identifications and core parameters contained in the dictionary to obtain the executable text of the newly added chunk.

The executable text is a code description of the newly added chunk, i.e. a code description of mathematical operations corresponding to a plurality of chunks are combined. The execution of the executable text executes a series of operations corresponding to the newly added chunk, and the process is the running process of the newly added chunk, so that the newly added chunk is tested for the user.

And continuously reconstructing each mathematical primitive by using a DAG data structure through mathematical language representation of graph theory on newly added mathematical primitive identifiers and core parameters contained in the dictionary, constructing a link relation between the mathematical primitive and other mathematical primitives for the reconstruction, and finally forming an executable text corresponding to the newly added chunk, wherein the executable text contains all executable sentences of the newly added chunk and is composed of code information of all mathematical primitives.

The new block can execute the acquisition of text, and the corresponding code information is deployed on the server for the realization of the new block, so that the subsequent artificial intelligence application is conveniently built and called.

Fig. 8 is a flow chart depicting step 650, according to a corresponding embodiment of fig. 7. In one exemplary embodiment, this step 650 includes:

in step 651, more than two mathematical primitive identifiers are mapped according to the newly added mathematical primitive identifiers in the dictionary.

As described above, the new mathematical primitive is obtained by packing and combining more than two mathematical primitives, so that the corresponding new mathematical primitive identifier in the dictionary of the new mathematical primitive is necessarily mapped to more than two mathematical primitive identifiers.

More than two math primitive identifications mapped by the newly added math primitive correspond to each math primitive forming the newly added math primitive, respectively.

In step 653, corresponding core parameters and data missing codes are obtained according to the two or more mathematical primitive identifiers, and the core parameters are filled into the data missing codes to obtain code information of the corresponding mathematical operation.

The server stores information related to codes for mathematical operations related to the artificial intelligence application algorithm. The code-related information stored for the execution of each data operation is data missing code in which the core parameters are missing. The data missing codes are stored correspondingly with mathematical primitive identifications as indexes. The data missing code missing the core parameters realizes the configuration of the executed mathematical operation by the filling of the core parameters.

Under the action of the data missing codes stored in the server, the server provides an artificial intelligence application algorithm for the user under the support of the data missing codes, so that the user can simply and intuitively build the artificial intelligence application through the block, and on the other hand, flexible and free algorithm realization is obtained under the support of the data missing codes, and the flexibility of building the artificial intelligence application is enhanced.

The core parameters in the dictionary correspond to the mathematical primitive identifications, so that after the data missing codes are obtained by the mathematical primitive identifications, the corresponding core parameters are also obtained by the mathematical primitive identifications, and the obtained core parameters are filled in the data missing codes to obtain complete code information for executing the corresponding mathematical operations.

The code information obtained by populating the core parameters is a mathematical primitive description at the code level. Under the action of the data missing codes and the code information, the user can build artificial intelligence application according to own requirements even if the user does not have programming intelligence and programming skills.

In step 655, the executable text of the artificial intelligence application is reconstructed from the mathematical language representation of the code information sequentially passing through graph theory from output to input according to the input dimensions and output dimensions indicated in the core parameters to which the mathematical primitive identification corresponds.

As indicated in the foregoing description, the core parameters are the mathematical operations that are correspondingly performed, and the input dimensions and the output dimensions are indicated for the implementation of the artificial intelligence algorithm by the code information, so that the core parameters can interface with other mathematical primitives based thereon.

It should be noted here that the mathematical language representation of the graph theory is for mathematical primitives and also for code information for performing the corresponding mathematical operation. The mathematical language representation of graph theory is to reconstruct the input, output and mathematical operation corresponding to the mathematical primitive rapidly and accurately by means of the DAG structure, and then reconstruct the DAG structure corresponding to the mathematical primitive continuously according to the sequence from the output to the input on the basis until all the reconstruction of the mathematical primitive corresponding to the dictionary is completed.

And after all mathematical primitives corresponding to the dictionary are reconstructed to obtain mathematical language representations of the graph theory, splicing the corresponding code information according to the mathematical language representations of the graph theory to obtain the executable text of the newly added block.

It should be understood that the new addition of the block for the artificial intelligence application is necessarily performed above the artificial intelligence application building, and the implementation of the artificial intelligence application building includes the following process, namely:

(1) Receiving a selection instruction of a block on a graphical interface, wherein the block is a graphical representation of a corresponding mathematical primitive;

(2) The method comprises the steps that a building area of a graphical interface is selected to be a building configuration block of an artificial intelligence application, and the blocks are mutually linked to form an artificial intelligence algorithm configuration of the built artificial intelligence application under user control;

(3) Converting the contained blocks into a dictionary according to an artificial intelligence algorithm configuration to obtain a dictionary formed by mathematical primitive identifiers corresponding to the blocks and core parameters, wherein the core parameters are configured corresponding to the blocks;

(4) And initiating decoding of the artificial intelligent application built by the user to the server through the dictionary, and triggering the artificial intelligent application to run on the server.

For the executed process (1), the selection instruction indicates the chunk selected by the user. As described above, the chunks are graphical representations of the corresponding mathematical primitives, and therefore, the chunks may be uniquely identified by the corresponding mathematical primitive identifications, the selection instruction carries the mathematical primitive identifications, so as to identify the chunks selected by the user, and corresponding responses are performed for this purpose.

Each time a block is selected, a process for realizing the construction of the artificial intelligence application receives and obtains a selection instruction of a corresponding block on the graphical interface. And so on, as the selection of the chunks proceeds, selection instructions corresponding to different chunks are continuously received.

In one exemplary embodiment, the performed process (1) includes: and in a plurality of blocks of the graphical interface initialization configuration, receiving and obtaining a selection instruction of the blocks on the graphical interface through user operation applied to the blocks until all the blocks for realizing the artificial intelligence application are selected.

Wherein, as previously described, the graphical interface initialization configures a plurality of chunks, e.g., the presence of chunks in the operating component selection field. The user can continuously apply user operations on the required blocks according to the requirements of artificial intelligence application construction, such as drag operations on the construction area, so as to generate selection instructions for the continuously selected blocks, and the corresponding process continuously receives the generated selection instructions until all the required blocks are selected.

And (2) for the executed process, along with the receiving of the selection instruction, continuously setting up configuration blocks for the artificial intelligent application in the building area of the graphical interface, wherein the configured blocks correspond to the mathematical primitive identifiers carried in the selection instruction.

Under the action of the selection instruction, building blocks corresponding to the related mathematical primitives are configured for the artificial intelligence application in the building area of the graphical interface, so that more than two building blocks exist in the building area of the graphical interface.

It should be appreciated that as the selection of the tiles on the graphical interface proceeds, the selected tiles are scattered throughout the build area of the graphical interface. At this time, the mutual linking between the blocks is performed under the control of the user, so as to obtain the artificial intelligence algorithm configuration of the constructed artificial intelligence application.

The artificial intelligence algorithm configuration is a network topology that describes mathematical operations performed in an artificial intelligence application, in other words, a network topology formed by deployed mathematical primitives. For interlinked math primitives, the output of the last math primitive will be the input of the next math primitive, and so on, to form the overall artificial intelligence algorithm configuration.

In one exemplary embodiment, the artificial intelligence algorithm configuration is in the form of a graph, i.e., it may be in the form of an artificial intelligence algorithm configuration graph. The artificial intelligence algorithm configuration indicates the chunks contained in the constructed network topology and the linking relationships between the contained chunks.

On a graphical interface for constructing the artificial intelligence application, the selected configuration of the blocks is freely carried out for the constructed artificial intelligence application along with the operation of a user, and the link relation among the selected configuration blocks is constructed, so that the user can freely carry out the operation according to the requirements of the artificial intelligence application, the graphical programming of the artificial intelligence application construction is realized, and the threshold is reduced to the greatest extent.

For the executed process (3), the artificial intelligence algorithm configuration indicates the chunks and the link relationships between the chunks contained in the network topology of the current built artificial intelligence application, and the chunks are corresponding to mathematical primitives, so that the conversion of a chunk into a dictionary according to the artificial intelligence algorithm configuration is the conversion of a chunk into dictionary data, and the obtained dictionary data is used to form a dictionary for implementing the built artificial intelligence application.

That is, for the configuration of the artificial intelligence algorithm, dictionary data is generated for the blocks, so that the conversion from the blocks contained in the configuration of the artificial intelligence algorithm to the dictionary is realized. Dictionary data generated for the chunks is used to indicate the corresponding mathematical primitive, and input and output control is performed on this mathematical primitive.

Optionally, the configuration of the artificial intelligence algorithm is a network topology formed by the blocks, a mathematical primitive identifier and a core parameter configured for the mathematical primitive are obtained according to the mathematical primitive corresponding to the block, the mathematical primitive identifier is used as an index item, the core parameter is used as an index value to construct dictionary data corresponding to the block, and the dictionary data corresponding to all the blocks form a dictionary of the constructed artificial intelligence application.

In one exemplary embodiment, for a number of algorithms involved in artificial intelligence applications, the core parameters corresponding to the mathematical primitives include key data such as hyper-parameters, input dimensions, and output dimensions to which the model applies. The configuration of the core parameters will ensure that the corresponding mathematical primitive defines a smooth execution of the mathematical operation.

And (4) for the executed process, after the dictionary is obtained by converting the current constructed artificial intelligence application, the decoding of the constructed artificial intelligence application can be initiated to the server through the dictionary, so that the operation of the artificial intelligence application on the server is triggered.

The dictionary is obtained by constructing the artificial intelligence application by a user, and the constructed artificial intelligence application can be obtained by a server side in the form of the dictionary through mathematical primitive identification and core parameters of the contained dictionary data record.

According to the method and the system for building the artificial intelligence application, the artificial intelligence application building based on the graphical interface is carried out on the user side realized by the user terminal, namely graphical programming is carried out, after the user completes the building of the artificial intelligence application, the built artificial intelligence application is informed of the building of the artificial intelligence application by the server side by means of a dictionary generated by facing mathematical primitives corresponding to the chunks, and then the artificial intelligence application is operated on the server side, so that the artificial intelligence requirement of the user is met, and functions which can be provided by the built artificial intelligence application are obtained.

For the construction of the artificial intelligence application, under the action of the graphic interface and the blocks corresponding to the mathematical primitives, the development of complex algorithms involved in the artificial intelligence application is realized for users, namely, the required complex algorithms are constructed through the selection and interlinking of different blocks, but the users do not need to have code development capability and learn artificial intelligence expertise, only need to roughly know the function of the mathematical primitive corresponding to each block, and the threshold of the construction of the artificial intelligence application is eliminated for the masses.

The exemplary embodiment described above encapsulates the algorithms involved in the artificial intelligence application, i.e., mathematical operations and their inputs and outputs, into mathematical primitives for presentation to a user in the form of chunks, based on which the construction of the artificial intelligence application above the graphical interface is enabled.

Optionally, for the executed process (2), firstly configuring the blocks indicated by the selection instruction in the building area of the graphical interface, then acquiring the corresponding core parameters for the blocks placed in the building area, and finally, along with the configuration of more than two blocks in the building area, performing the interlinking between the blocks to form an artificial intelligence algorithm configuration of the built artificial intelligence application under the control of a user.

And along with the receiving of the selection instruction, configuring the building area with the blocks according to the received selection instruction, so that the blocks indicated by the selection instruction are added to the building area of the graphical interface.

And (3) configuring the block in a construction area, namely adding needed mathematical primitives step by step for the currently constructed artificial intelligence application, so as to finally construct an artificial intelligence algorithm configuration for realizing the artificial intelligence application, namely realizing the network topology of the mathematical primitives corresponding to the artificial intelligence application.

Along with the progress of user operation, the selection instruction is continuously triggered to be generated, so that the building blocks selected by the user can be continuously added in the building area. For each block, the configuration and adjustment of the corresponding core parameters can be performed to adapt to the currently built artificial intelligence application.

In one exemplary embodiment, for a set of blocks placed in a build area, the configuration and adjustment of core parameters in the corresponding mathematical primitive is performed, and after the configuration and adjustment of core parameters for that set of blocks is completed, the configuration and adjustment of core parameters for other blocks may be performed.

For example, core parameter configuration and adjustment of a chunk in a build area may be initiated by selection of that chunk. Specifically, after a block in the building area is selected, a toolbar on the graphical interface is used for configuring and adjusting core parameters of the block, and at this time, a user only needs to configure and adjust setting parameters on the toolbar.

The link relation transformation of the corresponding chunks not only needs to transform other chunks linked by the chunks in the building area, but also should adjust the input dimension and/or the output dimension to adapt to the mathematical primitive of the linked chunk so as to adapt to the dynamically changing artificial intelligence application building.

And obtaining the core parameters corresponding to the blocks according to the core parameter configuration of the blocks by a user for the blocks selected by the graphic interface building area.

As indicated by the foregoing description, the build area is distributed with at least one chunk, and for any chunk, the corresponding core parameter configuration process may be initiated by its selection on the build area. Once a chunk is selected at the build area, the core parameter configuration for that chunk can be performed.

The core parameter configuration is the process of inputting parameters, adjusting parameters, selecting parameters and the like under the control of a user, and the different mathematical primitives corresponding to different blocks also correspond to different core parameter configuration processes.

Regardless, core parameters corresponding to the chunks selected for the building area are obtained through the performed core parameter configuration, and core parameters corresponding to all the chunks in the building area are obtained through the similar method, namely, the core parameter configuration of all the chunks in the building area is completed, and further, the operation of the built artificial intelligence application is ensured.

For more than two blocks stored in the building area, the blocks are linked under the control of a user, so that the blocks distributed in the building area can be built into an artificial intelligence algorithm configuration of an artificial intelligence application.

The link between the blocks refers to the connection line of the blocks in the building area, and the connection line indicates the input-output relationship between the connected blocks besides the mutual connection between the blocks; on the other hand, the links between chunks also indicate that the output of the last chunk will be the input of the next chunk linked.

And so on, with the end of the added blocks in the construction area and the arrangement of links among the blocks, an artificial intelligence algorithm configuration is formed. The artificial intelligence algorithm configuration is an algorithm description of an artificial intelligence application currently built by a user. For the artificial intelligence application currently built by the user, the running process of the application is a process of executing the algorithm according to the mathematical primitives and the interlinking relations indicated by the blocks in the corresponding artificial intelligence algorithm configuration.

In one exemplary embodiment, inter-linking between chunks is achieved by wiring between chunks performed under user operation. The obtained artificial intelligence algorithm configuration is realized under the control of a user, wherein the user control refers to the block selection operation which is required to be triggered by the user when the artificial intelligence application is built, the operation of connecting lines among the blocks and the like, the operation is not limited herein, and any operation capable of configuring the blocks for the built artificial intelligence application and constructing links among the configured blocks is the user operation which is triggered to form the artificial intelligence algorithm configuration.

By means of the method and the device, the artificial intelligence algorithm configuration is built for the currently built artificial intelligence application, so that algorithm development is achieved, the artificial intelligence algorithm configuration is built at will for self needs of a user, the flexibility is improved, meanwhile, the needed artificial intelligence application building is achieved, and the method and the device are not limited by the defects of specialized knowledge such as code knowledge, artificial intelligence algorithm, mathematical representation and the like.

Optionally, for the executed process (3), firstly, for the chunks included in the artificial intelligence algorithm configuration, obtaining mathematical primitive identifiers and core parameters corresponding to the chunks; and constructing a dictionary for artificial intelligence application by taking the mathematical primitive identification as an index item and the core parameter as an index value.

The artificial intelligence algorithm configuration is an algorithm logic description of a model and data adopted in the constructed artificial intelligence application, namely an algorithm logic formed by the introduced mathematical primitives, and as indicated by the previous description, the artificial intelligence algorithm configuration is a network topology formed by a plurality of blocks.

After the user completes the addition and linking of the blocks in the building area through the manipulation of the blocks, the blocks distributed in the building area and the relation among the blocks form a block framework of the artificial intelligence application, namely the mathematical primitives used for building the artificial intelligence application and the linking relation among the mathematical primitives are indicated.

Each mathematical primitive has a corresponding code description, i.e., code information including core parameters, for implementing the execution of the corresponding mathematical operation, so that the execution of the corresponding mathematical operation is achieved by the execution of the corresponding code description. The existence of corresponding mathematical primitives in an artificial intelligence algorithm configuration made up of chunks will indicate code description information for performing a series of mathematical operations in this artificial intelligence algorithm configuration, i.e., executable text made up of code information corresponding to mathematical primitive identifications and containing core parameters.

Based on this, for the build area where chunk addition and linking are completed, the build of the artificial intelligence algorithm configuration is completed accordingly. Under the configuration of the artificial intelligence algorithm, a dictionary is generated for the constructed artificial intelligence application by acquiring mathematical primitive identifiers and core parameters corresponding to each group of blocks.

It should be appreciated that the dictionary generated is used to communicate the user-adaptively built artificial intelligence algorithm configuration to the server to obtain the artificial intelligence application running on the server.

Each chunk has a corresponding mathematical primitive, i.e. the chunk is a graphical representation of the corresponding mathematical primitive, so that for the chunks under the artificial intelligence algorithm framework, the corresponding mathematical primitive identification can be obtained, and the core parameters can be obtained along with the core parameter configuration performed on the chunks. For a set of blocks, the obtained core parameters correspond to the mathematical primitive identifications.

Dictionary data are built for each block under the artificial intelligence algorithm configuration, namely dictionary data of the obtained block are built by taking the corresponding mathematical primitive mark as an index item and taking the core parameter as an index value, and similarly, dictionary data of all blocks under the artificial intelligence algorithm configuration form a dictionary of the artificial intelligence application.

Under the action of the dictionary, the image information of the block is converted into codes existing at the server side, namely executable text of the artificial intelligence application is obtained, and the dilemma that the artificial intelligence algorithm is complex and difficult to adapt to user development is solved.

In addition, through the mode of generating the dictionary, the core parameters of the artificial intelligence application built for the user are transferred, so that the core parameters of the user personalized configuration are provided for the server, and the artificial intelligence application can be built to be accurately suitable for the artificial intelligence application requirements of the user.

Optionally, for process (4), it first performs a string conversion on the mathematical primitive identifiers in the dictionary and the indexed core parameters; and transmitting the character string to the server by the artificial intelligent application constructed by the user, and decoding the character string by the server through the transmission initiation of the character string to obtain an executable text of the artificial intelligent application so as to run on the server.

The executable text is a code description of the artificial intelligence application built by the user. The execution of the executable text executes a series of operations configured by the user for realizing the artificial intelligence application, and the process is the running process of the artificial intelligence application.

Through the exemplary embodiment, the constructed artificial intelligence application can exist on the server side, and further the functions configured in the artificial intelligence application can be fully realized through excellent hardware conditions of the server side.

Correspondingly, the server implementation built for the artificial intelligence application comprises the following execution processes:

(a) In the artificial intelligent application building selected by a user, a server receives a character string corresponding to a dictionary, wherein the dictionary corresponding to the character string is used for describing a configured block of the artificial intelligent application building;

(b) Decoding the character string to obtain an executable text of the artificial intelligence application;

(c) And through the execution of the executable text, the user selects the constructed artificial intelligent application operation server.

In order to realize the construction of the artificial intelligence application in the user terminal, the deployed server side responds to the control cooperation of the user terminal to realize the construction of the artificial intelligence application, and obtains the artificial intelligence application existing in the server side for the user.

The operation of the artificial intelligence application is realized by the execution of a series of mathematical operations, namely, the operation required by a user is realized by taking a mathematical primitive as a unit. The execution of the mathematical operation is necessarily supported by the corresponding code information, so that the execution of the mathematical operation by the server side is controlled. Therefore, the server stores the information related to the codes for the deployed chunks correspondingly, and the information is required to be respectively called under the dictionary control generated by the user for the constructed artificial intelligent application so as to obtain executable texts capable of realizing the running of the artificial intelligent application.

Based on the above, as the artificial intelligence application construction performed by the user at the user terminal is completed, the dictionary generated for the constructed artificial intelligence application is converted into a character string and sent to the server, so that the server can obtain the artificial intelligence application construction performed by the user.

And decoding the character strings obtained by dictionary conversion to obtain mathematical primitive identifications and core parameters related in the constructed artificial intelligence application, continuously reconstructing each mathematical primitive by a DAG data structure through mathematical language representation of graph theory on the basis, constructing a link relation between the mathematical primitive and other mathematical primitives, and finally forming an executable text corresponding to the constructed artificial intelligence application. The executable text contains all executable sentences of the built artificial intelligence application and is composed of code information of all mathematical primitives.

After the executable text of the artificial intelligence application is obtained, the operation of the artificial intelligence application can be triggered, and at the moment, the functions deployed by the artificial intelligence application are realized through the execution of sentences in the executable text, so that the artificial intelligence requirement of a user is met.

On the one hand, the user can quickly and freely build the artificial intelligence application through the graphical interface, and on the other hand, the built artificial intelligence application exists and operates on the server side, so that the built artificial intelligence application can obtain excellent hardware performance and strong computing capacity, and the performance of the built artificial intelligence application is enhanced.

Optionally, for the execution of process (b), it comprises: firstly, decoding and converting a character string into a dictionary, wherein the dictionary comprises mathematical primitive identifiers and core parameters corresponding to chunks configured for artificial intelligent application; and reconstructing code information for executing corresponding mathematical operations through mathematical primitive identifiers and core parameters contained in the dictionary to obtain an executable text of the artificial intelligent application.

As described above, the set of dictionary data for the mathematical primitive identifier and the core parameter indicates the corresponding mathematical primitive and other mathematical primitives linked to the mathematical primitive, i.e., the mathematical primitive whose input dimension is the output dimension in the core parameter is linked to the current mathematical primitive.

Therefore, the reconstruction of the mathematical primitive can be performed by a piece of dictionary data contained in the dictionary, the reconstructed mathematical primitive will be expressed in the form of code information in terms of program execution, and the code information is obtained by the reconstruction, which is a program language for performing mathematical operations corresponding to the mathematical primitive. By analogy, all code information constitutes the executable text of an artificial intelligence application.

For the acquisition of artificial intelligence application executable text, in one exemplary embodiment, the method comprises the steps of:

obtaining a data missing code of the corresponding mathematical operation according to the mathematical primitive identification in the dictionary;

filling core parameters corresponding to the mathematical primitive identifiers into the obtained data missing codes to obtain code information for executing the corresponding mathematical operations;

and reconstructing the executable text of the artificial intelligence application from output to input sequentially through the mathematical language representation of the graph theory according to the input dimension and the output dimension indicated in the core parameters corresponding to the mathematical primitive identification.

The server is controlled by the artificial intelligence application built by the user terminal to obtain a dictionary oriented to the built artificial intelligence application, and the dictionary records mathematical primitive identifiers and core parameters corresponding to each mathematical primitive used by the built artificial intelligence application.

In addition, the server stores information related to the codes for mathematical operations to be performed. The code-related information stored for the execution of each mathematical operation is a data missing code in which the core parameters are missing. The data missing codes are stored correspondingly with mathematical primitive identifications as indexes. The data missing codes of the core parameters are missing, so that different core parameters can be filled adaptively along with the realization of different artificial intelligence applications, and different configurations of the executed mathematical operations can be realized, thereby being fully suitable for the realization of the artificial intelligence applications.

For example, FIG. 9 is a diagram illustrating a DAG structure corresponding to a simple mathematical primitive, according to one exemplary embodiment. This simple mathematical primitive is "i+w=o", i.e., takes "I" as input, "O" as output, performs an addition operation on the input data, and "W" is the parameter resulting from the training required.

The following tables are "I" and "O" data tables configured for the block Add in the graphical interface by the user, where the block Add indicates that the corresponding mathematical primitive will perform an addition operation on the input data, and specifically are shown in the following tables, namely:

I	O
		1	7
2	8
		3	9
4	10
		5	11

TABLE 1

By this exemplary embodiment, the construction of artificial intelligence applications is enabled by means of a DAG structure, which is fast to implement even when complex artificial intelligence algorithms are involved.

Optionally, for performing the process (c), comprising: in a training mode, sample data suitable for artificial intelligence application is obtained, wherein the training mode is selected by a user on a graphical interface;

executing the artificial intelligence application at the server through the execution of the executable text, and training parameters through sample data in the operation of the artificial intelligence application.

The exemplary embodiment provides a training process of the artificial intelligence application for the artificial intelligence application built by the user, so that the artificial intelligence application built by the user can complete iterative training of parameters through the set sample parameters, and the built artificial intelligence application can be normally used.

Sample data applicable to the artificial intelligence application is configured by a user in one exemplary embodiment, for example, the user performs sample data import through a toolbar set by a graphical interface, in addition, the setting of the sample data for the constructed artificial intelligence application can be realized through a chunk related to the data on an operation interface, and the selection of the sample data can be realized through a selection frame set in the toolbar.

That is, the sample data is collected by the user, may be stored in the server side and shared by other users, or may be preconfigured by the server side according to the building needs of various artificial intelligence applications, and is not limited herein.

After the construction of the artificial intelligence application is completed, that is, all the blocks for realizing the artificial intelligence application are added in the construction area and are mutually linked, a training mode can be selected to perform iterative training of parameters on the constructed artificial intelligence application.

That is, for the artificial intelligence application that is built, the user can control the operation of the artificial intelligence application by placing the artificial intelligence application in a training mode or an operation decision mode.

For example, a training mode or a decision mode on switch can be set on the graphical interface, and a user can select to enter a corresponding mode only by manipulating the switch.

In the training mode, sample data suitable for artificial intelligence application is obtained according to sample data configuration performed by a user, for example, sample data recommended by a server is obtained, sample data shared by other users is obtained, required sample data is imported in a graphical interface by the user, and the like.

Artificial intelligence applications are implemented based on machine learning, which is an algorithm that solves specific problems for users through machine learning. For example, the machine learning algorithm may be a neural network algorithm. Thus, the constructed artificial intelligence application necessarily involves implementation of training, and parameters required for normal operation of the artificial intelligence application can be obtained through iterative training.

According to the method and the device for obtaining the sample data, the sample data obtaining entrance is achieved for the built artificial intelligence application, even if a user cannot obtain sample data with huge data volume due to various situations, the sample data can be obtained by means of the server, training and subsequent use of the built artificial intelligence application are further guaranteed, and on the basis, the user can prepare the sample data according to the self requirements, so that the accuracy of subsequent decision of the built artificial intelligence application is guaranteed.

In addition, the sample data submitted to the server by the user can be hidden, namely stored as private data of the user according to the selection of the user, but the state of the sample data can be set to be public so as to be shared at the server.

For the server, besides the pre-configured sample data and the sample data shared by the user, the sample data may be obtained in a crowdsourcing manner, which is not limited herein.

In another exemplary embodiment, for process (c), the following steps may be further included:

acquiring data selected and input by a user, wherein the data comprises target data processed by an artificial intelligence application built by the user;

and executing the executable file for completing parameter training, and running the trained artificial intelligence application in the decision mode to complete the process of selecting input data by a user.

After the artificial intelligence application is built and trained, the user uploads the data processed by the artificial intelligence application to the graphic interface, wherein the data content and the data type are different according to the difference of the artificial intelligence application, for example, the uploaded data for requesting the artificial intelligence application to process can be text data, video data, even audio data and the like

According to the embodiment, the data decision of the user by using the built artificial intelligence application is realized, and the built artificial intelligence application really meets the data processing requirement of the user.

The user selects the input data, namely the data which the user needs to process by means of the built artificial intelligence application, and it is understood that the user builds the artificial intelligence application based on the processing requirements of the data.

Similar to the training mode, the user can select the input data by only turning on the decision mode under the control of the user.

Receiving data uploaded by a user to an artificial intelligence application built by the user on a graphical interface; determining artificial intelligent application built by a user according to the user identification carried in the data; triggering and operating the determined artificial intelligent application on target data which requests the artificial intelligent application to be built in the data.

After the artificial intelligence application is built and trained, the user uploads the data which is required to be processed by the artificial intelligence application on the graphical interface, and the data content and the data type of the data are different according to the difference of the artificial intelligence application, for example, the data which is uploaded to request the artificial intelligence application to process can be text data, video data, even audio data and the like.

The server cooperates with a plurality of user terminals to realize the construction of the artificial intelligence application on each user terminal, so that the constructed artificial intelligence application of different users is different. Therefore, for the server, the dictionary is stored by taking the user identification as an index, so that the artificial intelligence applications corresponding to different users are recorded.

The artificial intelligence application of building and training is completed, and the dictionary is stored in the server side for the user to call at any time. That is, the user may call the built and trained artificial intelligence application as needed with accessing the server, and iteratively optimize parameters in the artificial intelligence application by means of decision results in use of the artificial intelligence application, thereby continuously improving accuracy of decisions.

The user identification is used for uniquely marking the user, so that the user can realize the login of the user at the server through the user identification, and further, the built artificial intelligence application is invoked. After the uploading of the data to be processed is completed, the artificial intelligence corresponding to the user identification is requested to process the uploaded data, so that the artificial intelligence application is operated by taking the uploaded data as input.

The artificial intelligence application requirement of the user is the action recognition of the short video, and the method is combined for realizing the action recognition of the short video The rows illustrate.

The user needs to realize the video action recognition through the artificial intelligence application construction of the invention. Specifically, for a short video describing an action, the identification of the action can be realized through an artificial intelligence application built by a user, and an action classification result is output.

As described in the implementation environment corresponding to fig. 1, at the user terminal, the user may drag the required chunk in the selection field of the operation component of the graphical interface, so as to drag the required chunk to the operation interface, and build a link relationship between the chunks through a connection line until the construction of the artificial intelligence algorithm configuration is completed.

The neural network operation adopted in the constructed artificial intelligence algorithm configuration is the mathematical operation adopted by the artificial intelligence application realized by the user.

FIG. 10 is a graphical interface schematic shown according to an exemplary embodiment. In an exemplary embodiment, as shown in FIG. 10, an operational component selection field 810 sets numerous operational components, i.e., chunks associated with data, models; the operation interface 830 is a stay area where the user drags the required chunk, and the data and the model are built by dragging the icon, i.e., the assembly.

In the data and model construction performed, fine tuning of the data and model will also be performed through the toolbar 850 on the right.

The construction of the artificial intelligence application is completed through the process, and the iterative training of parameters is performed. Information related to the iterative training process, such as the number of iterations, and related iterative training results, are displayed via the operation result display area 870, where the training accuracy of the data and the model is also displayed, so as to ensure that the user can grasp the real performance of the artificial intelligence.

The user builds the artificial intelligence application capable of realizing the video action video, and can try different artificial intelligence algorithm configurations through continuous chunk dragging so as to find available artificial intelligence algorithm configurations from the artificial intelligence algorithm configurations.

For example, the operation component selection bar 810 may drag the block corresponding to the convolutional neural network operation to the operation interface, to construct a first layer convolutional neural network, fig. 11 is a schematic diagram of the block corresponding to the convolutional neural network operation on the operation interface, block I represents the input, block Dense represents the convolutional neural network operation in the deep learning neural network, block O represents the output, and drag two blocks corresponding to the convolutional neural network operation to the operation interface, to construct a three-layer convolutional neural network, thereby implementing convolutional feature extraction in the constructed artificial intelligence application, and fig. 12 is a block distribution and link schematic diagram of the three-layer convolutional neural network on the operation interface, according to the corresponding embodiment of fig. 11.

Through the interaction between the user terminal and the server, the user can continuously try to build the artificial intelligence application, the effect is continuously verified, the operation is simple, and the cost is low.

By the implementation of the invention, the user is helped to quickly realize the artificial intelligence application required by the user, and the application can be called at will when required.

For a person with knowledge zero base related to artificial intelligence, to realize autonomous development of an artificial intelligence application, the experienced process is approximately: firstly, completely learning code knowledge, having a certain code development capability, then completely learning artificial intelligence expertise, finally, selecting a proper programming language as a developer according to the specific requirement, and carrying out algorithm development in a compiler, wherein the time period required by the process is at least more than 7 years.

It should be appreciated that code knowledge, deep learning algorithms and mathematical representations related to machine learning, overall knowledge of artificial intelligence, etc. are difficult to popularize and the learning threshold for codes is also difficult for the masses to surmount.

In order to overcome the difficulties, the invention provides a user-friendly interface for the public, and encapsulates a plurality of technical details in the form of blocks and mathematical primitives to realize a standardized artificial intelligent model building process, thereby truly helping the public to use artificial intelligent technology.

Artificial intelligence algorithms differ from the simple algorithmic logic involved in similar to child programming in that they are a complex natural expression that is difficult to implement with the aid of graphical user programming.

In the invention, the code package is realized through the mathematical primitive so as to realize the needed mathematical operation, and the graphic block is defined on the graphic interface to represent the mathematical operation.

FIG. 13 is a schematic diagram illustrating interactions between the front end and the back end involved in the present invention, according to an example embodiment. In this exemplary embodiment, what is implemented on the user side is a web graphical interface, that is, for the user, the required artificial intelligence application is implemented through a web application, and the running of the web application can be implemented through the browser only by knowing the access address corresponding to the background.

It is needless to say that the present invention is not limited to this, and the present invention can be realized by various means such as issuing a terminal program.

As shown in FIG. 13, by the user manipulating the graphical interface of the website, the background server will transform the blocks configured by the user, i.e. the graphics, into codes, thereby obtaining the built artificial intelligence application, and return the result, or the artificial intelligence algorithm configuration diagram, to the user by means of the cloud computing capability provided by the server. On the web graphic interface, the user constructs an AI (Artificial Intelligence ) algorithm in the mathematical language representation of graph theory, namely the mathematical primitive pointed out above, and the related algorithm information is stored in a dictionary realized by JavaScript language. On the one hand, the user selects to train the algorithm on the website graphic interface, after that, the dictionary can be submitted to a background server in the form of a JSON character string, the server decodes the JSON character string into an operable AI code language from an output end to an input end after receiving the JSON character string, and the server completes training according to the AI code language and returns the result to the website graphic interface and displays the result.

It should be noted that mathematical primitives are mathematical operations corresponding to the mathematical operations, which may include basic mathematical operations (BMO, basic Mathematical Operations) such as addition, subtraction, multiplication, division, etc., and various complex mathematical operations, such as neural network operations, etc., mathematical language expressions of mathematical primitives on graph theory, are a DAG structure for expressing mathematical operations, which is also called MNG (Mathematical Network Graph). DAG (Directed Acyclic Graph) structure for describing the application logic of the expression in the form of nodes.

Fig. 14 shows a schematic implementation of the overall scheme of the present invention in an exemplary embodiment. In one exemplary embodiment, as shown in FIG. 14, for an artificial intelligence application that is developed autonomously by a user, it includes four major phases, namely: the construction period, the send graph period, the translation dictionary as MNG period and the return result period.

The construction period is a period when a user builds the needed artificial intelligence application, and in this period, the user can build an artificial intelligence algorithm and a new MNG graphic block through the chunk, and the mathematical primitive index and the core parameter corresponding to the artificial intelligence algorithm and the new MNG graphic block are marked and stored in the dictionary.

The send graph phase is performed by the build phase. The send graph stage is the stage of sending a dictionary for the constructed artificial intelligence algorithm or new MNG tile, and the term send graph stage refers to the artificial intelligence algorithm or MNG tile for which the dictionary corresponds, both corresponding to a network topology graph.

The type conversion of the character strings received by the server is realized in the stage of sending the diagram so as to eliminate the character strings and decode and convert the character strings back to the dictionary.

The send graph stage enters the translation dictionary into the MNG stage, where the server will perform MNG, i.e., recognition of mathematical primitives, to generate code via the DAG structure, obtain an executable AI code language, train and return the results.

The trained AI code language can be used to implement decisions on the data. The decision stage is implemented on the data, still running and returning the results. In this process, after the server has completed the method construction of the artificial intelligence application, the AI code language is obtained, at which point the model will be identified and checked for whether it has been trained, and if so, the AI code language is run with the data to be run, and the result returned.

Optionally, prior to making decisions on the data, the user will also be confirmed to want by presenting an artificial intelligence algorithm configuration map to the user, on which the previously trained AI code language can be run.

The artificial intelligence algorithm configuration diagram is constructed by the block when the user builds the artificial intelligence application. The process of the obtained AI code is the process of extracting important information from the artificial intelligence algorithm configuration diagram and analyzing the important information into codes, so that the important information is packaged, namely, each data element is packaged, and net-to-gate is realized, namely, the network structure is changed into single mathematical operation. By this it will be appreciated that any mathematical operation or mathematical equation is a complex of BMOs that are linked together to form a DAG structure, expressed as MNG. Each MNG has respective input and output, and the input and output between different MNGs can be linked with each other to form a complete logic network, so that the MNGs can be used as an equivalent transition form between the block and the code, and the information corresponding to the block can be accurately transferred to the code, thereby realizing the conversion of the graphical interface and the text programming language.

It should be appreciated that each of the mathematical primitives is implemented initially, and that they are also the basis for user manipulation, and that the network structure built by the user through the chunks will be converted to a dictionary as before. The dictionary is then converted to DAG format and parsed. The network structure then becomes a mathematical primitive which is then added to the server as a new mathematical primitive. The newly formed mathematical primitive may be used as well as other primitives to build other networks. As this loop is repeated, more complex algorithms can be developed without increasing the difficulty of developing the graphics-based programming language.

The following is an embodiment of the device for executing the newly added block in the artificial intelligence application building of the present invention. For details not disclosed in the embodiment of the apparatus of the present invention, please refer to an embodiment of a method for adding a block in the application and construction of artificial intelligence of the present invention.

Figure 15 is a block diagram illustrating an apparatus for adding chunks in artificial intelligence application building, according to an example embodiment. In an exemplary embodiment, as shown in fig. 15, the apparatus for adding a chunk in the artificial intelligence application building includes:

a new selection module 1010, configured to receive a new chunk selection instruction of a user for chunks configured in a graphical interface building area, where the new chunk selection instruction acts on the chunks that are linked with each other;

the dictionary generating module 1030 is configured to generate a dictionary according to the chunk correspondence acted by the newly added chunk selection instruction, where the dictionary includes a mathematical primitive identifier corresponding to the chunk and a core parameter;

the math representation module 1050 is configured to package the math primitive identifier and the core parameter in the dictionary into a new math primitive through the math language representation combination of graph theory;

and an updating module 1070, configured to update the new mathematical primitive to the graphical interface and the server.

In response thereto, FIG. 16 is a block diagram illustrating an apparatus for newly adding chunks in the construction of an artificial intelligence application implemented at the server side according to an exemplary embodiment. In an exemplary embodiment, as shown in fig. 16, the apparatus for adding a chunk in the artificial intelligence application building includes:

the update receiving module 1110 is configured to receive an update of a newly added mathematical primitive by a user, and obtain digital primitive definition information, where the mathematical primitive definition information includes a newly added mathematical primitive identifier and at least one mathematical base identifier and a core parameter corresponding to the newly added mathematical primitive package under the newly added mathematical primitive identifier;

and the storage module 1130 is used for storing the definition information of the mathematical primitive so as to enable the graphical interface to initialize and configure the newly added block.

Optionally, the present invention further provides an electronic device, which may be used in the implementation environment shown in fig. 1, to perform all or part of the steps of the method shown in any of fig. 3 to 8. The device comprises: a processor; a memory for storing processor-executable instructions;

wherein the processor is configured to perform a method implementing the previously indicated.

The specific manner in which the processor of the apparatus in this embodiment performs the operations has been described in detail in relation to the previous embodiments and will not be described in detail here.

In an exemplary embodiment, a storage medium is also provided, which is a computer-readable storage medium, such as may be a transitory and non-transitory computer-readable storage medium including instructions. Such as memory 204 including instructions executable by processor 218 of apparatus 200 to perform the methods described above.

It is to be understood that the invention is not limited to the precise arrangements and instrumentalities shown in the drawings, which have been described above, and that various modifications and changes may be effected without departing from the scope thereof. The scope of the invention is limited only by the appended claims.

Claims

1. A method for newly adding chunks in artificial intelligence application building, the method comprising:

respectively acquiring corresponding mathematical primitive identifiers and core parameters according to all the chunks acted by the newly added chunk selection instruction, wherein the core parameters record the output dimension and the input dimension of the mathematical primitive, each chunk uses the corresponding mathematical primitive identifier as an index item, and the core parameters construct dictionary data as index values so as to generate a dictionary corresponding to the newly added chunk;

The method comprises the steps of packaging mathematical primitive identifications and core parameters in a dictionary into newly added mathematical primitives through graph theory mathematical language expression combination, wherein the graph theory mathematical language is based on the input, output and mathematical operations corresponding to the mathematical primitives, and further rebuilding the corresponding DAG structure according to the sequence from output to input on the basis, until all the mathematical primitives corresponding to the dictionary are rebuilt, the newly added mathematical primitives are DAG structures, and the DAG structures enable the newly added mathematical primitives to have at least one input dimension and a single output dimension;

updating the newly added mathematical primitive to the graphical interface and the server;

storing definition information of mathematical primitives to enable the graphical interface to be initialized and configured with new blocks;

receiving operation selection related to the newly added block by a user, and obtaining a character string corresponding to the newly added block;

Decoding and converting the character string into a dictionary, wherein the dictionary comprises newly-added mathematical primitive identifiers and core parameters corresponding to the newly-added chunks;

obtaining more than two mathematical primitive identifiers according to the mapping of the newly added mathematical primitive identifiers in the dictionary;

respectively acquiring corresponding core parameters and data missing codes according to more than two mathematical primitive identifiers, and filling the core parameters into the data missing codes to obtain code information for executing corresponding mathematical operations;

reconstructing an executable text of the newly added block according to the input dimension and the output dimension indicated in the core parameter corresponding to the mathematical primitive identifier and the mathematical language representation of the code information sequentially passing through the graph theory from output to input;

and triggering the artificial intelligent application to run at the server through the running of the execution text.

2. The method of claim 1, wherein the updating the new mathematical primitive to the graphical interface and server comprises:

and adding a new added chunk to the graphical interface for the new added mathematical primitive according to the configured chunk name so as to update the added new added chunk in a plurality of chunks configured by initializing the graphical interface.

3. The method of claim 1, wherein the updating the new mathematical primitive to the graphical interface and server comprises:

generating a new mathematical primitive identifier for the new mathematical primitive;

generating mathematical primitive definition information by using mathematical primitive identifiers and core parameters which are packaged by the newly added mathematical primitive under the newly added mathematical primitive identifiers;

and updating the mathematical primitive definition information to the server, wherein the updating of the mathematical primitive definition information at the server enables the newly added mathematical primitive to be newly added with the corresponding chunk in the initialization of the graphical interface at the server.

4. An apparatus for newly adding a chunk in artificial intelligence application building, the apparatus comprising:

the dictionary generating module is used for respectively acquiring corresponding mathematical primitive identifications and core parameters according to all the chunks acted by the newly added chunk selection instruction, and the core parameters record the output dimension and the input dimension of the mathematical primitive; each group of blocks is marked with a corresponding mathematical primitive as an index item, and dictionary data is built by taking the core parameter as an index value so as to generate a dictionary corresponding to the newly added group of blocks;

The mathematical representation module is used for carrying out the combination and encapsulation of mathematical primitive identifications and core parameters in the dictionary into newly added mathematical primitives through mathematical language representation of graph theory, wherein the graph theory mathematical language is based on the input, output and mathematical operations corresponding to the mathematical primitives, and further, the reconstruction of the corresponding DAG structure is carried out according to the sequence from the output to the input on the basis until the reconstruction of all the mathematical primitives corresponding to the dictionary is completed, the newly added mathematical primitives are DAG structures, and the DAG structures enable the newly added mathematical primitives to have at least one input dimension and a single output dimension;

the updating module is used for updating the newly-added mathematical primitives to the graphical interface and the server;

the updating receiving module is used for receiving the updating of the newly-added mathematical primitive by a user to obtain digital primitive definition information, wherein the mathematical primitive definition information comprises a newly-added mathematical primitive identifier and at least one mathematical primitive identifier and a core parameter corresponding to the newly-added mathematical primitive package under the newly-added mathematical primitive identifier;

the storage module is used for storing definition information of mathematical primitives so as to enable the graphical interface to be configured with new blocks in an initialized mode;

5. A machine apparatus, comprising:

a processor; and

a memory having stored thereon computer readable instructions which, when executed by the processor, implement the method according to any of claims 1 to 3.