WO2022037627A1

WO2022037627A1 - Integrated management method, apparatus and system for artificial intelligence applications

Info

Publication number: WO2022037627A1
Application number: PCT/CN2021/113426
Authority: WO
Inventors: 南雨含; 黄缨宁; 孔维; 王昱森; 张钰
Original assignee: 第四范式（北京）技术有限公司
Priority date: 2020-08-20
Filing date: 2021-08-19
Publication date: 2022-02-24
Also published as: CN114077437A

Abstract

An integrated management method for artificial intelligence applications. The method comprises: acquiring an installation file, in a predetermined standard format, of a first artificial intelligence application, and installing the first artificial intelligence application on the basis of the installation file in the predetermined standard format; acquiring construction configuration information of a second artificial intelligence application, and constructing the second artificial intelligence application on the basis of the construction configuration information; and managing the installed first artificial intelligence application and the constructed second artificial intelligence application.

Description

An integrated management method, device and system for artificial intelligence applications

This disclosure claims the priority of the Chinese patent application with the application number 202010844690.4 and the application title "An integrated management method, device and system for artificial intelligence applications", which was filed with the China Patent Office on August 20, 2020, the entire contents of which are approved by References are incorporated in this disclosure.

technical field

The present disclosure relates to the field of artificial intelligence, and more particularly, to an integrated management method for artificial intelligence applications, an integrated management device for artificial intelligence applications, a system including at least one computing device and at least one storage device, and A computer-readable storage medium.

Background technique

With the development and popularization of artificial intelligence (AI, Artificial Intelligence) technology, more and more industry scenarios need to apply artificial intelligence applications. At present, due to the diverse business problems of enterprises and other users, most of the solutions on the market can only solve problems in specific fields, and most AI applications are highly customized for specific business problems. Subject to the difficulty of the user's own business problems, human efficiency and other issues, usually when the user is faced with the situation that the customized AI application provided by a single manufacturer cannot solve all the problems, he will choose to purchase AI tools and software from multiple manufacturers to meet the business needs. . When users' demands for software are booming, the data of various AI applications are not interoperable, the standards are different, and the operation and maintenance are complicated. The development and management of various AI applications will bring huge management costs to users, which is very unfavorable for enterprises and other users. AI intelligent transformation.

SUMMARY OF THE INVENTION

An object of the present disclosure is to provide a new technical solution for unified management of AI applications.

According to a first aspect of the present disclosure, an integrated management method for AI applications is provided, including:

Obtain an installation file in a predetermined standard format of the first AI application, and install the first AI application based on the installation file in the predetermined standard format;

obtaining construction configuration information of the second AI application, and constructing the second AI application based on the construction configuration information;

The installed first AI application and the constructed second AI application are managed.

According to a second aspect of the present disclosure, an integrated management device for AI applications is provided, including:

a first installation module, configured to obtain an installation file in a predetermined standard format of the first AI application, and install the first AI application based on the installation file in the predetermined standard format;

a second building module, configured to obtain the building configuration information of the second AI application, and build the second AI application based on the building configuration information;

The application management module is configured to manage the installed first AI application and the constructed second AI application.

According to a third aspect of the present disclosure, there is provided an artificial intelligence application integrated management system including at least one computing device and at least one storage device, wherein the at least one storage device is configured to store instructions, the instructions are When the at least one computing device operates, the at least one computing device is caused to perform the method according to the first aspect of the present disclosure.

According to a fourth aspect of the present disclosure, there is provided a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements the method according to the first aspect of the present disclosure.

Through the embodiments of the present disclosure, based on the integrated management method of AI applications, the first AI application installed and the second AI application constructed in the AI operating system can be managed in a unified manner, so that the development and management costs of AI applications can be reduced.

Other features of the present disclosure and advantages thereof will become apparent from the following detailed description of exemplary embodiments of the present disclosure with reference to the accompanying drawings.

Description of drawings

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

1 is a block diagram of one example of a hardware configuration of an electronic device that can be used to implement embodiments of the present disclosure;

2 is a schematic flowchart of an integrated management method for AI applications according to an embodiment of the present disclosure;

3 is a schematic diagram of a standard development process of a second AI application according to an embodiment of the present description;

4 is a block schematic diagram of an integrated management apparatus for AI applications according to an embodiment of the present disclosure;

5 is a block schematic diagram of a system according to an embodiment of the present disclosure.

detailed description

Various exemplary embodiments of the present disclosure will now be described in detail with reference to the accompanying drawings. It should be noted that the relative arrangement of the components and steps, the numerical expressions and numerical values set forth in these embodiments do not limit the scope of the present disclosure unless specifically stated otherwise.

The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the disclosure, its application or uses in any way.

Techniques, methods, and apparatus known to those of ordinary skill in the relevant art may not be discussed in detail, but where appropriate, such techniques, methods, and apparatus should be considered part of the specification.

In all examples shown and discussed herein, any specific values should be construed as illustrative only and not limiting. Accordingly, other instances of the exemplary embodiment may have different values.

It should be noted that like numerals and letters refer to like items in the following figures, so once an item is defined in one figure, it does not require further discussion in subsequent figures.

Hereinafter, various embodiments and examples according to embodiments of the present disclosure will be described with reference to the accompanying drawings.

FIG. 1 is a block diagram showing a hardware configuration of an electronic device 1000 that can implement an embodiment of the present disclosure.

The electronic device 1000 may be a laptop computer, a desktop computer, a cell phone, a tablet computer, or the like. As shown in FIG. 1 , the electronic device 1000 may include a processor 1100, a memory 1200, an interface device 1300, a communication device 1400, a display device 1500, an input device 1600, a speaker 1700, a microphone 1800, and the like. The processor 1100 may be a central processing unit CPU, a microprocessor MCU, or the like. The memory 1200 includes, for example, a ROM (Read Only Memory), a RAM (Random Access Memory), a nonvolatile memory such as a hard disk, and the like. The interface device 1300 includes, for example, a USB interface, an earphone interface, and the like. The communication device 1400 is capable of, for example, wired or wireless communication, and may specifically include Wifi communication, Bluetooth communication, 2G/3G/4G/5G communication, and the like. The display device 1500 is, for example, a liquid crystal display, a touch display, or the like. The input device 1600 may include, for example, a touch screen, a keyboard, a somatosensory input, and the like. The user can input/output voice information through the speaker 1700 and the microphone 1800 .

The electronic device shown in FIG. 1 is merely illustrative and in no way imply any limitation on the present disclosure, its application, or use. Applied to the embodiments of the present disclosure, the memory 1200 of the electronic device 1000 is configured to store instructions, and the instructions are configured to control the processor 1100 to operate to execute any one of the AI applications provided by the embodiments of the present disclosure integrated management approach. Those skilled in the art should understand that although the electronic device 1000 shows multiple devices in FIG. 1 , the present disclosure may only relate to some of the devices. For example, the electronic device 1000 only relates to the processor 1100 and the storage device 1200 . A skilled person can design instructions according to the solutions disclosed in the present disclosure. How the instruction controls the processor to operate is well known in the art, so it will not be described in detail here.

In this embodiment, an integrated management method for AI applications is provided. The management method of the AI application may be implemented by an electronic device. The electronic device may be the electronic device 1000 shown in FIG. 1 . The electronic device 1000 of this embodiment may be pre-deployed with a device of an integrated management system of an AI application.

In order to solve the current problems of non-interoperability of data of various AI applications, different standards, complicated operation and maintenance, and huge management costs brought about by the development and management of various AI applications, this disclosure creatively proposes that various AI applications should be implemented according to certain requirements. The design concept of integrating the standards for unified management, proposes an integrated management system for AI applications, which can also be called AI operating system, or AIOS for short, which can integrate various APPs through certain standards, and can Unified management of integrated APPs. Specifically, AI applications can be installed on AIOS, and AI applications can also be built through AIOS. AIOS also provides various systems and application management tools to manage and control the hardware resources required for running AI applications. and software resources.

In an example, the electronic device in this embodiment may be a computer cluster, and then the AIOS may be deployed in the computer cluster.

In one embodiment of the present disclosure, the AIOS can provide a running environment for the AI application, where the running environment specifically includes services and interfaces necessary for the management and development of the AI application.

The AIOS in this embodiment can also be used to manage and control the hardware resources and software resources required for the running of the AI application, so that all the resources required by the AI application can be used to the maximum extent.

Further, AIOS can also provide a user operation interface that conforms to operating system cognition, which is convenient for users to manage AI applications in AIOS.

As shown in FIG. 2 , the integrated management method for AI applications in this embodiment may include the following steps S2100 to S2300:

Step S2100: Acquire an installation file in a predetermined standard format corresponding to the AI application, and install the first AI application based on the installation file in the predetermined standard format.

In one embodiment of the present disclosure, the first AI application may be installed in the AIOS based on an installation file in a predetermined standard format. Then, the first AI application in this embodiment may be an AI application installed in the AIOS according to the installation file.

In order to integrate one or more first AI applications on the AIOS, the standard format of the installation files of the first AI applications needs to be pre-specified. In order to be installed in the AIOS, the installation files of different first AI applications need to meet the same requirements. Standard format for installation files.

In this embodiment, the predetermined standard format may specifically be a file format set in advance according to the AIOS, and the corresponding first AI application is installed with an installation file based on the predetermined standard format. For example, the predetermined standard format may be the Docker standard format.

In an embodiment of the present disclosure, acquiring the installation file in the predetermined standard format of the first AI application includes steps S2110 to S2120 as shown below:

Step S2110: Obtain the original installation file of the first AI application.

In this embodiment, the original installation file may be an application installation file provided by a third-party service, for example, an installation file provided by a developer of the first AI application. The original installation file may be downloaded from the application store provided by the AIOS of this embodiment, may also be downloaded from a designated download address, or may be obtained from other electronic devices that pre-store the original installation file, It may also be obtained from the local disk of the electronic device implementing this embodiment. The present disclosure does not limit the source of the original installation file.

Step S2120: Convert the format of the original installation file into a predetermined standard format to obtain an installation file in the predetermined standard format.

In this embodiment, after the installation file in the predetermined standard format is obtained by converting the format of the original installation file into a predetermined standard format, the corresponding first AI application can be installed based on the installation file in the predetermined standard format.

In one embodiment of the present disclosure, the method may further include: providing an application store in which at least one first AI application is provided.

In one example, an installation file in a predetermined standard format of the first AI application may be provided in the application store. Then, acquiring the original installation file of the first AI application may include: downloading the installation file of the first AI application in a predetermined standard format from the application store.

In another example, the original installation file of the first AI application may be provided in the application store. Then, obtaining the original installation file of the first AI application may include: downloading the original installation file of the first AI application from the application store, and converting the format of the original installation file into a predetermined standard format to obtain an installation file in the predetermined standard format.

In this embodiment, by installing the first AI application in the AIOS, the installed first AI application can run in the AIOS. Moreover, AIOS can also manage the installed first AI application.

Step S2200: Acquire construction configuration information of the second AI application, and build a second AI application based on the construction configuration information.

AIOS can not only install the above-mentioned first AI application, such as applications developed by third-party manufacturers or customized development, but also provide the ability to build a second AI application. Users can use the services and tools provided by AIOS to build A second AI application that matches business needs.

In one embodiment of the present disclosure, the configuration information may include storage configuration information and computing configuration information. Then, based on the construction configuration information, constructing the second AI application may include: constructing the second AI application based on the storage configuration information and the computing configuration information.

AIOS can provide a standard process for building AIOS. In one example, the storage configuration information and computing configuration information related to the application to be built can be received or automatically obtained based on a predetermined configuration process, and AIOS can perform AI application based on the configuration information. Construction and subsequent management. The predetermined configuration process may be a standard application construction process provided by AIOS, and a user (for example, an application developer) only needs to configure (input) the configuration information required to construct the second AI application according to this standard application construction process (input). . For example, a standard application building process interface and guide can be provided to the user, and the user only needs to input the configuration information required to build the second AI application according to the interface and guide. According to an exemplary embodiment of the present disclosure, the information of the configuration required to construct the second AI application may be classified into storage configuration information and computing configuration information according to information usage.

According to an exemplary embodiment of the present disclosure, the storage configuration information may include, but is not limited to, at least one of input source information, application meta information, log and monitoring indicator collection information, and storage declaration information. Hereinafter, the storage configuration information according to an exemplary embodiment of the present disclosure will be described in detail.

According to an exemplary embodiment of the present disclosure, the input source information may include the identification of one or more standard Docker images. Here, Docker can be selected as the general input convention. Docker has the most extensive developer base and is the most commonly used containerization technology, which conforms to the evolution design of advanced enterprise architecture. In addition, Docker also has the following advantages: faster startup speed than virtual machines, consistent operating environment, easier migration, performance closer to native applications than virtual machines, easier reuse, Docker ecosystem has a large number of high-quality high-quality The official mirror image greatly reduces the cost of mirror production. In terms of generality, Docker can adapt to mainstream cloud vendors at home and abroad, including AWS, Google Cloud, Microsoft Azure, Alibaba Cloud, Tencent Cloud, Huawei Cloud, etc., which means that any cloud user can contribute Docker applications. Docker is compatible with resource management and scheduling systems commonly used by enterprises, such as Kubernetes (K8s), Mesos, etc., which means that almost any enterprise customer can contribute to Docker applications. Docker is compatible with operating systems commonly used in the industry, and can run on Linux, Mac, and Windows, which means that developers of any operating system can contribute to Docker applications.

According to an exemplary embodiment of the present disclosure, the application meta information is information required for unified operation and management of the second AI application. The application meta information may include, but is not limited to, at least one of the following items: unique application identification, authentication information, unique application name, unique display name, application icon, description document, creation time, update time, creator information, Version number, application source type, startup resource consumption information, data source, application storage address, image address, and startup method. Specifically, the unique application identifier may be generated by AIOS, and the user (in this context, the user may be an application developer) may be unaware. The authentication information can be passed into the system key by the user, and the correctness is verified by AIOS. The unique application name can be passed in by the developer and verified by AIOS for uniqueness. The unique display name can be passed in by the developer and verified by AIOS for uniqueness. The application icon can be passed in by the developer and verified by AIOS for uniqueness. The description document can be imported by the developer, and the completeness and internationalization specifications are verified by AIOS. The second level needs to include introduction and help documents. The creation time can be generated by AIOS, and the user can not perceive it. The update time can be generated by AIOS, and the user may not perceive it. The creator information can be passed in by the user, and can be checked for correctness by AIOS. The version number can be passed in by the user. The application source type can be generated by the system, and the user may not perceive it, and is used to indicate whether the application is an official preset application or a third-party application. The startup resource consumption information can be passed in by the user to indicate the number of instances, whether to automatically expand or shrink, CPU, GPU, MEM, FPGA, etc. The data source can be an official preset data form, a standard data source provided by the underlying storage system, or another Docker image to connect to a third-party data source. The application storage address refers to the data storage space in which the second AI application is running. The image address can be passed in by the developer, and is used to indicate the startup path of the second AI application. The startup mode can be passed in by the developer to indicate whether the same second AI application can be repeatedly opened by different application users during operation. Of course, the application meta information may not be limited to the above-mentioned information, and may also include information required by the AIOS to manage the second AI application.

According to an exemplary embodiment of the present disclosure, the log and monitoring indicator collection information may be input by the user to indicate how to obtain the log and monitoring indicators. For example, if the user gets the logs using the standard error stream, no awareness is required. If the user obtains the monitoring indicators through AIOS, it can also be insensitive. If users need to customize monitoring indicators, they can use the software development kit (SDK) provided by AIOS for customized development.

According to an exemplary embodiment of the present disclosure, the storage claim information may be input by the user to claim the storage system used, and a corresponding time-to-live (TTL) policy and the like may be configured. For example, a storage system may include, but is not limited to, file storage, database storage, and the like.

According to an exemplary embodiment of the present disclosure, the computing configuration information may include, but is not limited to, at least one of service startup mode information, dynamic resource information, and computing capability information. Hereinafter, the computing configuration information according to the exemplary embodiment of the present disclosure will be described in detail.

According to an exemplary embodiment of the present disclosure, the service launch mode information may be input by the user to declare the application launch mode. Users do not need to perceive the startup process of the second AI application, but only need to configure the service startup method information, and the system can automatically initiate application startup to the underlying resource management and altitude platforms.

According to an exemplary embodiment of the present disclosure, the dynamic resource information may be input by the user to declare the requested application dynamic resource. For example, a dynamic resource can include two parts. The first part can be a dynamic service developed by the user or its own resource that the task can control and can be exposed to the application user for modification. The second part can be a dynamic service or task developed by the user in the A resource that can be additionally requested during operation to start a new service or task.

According to an exemplary embodiment of the present disclosure, the computing capability information is configurable by a user to design application computing capabilities. For example, a user can use the SDK provided by AIOS to call a wide range of algorithms (eg, AI algorithms), that is, one or more algorithms can be called by using the SDK. Here, AI algorithms may include, but are not limited to, binary classification, multi-classification, regression, clustering, time series analysis, image classification, image detection, optical character recognition (OCR), speech recognition, natural language processing (NLP), migration Learning, reinforcement learning, federated learning, automatic machine learning, parameter tuning, etc. For another example, if the computing power required by the user is not within the computing power provided by the AIOS, the user can also perform custom development through the AIOS, and enjoy the ability to mix with the built-in computing power.

According to an exemplary embodiment of the present disclosure, the above-mentioned storage configuration information and computing configuration information may be configured according to the standard development process provided by AIOS. For example, the standard development process according to the exemplary embodiment of the present disclosure may be as shown in FIG. 3 . 3 is a schematic diagram of a standard development process according to an exemplary embodiment of the present disclosure. That is, the user can configure information in the following order: in step S201, receiving input source information; in step S202, receiving application meta information; in step S203, receiving service startup mode information; in step S204, receiving dynamic resource information; in step S205 , receive log and monitoring instruction collection information; in step S206 , receive storage declaration information; in step S207 , receive computing capability information. Of course, the standard development process is not limited to the above sequence, and can also include any other feasible sequence. For example, steps S201 and S202 are the first and second steps, and steps S203 to S207 can be performed in any sequence; The above-mentioned steps S201 to S207 are executed according to a user-defined process, that is, the above-mentioned steps S201 to S207 can be executed in any possible order according to user requirements.

According to an exemplary embodiment of the present disclosure, a part of the above-mentioned storage configuration information and computing configuration information may be obtained by a user configuring a trait provided by the AIOS. Here, the characteristic capability may refer to an entity in AIOS that has encapsulated available resources for directly invoking related capabilities. For example, the AIOS may include a feature capability library including at least one feature capability. The user can configure the feature capability in the feature capability library provided by the AIOS to obtain corresponding information as required, and the AIOS can call the corresponding feature capability based on the corresponding information to construct the second AI application.

For example, when the input source information includes identifications of multiple standard Docker images, the input source may be an arrangement of multiple standard Docker images. The orchestration of multiple standard Docker images can be obtained by configuring the image orchestration feature provided by AIOS. Here, the orchestration of images may include online and offline hybrid orchestration when multiple images include online images and offline images, online orchestration when multiple images are all online images, and offline orchestration when multiple images are all offline images.

For example, the service startup mode information can be obtained by the user configuring the service startup mode feature capability provided by the AIOS.

For example, the storage claim information can be obtained by the user configuring the storage feature capability provided by the AIOS. Of course, in addition to using storage feature capabilities to obtain storage declaration information, storage solutions can also be replaced with open-source, enterprise-level storage solutions, such as Amazon's S3 storage, storage connected to Microsoft Cloud, and so on.

For example, the parameter tuning information can be obtained by configuring the storage feature capability and parameter tuning feature capability provided by the AIOS by the user.

According to an exemplary embodiment of the present disclosure, when a part of the information in the storage configuration information and the calculation configuration information is obtained by configuring the characteristic capability by the user, the corresponding characteristic capability may be invoked based on the part of the information for constructing the second characteristic capability. AI applications.

Step S2300: Manage the installed first AI application and the constructed second AI application.

In one embodiment of the present disclosure, a management tool for managing AI applications may be provided in AIOS. The management tool may include, for example, a resource management module, a license verification module, an authority management module, an audit module, a notification module and a log module, a file management module, a data center, an application management module, and a process management module.

The resource management module can be used to monitor the computing resources and/or storage resources consumed by the AI application during the running process, and can also be used to monitor the server and related hardware resources required by the AI application during the running process. The user can configure and manage the resources consumed by the AI application during the running process in the AIOS through the resource management module. Resource monitoring and management through the resource management module can ensure the stable operation of AIOS and AI applications and ensure business continuity.

The license verification module can provide management of software licenses for the use of AIOS and/or AI applications. The user can flexibly restrict the usage time, application scope and functional modules of AIOS and/or AI applications through the license verification module, so as to Effective intellectual property protection for developers of AIOS and/or AI applications.

The authority management module is used to manage the user's operation authority and operation capability scope on the AIOS. Specifically, the user who uses the AIOS, the tools that the user can use in the AIOS, and the management area of the user resources can be managed through the rights management module.

The audit module provides users with the audit function of AIOS, and audit records for key user operation behaviors to ensure the stability and traceability of AIOS. In the audit module, each user's operation and behavior in each AI application can be monitored Records, provides background functions for viewing, and is supplemented by a comprehensive monitoring panel to analyze the operation of the entire AIOS.

The notification module can be used to set and manage the notification content and notification method for AIOS and each AI application. In one example, the notification module can be used for alarming, and alarm management provides users with the ability to individualize alarms for monitoring items. Users can flexibly formulate alarm rules for specified monitoring items, and set custom alarm policies for serious AIOS problems. , when a business indicator reaches a critical value or there is an abnormality in the AIOS, the notification module can send an alarm in the form of SMS and email as expected by the user, prompting the user to know and deal with the abnormality in time, thereby improving the availability of AIOS and business applications.

The log module is a tool for unified collection, management, screening, and retrieval of various operating data of AI applications. Users can monitor and analyze operations and maintenance based on logs, and monitor and track call chains to meet various business analysis and other digital operation and maintenance and operation scenarios.

In AIOS, there is generally a file system for users to store, query and manage the files of the entire AIOS. The file management module is used to support user uploads and data generated by other AI applications that users want to view and consume. Among them, the basic data types that can be managed include tables, pictures, voice, text, video, and various user-defined data formats. Users upload the defined data types from other data sources, such as data tables, data groups, etc. After uploading AIOS, they can preview, view the meta information of the data, delete and other operations in the file management module. The data stored in the file management module can be consumed in AI applications for tasks such as model training. At the same time, AI applications may also generate a large number of logs, reports, various short-term intermediate data, etc. These data can be viewed through the back-end view of the file management module to view the data generated in the entire AIOS and stored in HDFS and other media such as object storage engines. data, and perform tasks such as data cleaning.

The data center can be used to manage and store the relationship of multiple basic data, and the abstract data form of governance logic. In the existing technology, the construction of AI applications needs to define goals according to business requirements, collect data, perform feature engineering, and obtain a possible result after model training. The intermediate operations greatly depend on the quality of the data itself. level and completeness of the application environment. Although this method can maximize the modeling effect, its inefficient application generation capability is also one of the thresholds that hinder the popularization of current AI technology. Therefore, the embodiments of the present disclosure provide a data center in AIOS, and AI data related to AI applications can be obtained through the data center. The data form of AI data can be regarded as a file format, similar to the traditional DOC file format.

The data center can provide the use of AI data based on the agreement with the AI application, which can be opened by the AI application specified in the data form, and combined with the functions of the AI application to solve specific problems. The data center can include data usage methodologies for application problems in specific business areas, such as data acquisition methods, data governance methods, and data relationships. The data center can also display the basic file meta format obtained by AIOS.

The application management module is a management tool that carries all AI applications that different users have access rights to, and its meta information can be displayed in the application management module (for example, it can be an application logo, etc.). For application developers, the application management module can provide a complete set of registration procedures for standard AI applications, including application for resources, control of start and stop, application for storage space, connection of permission system, integration of monitoring, etc. process.

The process management module can be used to manage and monitor the running process and resources of the AI application. During the running process of AI applications, active processes will be generated. For AI applications, process management is the management of running program entities, such as all system resources occupied by the running AI applications. to manage. In the process of running AI applications, the resources that AI applications need to use and schedule are usually very complex, involving scheduling among multiple computing units. Concepts such as "sub" and "operation stage", disassemble the operation of AI applications according to the dimension of business organization, convert monitoring from the perspective of layered and machine to service-centered perspective, and understand the relationship between services within the monitoring scope. Dependencies can simplify the difficulty of the process management module to manage and monitor the operation of AI applications.

While AIOS provides convenience services for AI applications, it also provides development tools for AI applications, such as Studio and HyperCycle, to meet the mainstream development methods and development experience in the market.

Studio development tools can provide visual AI application development modules, including modeling IDE (Integrated Development Environment) module, application IDE module, model management and application management. Organize and connect data, algorithms, evaluation, and debugging to help modeling users clearly and intuitively define the entire process of various types of model development, model management and evaluation, and application construction, orchestration, and debugging.

Modeling IDE can provide machine learning model research and development tools. Users can process data and produce models within this tool. The modeling IDE can provide two modeling environments for interactive logic: DAG (Directed Acyclic Graph) and Notebook. DAG is a low-threshold visual modeling tool. Users can perform visual operations on the DAG canvas and connect various data processing and machine learning operators to define the entire machine learning process. Notebook is a professional command-line modeling tool that allows users to quickly validate ideas and get feedback by editing code block by block.

The models produced in the Studio development tools are all taken over by the model management, and also support the import of external models for overall management. At the same time, with the model as the core, the evaluation and interpretation reports related to the model are summarized to help users understand the model from more perspectives.

The application IDE provides tools for building, arranging, debugging, and packaging AI applications, and can provide users with complete, reliable, and highly flexible enterprise-level AI application development capabilities. In the application IDE, users can complete the creation of applications according to the needs of actual business scenarios, through selecting models/DAGs, configuring resources and operating strategies, and packaging applications. .

After users develop AI applications that meet business requirements, they are deployed through application management. Application management can realize all-round management and control of AI applications, including operation and maintenance monitoring, debugging experiments, dynamic resource configuration, version management, etc., to meet all the work requirements of users for the online production environment of AI applications.

HyperCycle development tool is a low-threshold AI application construction tool (including HC scenario definition, machine learning HyperCycle, image classification HyperCycle, etc., tools for various business problems), through the division of data, business, IT and automatic model strategy Blessing helps enterprises solve AI business problems on a large scale, and realize the application and implementation of rapid AI scenarios.

Specifically, data personnel can collect and sort out the data streams and data relationships that need to be accessed according to business requirements, define a HyperCycle scenario dedicated to a certain business, and configure learning tasks and learning frequencies through the HyperCycle development tool. The HyperCycle development tool automatically matches the corresponding automatic learning strategy according to the algorithm type and data type, and updates the online business data according to the set update frequency to prevent attenuation. Users can monitor the performance indicators through the product front desk to observe the operation of the model in the business. Operation and maintenance personnel can judge the health of business and data according to the activity performance and index in the process manager.

In an embodiment of the present disclosure, AIOS can also provide users with an extremely concise operation experience, which is logically in line with the user's pre-judgment of AIOS, and can also increase the understandability of AIOS with the help of the user's experience in using other operating systems With improved flexibility and operability, users can better focus on their own business goals, lower the threshold for understanding AIOS, and improve work efficiency.

Based on the above description, in an embodiment of the present disclosure, managing the installed first AI application and the constructed second AI application includes at least one of the following:

Obtain the data produced by the target AI application during the running process, and store the produced data;

Collect the running data of the target AI application to obtain the running log of the target AI application;

Monitor at least one of the currently running processes, operators, workflows and running stages of the target AI application to obtain resources occupied by at least one of the processes, operators, workflows and running stages;

Monitor the user's operation behavior for the target AI application, and obtain the audit records of the first AI application and the second AI application;

Monitor the specified indicators of the target AI application during the running process, and issue an alarm notification when the specified indicators are abnormal;

In response to the operation of running the target AI application, verify the user's operation permission for the target AI application, and run the target AI application if the user has the operation permission for the target AI application;

Accept and store the data uploaded by the user, and provide the uploaded data to the target AI application for consumption.

In this embodiment, the target AI application may include at least one first AI application and at least one second AI application.

In the embodiment in which the management of the installed first AI application and the constructed second AI application includes acquiring the data produced by the target AI application during the running process, and storing the produced data, the target AI application may The data uploaded to AIOS is consumed for consumption, and the output data is obtained, which can include tasks such as model training, prediction, classification, recommendation, and natural language processing. The data types of the data uploaded by the user may include tables, pictures, voice, text, video, and various data formats customized by the user.

In this embodiment, the file management module may be used to store, view and manage the data uploaded by the user to the AIOS and the data produced by the target AI application during the running process.

Based on this embodiment, the method may further include: in response to the operation of viewing the generated data, displaying the generated data for the user to view.

Specifically, the user can view and manage the data produced by the target AI application during the running process by opening the file management module.

In the embodiment where the management of the installed first AI application and the constructed second AI application includes collecting the running data of the target AI application to obtain the running log of the target AI application, the target AI may be collected through the log module in the AIOS The running data of the application is generated, and the running log of the target AI application is generated.

After obtaining the operation log of the target AI application, the method may further include: in response to the operation of viewing the operation log, displaying the operation log for the user to view.

In this embodiment, the user may trigger the operation of viewing the running log by opening the log module, and the AIOS displays the running log in response to the operation of viewing the running log for the user to view and analyze.

The management of the installed first AI application and the built second AI application includes monitoring at least one of the currently running process, operator, workflow and running stage of the target AI application, so as to obtain the current running process, In the embodiment of the resource occupied by at least one of the operator, the workflow, and the running phase, the resources occupied by at least one of the currently running process, operator, workflow, and running phase of the target AI application may include computing resources, storage resources and/or hardware resources.

In this embodiment, a process is a running activity of a program in a computer on a certain data set, a basic unit for resource allocation and scheduling by a system, and a basis for an operating system structure. A program is a description of instructions, data, and their organization, and a process is the entity of a program. By monitoring and managing the current process of the AI application during the running process, the difficulty of monitoring and managing the AI application during the running process can be simplified.

In one example, the computing resource may include a hadoop cluster, the storage resource may include MySQL, and the hardware resource may include at least one of a CPU, a GPU, and a memory.

By monitoring the computing resources and/or storage resources consumed by the AI application during the running process, it is convenient for the user to manage the current activities of the AI application during the running process, so as to ensure the stable operation of the AIOS and the AI application.

After acquiring the resources occupied by at least one of the process, operator, workflow and running phase, the method further includes: in response to the operation of viewing the occupied resources, displaying the occupied resources for the user to view. In this way, when the resources occupied by at least one of the process, operator, workflow and running phase are too large, the user can stop running the target AI application according to his own needs, or adjust the target AI application to improve the target AI application. availability.

In this embodiment, by monitoring at least one of the currently running process, operator, workflow and running phase of the target AI application, the dependencies between the services involved in the running process of the target AI application can be known, It can simplify the difficulty of the process management module to monitor the operation and management of AI applications.

In the embodiment where the management of the installed first AI application and the constructed second AI application includes monitoring the user's operation behavior with respect to the target AI application, and obtaining the audit records of the first AI application and the second AI application, the AIOS may The audit module in the system obtains the user's operation behavior for the target AI application and obtains the audit record.

After obtaining the audit record of the target AI application, the method may further include: in response to the operation of viewing the audit record, displaying the audit record for the user to view.

The user can trigger the operation of viewing the audit record by opening the audit module, and AIOS displays the audit record in response to the operation of viewing the audit record for the user to view and analyze. In this way, the traceability of the target AI application can be guaranteed.

In the embodiment where the management of the installed first AI application and the constructed second AI application includes monitoring the specified indicators of the target AI application during the running process, when the specified indicators are abnormal, an alarm notification is issued, which may be Through the notification module in AIOS, set the effective range of the index value of the specified index of the target AI application during the running process, and monitor the specified index of the target AI application during the running process. When the actual index value of the specified index exceeds the preset value If it is within the valid range, it is determined that the specified indicator is abnormal, and an alarm notification is issued according to the notification mode set by the user in the notification module. The user can view the alarm notifications issued in the history by opening the notification module.

Managing the installed first AI application and the built second AI application includes verifying the user's operation permission for the target AI application in response to the operation of running the target AI application, in the case that the user has the operation permission for the target AI application , In the embodiment of running the target AI application, whether the user has the right to the target AI application can be determined by verifying whether the current time when the operation of running the target AI application in the AIOS is received is within the permitted usage time of the target AI application through the license verification module. Permission to operate the AI application. If the user has the operation permission for the target AI application, run the target AI application in the AIOS; if the user does not have the operation permission for the target AI application, the target AI application may be prohibited from running in the AIOS. In this way, effective intellectual property protection can be performed for developers of target AI applications.

In the embodiment in which the management of the installed first AI application and the constructed second AI application includes accepting and storing the data uploaded by the user, and providing the uploaded data to the target AI application for consumption, in this embodiment, it is possible to It is through the file management module to store, view and manage the data uploaded by users to AIOS. The data type of the data uploaded by the user may include table, picture, voice, text, video, and various data formats defined by the user.

The target AI application can be consumed according to the data uploaded by the user, which can specifically include tasks such as model training, prediction, classification, recommendation, and natural language processing.

In one embodiment of the present disclosure, providing the AI data to the target AI application for consumption further includes:

Obtain AI data related to the target AI application from the uploaded data, and provide the relevant AI data to the target AI application for consumption.

In this embodiment, the user can set the data format of the AI data related to the target AI application through the data center in advance, so that the data center filters the data stored in the file management module according to the data format, and only sends the data to the target AI application. AI applications provide AI data related to them.

Based on this embodiment, the method may further include: displaying the AI data in response to the operation of viewing the AI data.

Specifically, the user may open the data center and trigger the operation of viewing the data center. In response to the operation of the viewing data center, AIOS displays AI data related to AI applications for users to view and select.

In another embodiment of the present disclosure, providing the AI data to the target AI application for consumption further includes:

Obtain AI data related to the target AI application from the uploaded data, filter the data that matches the application scenario corresponding to the target AI application from the AI data, and provide it to the target AI for consumption.

Specifically, the user may specify the application scenario corresponding to the target AI application in advance according to actual needs.

The data center in AIOS can filter out the data that matches the application scenario from the AI data. In this way, AI data that matches the application scenario can be automatically provided to the target AI application, without the need for manual screening by the user, which can reduce user operations and improve user experience.

In an embodiment of the present disclosure, the method may further include:

Determine the user's operation authority for the integrated AI application, and provide the corresponding AI application according to the operation authority.

The AI applications mentioned in the embodiments of the present disclosure may include at least one first AI application and/or at least one second AI application.

Specifically, for users with different operation rights, the AI applications provided by AIOS may be different. In this embodiment, the user can only use the AI application provided according to the corresponding operation authority.

For example, corresponding to the first user and the second user with different operation permissions, the AI application provided to the first user may include AI application A and AI application B, and the AI application provided to the second user may include AI Application B and AI Application C. Then, the first user can use AI application A and AI application B, but cannot use AI application C; the second user can use AI application B and AI application C, but cannot use AI application A.

In addition, the method may further include: providing corresponding computing resources and/or storage resources in the AIOS according to the operation authority.

In an embodiment of the present disclosure, the method may further include:

Provides a management desktop, on which there are shortcuts of some or all of the integrated AI applications and/or management tool shortcuts and/or file shortcuts; in response to editing operations on the desktop, edit the shortcuts provided in the desktop .

The editing operations in this embodiment may include operations such as adding, moving, deleting, and storing. Users can add AI application shortcuts and/or management tool shortcuts and/or file shortcuts to the desktop according to their own needs, or delete AI application shortcuts and/or management tool shortcuts and/or shortcuts displayed on the desktop. Or file shortcuts, or adjust the shortcuts of AI applications displayed on the desktop and/or the placement of management tool shortcuts and/or file shortcuts on the desktop, or change the desktop used to accommodate the AI displayed on the desktop Folders of application shortcuts and/or management tool shortcuts and/or file shortcuts, so that users can manage the desktop more efficiently and conveniently.

In this embodiment, the user can add any accessible shortcut on the desktop according to his own preferences and work habits, to help the user open the corresponding management tool, AI application or file more conveniently.

Through the integrated management method in this embodiment, the AI applications deployed in the AIOS can be integrated and managed in a unified manner, thereby reducing the development and management costs of the AI applications.

In this embodiment, an integrated management apparatus 3000 for AI applications is provided. As shown in FIG. 4 , the integrated management apparatus 3000 may include a first installation module 3100 , a second construction module 3200 and an application management module 3300 . The first installation module 3100 is configured to obtain an installation file in a predetermined standard format of the first AI application, and install the first AI application based on the installation file in the predetermined standard format; the second construction module 3200 is configured to obtain a second AI application Based on the construction configuration information, the second AI application is constructed based on the construction configuration information; the application management module 3300 is configured to manage the installed first AI application and the constructed second AI application.

In an embodiment of the present disclosure, the first installation module 3100 may be specifically configured as:

Obtain the original installation file of the target AI application, and convert the format of the original installation file into a predetermined standard format.

In one embodiment of the present disclosure, the predetermined standard format includes a Docker standard format.

In one embodiment of the present disclosure, constructing the configuration information includes storing the configuration information and computing the configuration information;

The second building block 3200 can be configured to:

A second AI application is constructed based on the storage configuration information and the computing configuration information.

In an embodiment of the present disclosure, the integrated management apparatus 3000 may further include:

an application store providing module configured to provide an application store where at least one first AI application is provided;

The first installation module 3100 can also be configured to:

Download a predetermined standard format installation file of the first AI application from the application store; or

Download the original installation file of the first AI application from the application store, and convert the format of the original installation file into a predetermined standard format.

In one embodiment of the present disclosure, the application management module 3300 is configured to perform at least one of the following:

Obtain the data produced by the first AI application and the second AI application during the running process, and store the produced data;

Collect the operation data of the first AI application and the second AI application, and obtain the operation logs of the first AI application and the second AI application;

monitoring at least one of the currently running processes, operators, workflows and running phases of the first AI application and the second AI application to obtain resources occupied by at least one of the processes, operators, workflows and running phases;

Monitor the user's operation behavior for the first AI application and the second AI application, and obtain the audit records of the first AI application and the second AI application;

Monitor the specified indicators during the operation of the first AI application and the second AI application, and issue an alarm notification when the specified indicators are abnormal;

In response to the operation of running the first AI application and the second AI application, verify the user's operation permission for the first AI application and the second AI application, if the user has the operation permission for the first AI application and the second AI application , run the first AI application and the second AI application;

Accept and store the data uploaded by the user, and provide the uploaded data to the first AI application and the second AI application for consumption.

In an embodiment of the present disclosure, the integrated management apparatus 3000 may further include a presentation module configured to:

After storing the output data, display the output data in response to the operation of viewing the output data;

or,

After obtaining the audit records of the first AI application and the second AI application, in response to the operation of viewing the audit records, displaying the audit records;

or,

After obtaining the operation logs of the first AI application and the second AI application, displaying the operation logs in response to the operation of viewing the operation logs;

or,

After acquiring the resources occupied by at least one of the process, the operator, the workflow and the running phase, the occupied resources are displayed in response to the operation of viewing the occupied resources.

In one embodiment of the present disclosure, the application management module 3300 may be configured to:

Obtain AI data related to the first AI application and the second AI application from the uploaded data, and provide the AI data to the first AI application and the second AI application for consumption;

or,

Obtain AI data related to the first AI application and the second AI application from the uploaded data, filter out the data matching the application scenarios corresponding to the first AI application and the second AI application from the AI data, and provide the data to The first AI is applied and the second AI is consumed.

The authority management module is configured to determine the user's operation authority to the integrated AI application, and provide the corresponding AI application according to the operation authority.

The desktop management module, configured as:

Provide a management desktop with integrated shortcuts to some or all of the AI applications and/or management tool shortcuts and/or file shortcuts;

In response to an editing operation on the administrative desktop, the shortcut provided in the desktop is edited.

In an embodiment of the present disclosure, a part of the information in the storage configuration information and the computing configuration information is obtained by configuring a feature capability by a user, where the feature capability refers to an entity that has encapsulated available resources for directly invoking related capabilities;

The second building module 3200 may also be configured to: call the corresponding characteristic capability based on a part of the storage configuration information and the computing configuration information to build the second AI application.

In an embodiment of the present disclosure, the storage configuration information may include at least one of input source information, application meta information, log and monitoring indicator collection information, and storage declaration information;

The computing configuration information includes at least one of service startup mode information, dynamic resource information, and computing capability information.

In an embodiment of the present disclosure, constructing a second AI application based on the storage configuration information and the computing configuration information includes at least one of the following steps:

Based on the input source information, docking the input source;

Manage the second AI application based on the application meta information;

Based on the service startup method information, initiate an application startup instruction to the underlying resource management and scheduling platform;

Based on dynamic resource information, perform dynamic resource management;

Provide logs and monitoring information of the second AI application based on the collection of information from logs and monitoring indicators;

Manage the storage of the second AI application based on the storage declaration information;

Based on the computing capability information, the computing capability of the second AI application is configured.

In one embodiment of the present disclosure, the input source information includes identifiers of one or more standard Docker images;

When the input source information includes the identifiers of multiple standard Docker images, the input source is the arrangement of the multiple standard Docker images, wherein the arrangement based on the multiple standard Docker images is obtained by configuring the feature capability of the image arrangement by the user.

In one embodiment of the present disclosure, the application meta information includes at least one of the following items:

Unique application ID, authentication information, unique application name, unique display name, application icon, description document, creation time, update time, creator information, version number, application source type, startup resource consumption information, data source, application storage address , mirror address, startup method.

In one embodiment of the present disclosure, the service startup mode information is obtained by a user configuring a service startup mode characteristic capability.

In one embodiment of the present disclosure, the storage claim information is obtained by a user configuring storage characteristic capabilities.

In one embodiment of the present disclosure, the computing capability information is obtained by invoking one or more algorithms by a user using a software development kit, and/or obtained by custom development of an operator platform;

The parameter tuning information in the computing capability information is obtained by configuring the parameter tuning feature capability of the user.

Those skilled in the art should understand that the integrated management apparatus 3000 for AI applications can be implemented in various ways. For example, the integrated management apparatus 3000 of the AI application may be implemented by configuring the processor through instructions. For example, the instructions may be stored in ROM, and when the device is started, the instructions are read from the ROM into the programmable device to implement the integrated management apparatus 3000 of the AI application. For example, the integrated management device 3000 of the AI application can be built into a dedicated device (eg, ASIC). The integrated management apparatus 3000 of the AI application may be divided into mutually independent units, or they may be combined together for implementation. The integrated management apparatus 3000 of the AI application may be implemented by one of the above various implementation manners, or may be implemented by a combination of two or more of the above various implementation manners.

In this embodiment, the integrated management apparatus 3000 for AI applications may have various implementation forms. For example, the integrated management apparatus 3000 for AI applications may be any software product that provides management services for AI applications or functional modules running in an application program , or peripheral embeds, plug-ins, patches, etc. of these software products or applications, or these software products or applications themselves.

In this embodiment, as shown in FIG. 5 , an AI application integrated management system 4000 including at least one computing device 4100 and at least one storage device 4200 is also provided, which may also be referred to as an AI operating system. The at least one storage device 4200 is configured to store executable instructions; the instructions, when executed by the at least one computing device, cause the at least one computing device 4100 to execute an integrated management method for an AI application according to any embodiment of the present disclosure.

In this embodiment, the system 4000 may be a mobile phone, a tablet computer, a palmtop computer, a desktop computer, a notebook computer, a workstation, a game console, etc., or a distributed system composed of multiple devices.

<Computer-readable storage medium>

In this embodiment, a computer-readable storage medium is also provided, on which a computer program is stored, and when the computer program is executed by a processor, implements the integrated management method of an AI application according to any embodiment of the present disclosure.

The present disclosure may be an apparatus, method and/or computer program product. The computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for causing a processor to implement various aspects of the present disclosure.

A computer-readable storage medium may be a tangible device that can hold and store instructions for use by the instruction execution device. The computer-readable storage medium may be, for example, but not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. More specific examples (non-exhaustive list) of computer readable storage media include: portable computer disks, hard disks, random access memory (RAM), read only memory (ROM), erasable programmable read only memory (EPROM) or flash memory), static random access memory (SRAM), portable compact disk read only memory (CD-ROM), digital versatile disk (DVD), memory sticks, floppy disks, mechanically coded devices, such as printers with instructions stored thereon Hole cards or raised structures in grooves, and any suitable combination of the above. Computer-readable storage media, as used herein, are not to be construed as transient signals per se, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (eg, light pulses through fiber optic cables), or through electrical wires transmitted electrical signals.

The computer readable program instructions described herein may be downloaded to various computing/processing devices from a computer readable storage medium, or to an external computer or external storage device over a network such as the Internet, a local area network, a wide area network, and/or a wireless network. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and/or edge servers. A network adapter card or network interface in each computing/processing device receives computer-readable program instructions from a network and forwards the computer-readable program instructions for storage in a computer-readable storage medium in each computing/processing device .

Computer program instructions for carrying out operations of the present disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state setting data, or instructions in one or more programming languages. Source or object code, written in any combination, including object-oriented programming languages, such as Smalltalk, C++, etc., and conventional procedural programming languages, such as the "C" language or similar programming languages. The computer readable program instructions may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer, or entirely on the remote computer or server implement. In the case of a remote computer, the remote computer may be connected to the user's computer through any kind of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (eg, using an Internet service provider through the Internet connect). In some embodiments, custom electronic circuits, such as programmable logic circuits, field programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), can be personalized by utilizing state information of computer readable program instructions. Computer readable program instructions are executed to implement various aspects of the present disclosure.

Aspects of the present disclosure are described herein with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the disclosure. It will be understood that each block of the flowchart illustrations and/or block diagrams, and combinations of blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer readable program instructions.

These computer readable program instructions may be provided to a processor of a general purpose computer, special purpose computer or other programmable data processing apparatus to produce a machine that causes the instructions when executed by the processor of the computer or other programmable data processing apparatus , resulting in means for implementing the functions/acts specified in one or more blocks of the flowchart and/or block diagrams. These computer readable program instructions can also be stored in a computer readable storage medium, these instructions cause a computer, programmable data processing apparatus and/or other equipment to operate in a specific manner, so that the computer readable medium on which the instructions are stored includes An article of manufacture comprising instructions for implementing various aspects of the functions/acts specified in one or more blocks of the flowchart and/or block diagrams.

Computer readable program instructions can also be loaded onto a computer, other programmable data processing apparatus, or other equipment to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus, or other equipment to produce a computer-implemented process , thereby causing instructions executing on a computer, other programmable data processing apparatus, or other device to implement the functions/acts specified in one or more blocks of the flowcharts and/or block diagrams.

The flowchart and block diagrams in the Figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of instructions, which comprises one or more functions for implementing the specified logical function(s) executable instructions. In some alternative implementations, the functions noted in the blocks may occur out of the order noted in the figures. For example, two blocks in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It is also noted that each block of the block diagrams and/or flowchart illustrations, and combinations of blocks in the block diagrams and/or flowchart illustrations, can be implemented in dedicated hardware-based systems that perform the specified functions or actions , or can be implemented in a combination of dedicated hardware and computer instructions. It is well known to those skilled in the art that implementation in hardware, implementation in software, and implementation in a combination of software and hardware are all equivalent.

Various embodiments of the present disclosure have been described above, and the foregoing descriptions are exemplary, not exhaustive, and not limiting of the disclosed embodiments. Numerous modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The terminology used herein was chosen to best explain the principles of the embodiments, the practical application or technical improvement in the marketplace, or to enable others of ordinary skill in the art to understand the embodiments disclosed herein. The scope of the present disclosure is defined by the appended claims.

Industrial Applicability

Through the embodiment of the present disclosure, the installation file of the first artificial intelligence application in a predetermined standard format is obtained, and the first artificial intelligence application is installed based on the installation file in the predetermined standard format; the construction configuration information of the second artificial intelligence application is obtained, and based on the construction configure information, and construct a second artificial intelligence application; manage the installed first artificial intelligence application and the constructed second artificial intelligence application. The installation of the first artificial intelligence application and the constructed second artificial intelligence application in the artificial intelligence operating system can be managed uniformly, so that the development and management costs of the artificial intelligence application can be reduced. Therefore, the present disclosure has strong industrial applicability.

Claims

An integrated management method for artificial intelligence applications, including:

obtaining an installation file in a predetermined standard format of the first artificial intelligence application, and installing the first artificial intelligence application based on the installation file in the predetermined standard format;

obtaining construction configuration information of a second artificial intelligence application, and constructing the second artificial intelligence application based on the construction configuration information;

The installed first artificial intelligence application and the constructed second artificial intelligence application are managed.
According to the method according to claim 1, the obtaining the installation file of the predetermined standard format of the first artificial intelligence application comprises:

Obtain the original installation file of the first artificial intelligence application, and convert the format of the original installation file into the predetermined standard format.
The method of claim 1 or 2, wherein the predetermined standard format comprises a Docker standard format.
The method according to any one of claims 1 to 3,

The construction configuration information includes storage configuration information and computing configuration information;

The building the second artificial intelligence application based on the building configuration information includes:

The second artificial intelligence application is constructed based on the storage configuration information and the computing configuration information.
The method of any one of claims 1 to 4, further comprising:

providing an application store in which at least one first artificial intelligence application is provided;

The obtaining the installation file in the predetermined standard format of the first artificial intelligence application includes:

Download an installation file in a predetermined standard format of the first artificial intelligence application from the application store; or

Download the original installation file of the first artificial intelligence application from the application store, and convert the format of the original installation file into the predetermined standard format.
The method according to any one of claims 1 to 5, wherein the managing the installed first artificial intelligence application and the constructed second artificial intelligence application comprises at least one of the following:

Acquiring data produced by the first artificial intelligence application and the second artificial intelligence application during operation, and storing the produced data;

collecting the operation data of the first artificial intelligence application and the second artificial intelligence application to obtain the operation logs of the first artificial intelligence application and the second artificial intelligence application;

Monitoring at least one of currently running processes, operators, workflows, and running phases of the first artificial intelligence application and the second artificial intelligence application to obtain the processes, the operators, the workflows, and the running phases. resources occupied by at least one of the running phases;

Monitoring the operation behavior of the user with respect to the first artificial intelligence application and the second artificial intelligence application, and obtaining audit records of the first artificial intelligence application and the second artificial intelligence application;

monitoring the specified indicators during the running process of the first artificial intelligence application and the second artificial intelligence application, and issuing an alarm notification when the specified indicators are abnormal;

In response to the operation of running the first artificial intelligence application and the second artificial intelligence application, the user's operating permission for the first artificial intelligence application and the second artificial intelligence application is verified, and the user has access to all the artificial intelligence applications. run the first artificial intelligence application and the second artificial intelligence application under the operating permission of the first artificial intelligence application and the second artificial intelligence application;

Accept and store the data uploaded by the user, and provide the uploaded data to the first artificial intelligence application and the second artificial intelligence application for consumption.
The method of claim 6,

After storing the output data, the method further includes: in response to the operation of viewing the output data, displaying the output data;

or,

After acquiring the audit records of the first artificial intelligence application and the second artificial intelligence application, the method further includes: displaying the audit records in response to the operation of viewing the audit records;

or,

After obtaining the operation logs of the first artificial intelligence application and the second artificial intelligence application, the method further includes: displaying the operation logs in response to the operation of viewing the operation logs;

or,

After acquiring the resources occupied by at least one of the process, the operator, the workflow and the running phase, the method further includes: in response to an operation of viewing the occupied resources, displaying the resources occupied.
The method according to claim 6 or 7, the providing the uploaded data to the first artificial intelligence application and the second artificial intelligence application for consumption, comprising:

Obtain artificial intelligence data related to the first artificial intelligence application and the second artificial intelligence application from the uploaded data, and provide the related artificial intelligence data to the first artificial intelligence application and the second artificial intelligence application. The second artificial intelligence application for consumption;

or,

Obtain artificial intelligence data related to the first artificial intelligence application and the second artificial intelligence application from the uploaded data, and filter out the artificial intelligence data related to the first artificial intelligence application and the second artificial intelligence application from the relevant artificial intelligence data. The data matched by the application scenario corresponding to the second artificial intelligence application is provided to the first artificial intelligence application and the second artificial intelligence for consumption.
The method of any one of claims 1 to 8, further comprising:

Determine the user's operation authority for the integrated artificial intelligence application, and provide the corresponding artificial intelligence application according to the operation authority.
The method of any one of claims 1 to 9, further comprising:

Provide a management desktop on which shortcuts to some or all of the integrated artificial intelligence applications and/or management tool shortcuts and/or file shortcuts are provided;

In response to an editing operation on the desktop, the shortcut provided in the desktop is edited.
The method according to claim 4, wherein a part of the information in the storage configuration information and the computing configuration information is obtained by configuring a characteristic capability by a user, wherein the characteristic capability means that the available resources have been encapsulated for direct The entity that invokes the relevant capability;

The constructing the second artificial intelligence application based on the storage configuration information and the computing configuration information includes: calling a corresponding feature based on the storage configuration information and the part of the computing configuration information capabilities to build the second artificial intelligence application.
The method according to claim 4 or 11, wherein the storage configuration information includes at least one of input source information, application meta information, log and monitoring indicator collection information, and storage declaration information;

The computing configuration information includes at least one of service startup mode information, dynamic resource information, and computing capability information.
An integrated management device for artificial intelligence applications, comprising:

a first installation module, configured to obtain an installation file in a predetermined standard format of a first artificial intelligence application, and install the first artificial intelligence application based on the installation file in the predetermined standard format;

a second building module, configured to obtain building configuration information of a second artificial intelligence application, and build the second artificial intelligence application based on the building configuration information;

The application management module is configured to manage the installed first artificial intelligence application and the constructed second artificial intelligence application.
According to the integrated management apparatus of claim 13, the first installation module is configured to: obtain an original installation file of the target artificial intelligence application, and convert the format of the original installation file into the predetermined standard format.
The integrated management apparatus according to claim 13 or 14, wherein the predetermined standard format includes a Docker standard format.
The integrated management device according to any one of claims 13 to 15,

The construction configuration information includes storage configuration information and computing configuration information;

The second building module is configured to build the second artificial intelligence application based on the storage configuration information and the computing configuration information.
The integrated management device according to any one of claims 13 to 16, further comprising:

an application store providing module configured to provide an application store in which at least one first artificial intelligence application is provided;

The first installation module is configured to include:

Download an installation file in a predetermined standard format of the first artificial intelligence application from the application store; or

Download the original installation file of the first artificial intelligence application from the application store, and convert the format of the original installation file into the predetermined standard format.
The integrated management apparatus according to any one of claims 13 to 17, the application management module is configured to perform at least one of the following:

Acquiring data produced by the first artificial intelligence application and the second artificial intelligence application during operation, and storing the produced data;

collecting the operation data of the first artificial intelligence application and the second artificial intelligence application to obtain the operation logs of the first artificial intelligence application and the second artificial intelligence application;

Monitoring at least one of currently running processes, operators, workflows, and running phases of the first artificial intelligence application and the second artificial intelligence application to obtain the processes, the operators, the workflows, and the running phases. resources occupied by at least one of the running phases;

Monitoring the operation behavior of the user with respect to the first artificial intelligence application and the second artificial intelligence application, and obtaining audit records of the first artificial intelligence application and the second artificial intelligence application;

monitoring the specified indicators during the running process of the first artificial intelligence application and the second artificial intelligence application, and issuing an alarm notification when the specified indicators are abnormal;

In response to the operation of running the first artificial intelligence application and the second artificial intelligence application, the user's operating permission for the first artificial intelligence application and the second artificial intelligence application is verified, and the user has access to all the artificial intelligence applications. run the first artificial intelligence application and the second artificial intelligence application under the operating permission of the first artificial intelligence application and the second artificial intelligence application;

Accept and store the data uploaded by the user, and provide the uploaded data to the first artificial intelligence application and the second artificial intelligence application for consumption.
The integrated management device of claim 18, further comprising a presentation module configured to:

After storing the output data, in response to the operation of viewing the output data, displaying the output data;

or,

After obtaining the audit records of the first artificial intelligence application and the second artificial intelligence application, in response to the operation of viewing the audit records, displaying the audit records;

or,

After obtaining the operation logs of the first artificial intelligence application and the second artificial intelligence application, in response to the operation of viewing the operation logs, displaying the operation logs;

or,

After acquiring the resources occupied by at least one of the process, the operator, the workflow and the running phase, the occupied resources are displayed in response to the operation of viewing the occupied resources.
According to the integrated management device of claim 17 or 18, the application management module is configured to:

Obtain artificial intelligence data related to the first artificial intelligence application and the second artificial intelligence application from the uploaded data, and provide the related artificial intelligence data to the first artificial intelligence application and the second artificial intelligence application. The second artificial intelligence application for consumption;

or,

Obtain artificial intelligence data related to the first artificial intelligence application and the second artificial intelligence application from the uploaded data, and filter out the artificial intelligence data related to the first artificial intelligence application and the second artificial intelligence application from the relevant artificial intelligence data. The data matched by the application scenario corresponding to the second artificial intelligence application is provided to the first artificial intelligence application and the second artificial intelligence for consumption.
The integrated management device according to any one of claims 13 to 20, further comprising:

The authority management module is configured to determine the operation authority of the user to the integrated artificial intelligence application, and provide the corresponding artificial intelligence application according to the operation authority.
The integrated management device according to any one of claims 13 to 21, further comprising a desktop management module configured to:

Provide a management desktop on which shortcuts to some or all of the integrated artificial intelligence applications and/or management tool shortcuts and/or file shortcuts are provided;

In response to an editing operation on the management desktop, the shortcut provided in the desktop is edited.
The integrated management device according to claim 16, wherein a part of the information in the storage configuration information and the computing configuration information is obtained by configuring a characteristic capability by a user, wherein the characteristic capability means that the available resources have been encapsulated to entities for direct invocation of related capabilities;

The second building module is configured to: call a corresponding characteristic capability based on the storage configuration information and the part of the computing configuration information to build the second artificial intelligence application.
The integrated management device according to claim 16 or 23, wherein the storage configuration information includes at least one of input source information, application meta information, log and monitoring indicator collection information, and storage declaration information;

The computing configuration information includes at least one of service startup mode information, dynamic resource information, and computing capability information.
An integrated management system for artificial intelligence applications comprising at least one computing device and at least one storage device, wherein the at least one storage device is configured to store instructions that, when executed by the at least one computing device, cause all of the The at least one computing device performs the following steps of integrated management of artificial intelligence applications:

obtaining an installation file in a predetermined standard format of the first artificial intelligence application, and installing the first artificial intelligence application based on the installation file in the predetermined standard format;

obtaining construction configuration information of a second artificial intelligence application, and constructing the second artificial intelligence application based on the construction configuration information;

The installed first artificial intelligence application and the constructed second artificial intelligence application are managed.
According to the system according to claim 25, the obtaining the installation file of the predetermined standard format of the first artificial intelligence application comprises:

Obtain the original installation file of the first artificial intelligence application, and convert the format of the original installation file into the predetermined standard format.
The system of claim 25 or 26, the predetermined standard format comprising the Docker standard format.
A system according to any one of claims 25 to 27,

The construction configuration information includes storage configuration information and computing configuration information;

The building the second artificial intelligence application based on the building configuration information includes:

The second artificial intelligence application is constructed based on the storage configuration information and the computing configuration information.
28. The system of any one of claims 25 to 28, the instructions, when executed by the at least one computing device, further cause the at least one computing device to perform the steps of:

providing an application store in which at least one first artificial intelligence application is provided;

The obtaining the installation file in the predetermined standard format of the first artificial intelligence application includes:

Download an installation file in a predetermined standard format of the first artificial intelligence application from the application store; or

Download the original installation file of the first artificial intelligence application from the application store, and convert the format of the original installation file into the predetermined standard format.
The system according to any one of claims 25 to 29, wherein the managing the installed first artificial intelligence application and the constructed second artificial intelligence application comprises at least one of the following:

Acquiring data produced by the first artificial intelligence application and the second artificial intelligence application during operation, and storing the produced data;

collecting the operation data of the first artificial intelligence application and the second artificial intelligence application to obtain the operation logs of the first artificial intelligence application and the second artificial intelligence application;

Monitoring at least one of currently running processes, operators, workflows, and running phases of the first artificial intelligence application and the second artificial intelligence application to obtain the processes, the operators, the workflows, and the running phases. resources occupied by at least one of the running phases;

Monitoring the operation behavior of the user with respect to the first artificial intelligence application and the second artificial intelligence application, and obtaining audit records of the first artificial intelligence application and the second artificial intelligence application;

monitoring the specified indicators during the running process of the first artificial intelligence application and the second artificial intelligence application, and issuing an alarm notification when the specified indicators are abnormal;

In response to the operation of running the first artificial intelligence application and the second artificial intelligence application, the user's operating permission for the first artificial intelligence application and the second artificial intelligence application is verified, and the user has access to all the artificial intelligence applications. run the first artificial intelligence application and the second artificial intelligence application under the operating permission of the first artificial intelligence application and the second artificial intelligence application;

Accept and store the data uploaded by the user, and provide the uploaded data to the first artificial intelligence application and the second artificial intelligence application for consumption.
The system of claim 30,

After storing the data for the output, the instructions, when executed by the at least one computing device, further cause the at least one computing device to perform the step of performing a recommendation regarding a configuration file: in response to viewing the output The operation of the data, showing the output data;

or,

After obtaining the audit records of the first artificial intelligence application and the second artificial intelligence application, the instructions, when executed by the at least one computing device, further cause the at least one computing device to perform a recommendation on a configuration file The steps of: displaying the audit record in response to the operation of viewing the audit record;

or,

After the operation logs of the first artificial intelligence application and the second artificial intelligence application are obtained, the instructions, when executed by the at least one computing device, further cause the at least one computing device to perform a recommendation on a configuration file The steps of: displaying the operation log in response to the operation of viewing the operation log;

or,

After acquiring the resources occupied by at least one of the process, the operator, the workflow and the running phase, the instruction, when executed by the at least one computing device, further causes the at least one The computing device performs the step of presenting the occupied resources in response to the operation of viewing the occupied resources.
The system of claim 30 or 31, said providing the uploaded data to the first artificial intelligence application and the second artificial intelligence application for consumption, comprising:

Obtain artificial intelligence data related to the first artificial intelligence application and the second artificial intelligence application from the uploaded data, and provide the related artificial intelligence data to the first artificial intelligence application and the second artificial intelligence application. The second artificial intelligence application for consumption;

or,

Obtain artificial intelligence data related to the first artificial intelligence application and the second artificial intelligence application from the uploaded data, and filter out the artificial intelligence data related to the first artificial intelligence application and the second artificial intelligence application from the relevant artificial intelligence data. The data matched by the application scenario corresponding to the second artificial intelligence application is provided to the first artificial intelligence application and the second artificial intelligence for consumption.
32. The system of any one of claims 25 to 32, the instructions, when executed by the at least one computing device, further cause the at least one computing device to perform the steps of:

Determine the user's operation authority to the integrated artificial intelligence application, and provide the corresponding artificial intelligence application according to the operation authority.
33. The system of any one of claims 25 to 33, the instructions, when executed by the at least one computing device, further cause the at least one computing device to perform the steps of:

Provide a management desktop on which shortcuts to some or all of the integrated artificial intelligence applications and/or management tool shortcuts and/or file shortcuts are provided;

In response to an editing operation on the desktop, the shortcut provided in the desktop is edited.
The system according to claim 28, wherein a part of the information in the storage configuration information and the computing configuration information is obtained by configuring a feature capability by a user, wherein the feature capability means that the available resources have been encapsulated for direct The entity that invokes the relevant capability;

Wherein, the constructing the second artificial intelligence application based on the storage configuration information and the computing configuration information includes: calling a corresponding characteristic based on the storage configuration information and the part of the computing configuration information capabilities to build the second artificial intelligence application.
The system according to claim 28 or 35, wherein the storage configuration information includes at least one of input source information, application meta information, log and monitoring indicator collection information, and storage declaration information;

The computing configuration information includes at least one of service startup mode information, dynamic resource information, and computing capability information.
A computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements the method of any one of claims 1 to 12.