WO2022267874A1

WO2022267874A1 - Troubleshooting method and system, electronic device, and computer readable storage medium

Info

Publication number: WO2022267874A1
Application number: PCT/CN2022/097234
Authority: WO
Inventors: 赵亮; 杜永生; 林礼剑
Original assignee: 中兴通讯股份有限公司
Priority date: 2021-06-24
Filing date: 2022-06-06
Publication date: 2022-12-29
Also published as: CN115529256A

Abstract

Embodiments of the present application relate to the technical field of communications, and in particular to a troubleshooting method and system, an electronic device, and a computer readable storage medium. The method comprises: if natural language information used for representing a troubleshooting intent for a network is received, inputting the natural language information into a preset natural language processing (NLP) model to obtain keyword information corresponding to the natural language information; invoking an application related to the keyword information to optimize the network; and if it is detected that the optimized network does not have a network fault corresponding to the troubleshooting intent, generating first feedback information, wherein the first feedback information is used for representing that the network fault has been removed.

Description

Troubleshooting method, system, electronic device and computer readable storage medium

Cross References to Related Applications

This application is based on the Chinese patent application with the application number "202110706128.X" and the filing date is June 24, 2021, and claims the priority of the Chinese patent application. The entire content of the Chinese patent application is incorporated herein by reference. into this application.

technical field

The embodiments of the present application relate to the technical field of communications, and in particular, to a troubleshooting method, system, electronic device, and computer-readable storage medium.

Background technique

With the rapid development of communication technology, communication networks have become an indispensable part of human production and life, and communication access equipment and communication services have grown exponentially. Humans have higher and higher requirements for the quality of communication networks. Faults in the network will affect the quality of the communication network. These faults need to be discovered and resolved in a timely manner. Therefore, network operation and maintenance troubleshooting has become an important part of ensuring communication quality.

However, when operators perform network operation and maintenance troubleshooting, network optimization and maintenance personnel need to manually determine which areas of the communication network have faults, and enter the interactive interface of the product module corresponding to the fault according to the field to which the fault belongs to interact to solve the problem. For this fault, finally enter the interactive interface of some product modules to check whether the fault has been eliminated. The whole process requires a lot of human labor and time. There are too many human-computer interactions, and the efficiency of network operation and troubleshooting is very low.

Contents of the invention

An embodiment of the present application provides a troubleshooting method, the method comprising: if receiving natural language information representing a troubleshooting intention for the network, inputting the natural language information into a preset natural language processing ( Natural Language Processing (abbreviation: NLP) model, obtains keyword information corresponding to the natural language information; invokes an application related to the keyword information to optimize the network; if the optimized network is detected If there is no network fault corresponding to the troubleshooting intention, first feedback information is generated; where the first feedback information is used to indicate that the network fault has been eliminated.

The embodiment of the present application also provides a troubleshooting system, including an acquisition module, an intent translation module, an execution module, an application module, and a feedback module. The intention translation module includes a preset natural language processing NLP model; the acquisition module is used for Obtaining natural language information used to characterize network troubleshooting intentions, and sending the natural language information to the intention translation module; the intention translation module is used to input the natural language information into a preset natural language Processing the NLP model, obtaining keyword information corresponding to the natural language information, and sending the keyword information to the execution module; the execution module is used to call an application related to the keyword information, and execute the The network is optimized; the application module is used for accepting the call of the execution module; the feedback module is used for generating a first Feedback information; wherein, the first feedback information is used to indicate that the network fault has been eliminated.

The embodiment of the present application also provides an electronic device, including: at least one processor; and a memory connected in communication with the at least one processor; wherein, the memory stores information that can be executed by the at least one processor. Instructions, the instructions are executed by the at least one processor, so that the at least one processor can execute the above troubleshooting method.

The embodiment of the present application also provides a computer-readable storage medium storing a computer program, and implementing the above troubleshooting method when the computer program is executed by a processor.

Description of drawings

FIG. 1 is a flowchart one of a troubleshooting method according to an embodiment of the present application;

FIG. 2 is a second flowchart of a troubleshooting method according to another embodiment of the present application;

Fig. 3 is a flowchart of optimizing the network by invoking applications related to keyword information according to an embodiment of the present application;

FIG. 4 is a flow chart of inputting natural language information into a preset NLP model and obtaining keyword information corresponding to the natural language information according to one embodiment of the present application;

FIG. 5 is a flow chart for obtaining empirical rules for operation, maintenance and troubleshooting provided in an embodiment of the present application;

FIG. 6 is a flow chart of receiving natural language information used to characterize network troubleshooting intentions according to an embodiment of the present application;

FIG. 7 is a schematic structural diagram of a troubleshooting system according to another embodiment of the present application;

Fig. 8 is a schematic structural diagram of an electronic device according to another embodiment of the present application.

detailed description

The main purpose of the embodiment of the present application is to propose a troubleshooting method, system, electronic device, and computer-readable storage medium, aiming at shortening the time for network operation and maintenance troubleshooting, reducing the number of human-computer interactions, and improving network operation and maintenance troubleshooting s efficiency.

In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the embodiments of the present application will be described in detail below with reference to the accompanying drawings. However, those of ordinary skill in the art can understand that in each embodiment of the application, many technical details are provided for readers to better understand the application. However, even without these technical details and various changes and modifications based on the following embodiments, the technical solutions claimed in this application can also be realized. The division of the following embodiments is for the convenience of description, and should not constitute any limitation to the specific implementation of the present application, and the embodiments can be combined and referred to each other on the premise of no contradiction.

An embodiment of the present application relates to a troubleshooting method applied to an electronic device, where the electronic device may be a terminal or a server. In this embodiment and the following embodiments, the electronic device is described by taking the server as an example. The implementation details of the troubleshooting method of this embodiment are described in detail below, and the following content is only the implementation details provided for the convenience of understanding, and is not necessary for implementing the solution.

The specific process of the troubleshooting method in this embodiment can be shown in Figure 1, including:

Step 101, if the natural language information representing the network troubleshooting intention is received, input the natural language information into a preset NLP model to obtain keyword information corresponding to the natural language information.

In a specific implementation, the server can receive the natural language information input by the user in real time to represent the intention of troubleshooting the network, and after receiving the natural language information used to represent the intention of troubleshooting the network, the server can convert the natural language information to Input to the preset NLP model to obtain keyword information corresponding to natural language information. Wherein, the preset NLP model may be one of the NLP models trained in advance by those skilled in the art, or an open source NLP model obtained from the Internet, which is not specifically limited in the embodiment of the present application. The preset NLP model is pre-stored in a memory inside the server, or in an external memory connected to the server.

In an example, the server can obtain in real time the natural language information used to characterize the network troubleshooting intention entered by the user in text form, and directly input the textual natural language information used to characterize the network troubleshooting intention into the In the preset NLP model.

In another example, the server can obtain the natural language information input by the user in voice form in real time to characterize the intention to troubleshoot the network. After the language information, the server can convert the natural language information in the form of speech used to represent the intention to troubleshoot the network into natural language information in the form of text used to represent the intention to troubleshoot the network, and convert the converted natural language information Input to the preset NLP model.

In one example, after the server inputs the natural language information into the preset NLP model, the acquired keyword information corresponding to the network failure includes at least but not limited to: the business field to which the network failure corresponding to the troubleshooting intention belongs, the network failure occurrence , the logical and/or geographic object corresponding to the network failure, the required action and the target corresponding to the required action. Keyword information includes domain, time, object, operation and operation target, which can be more convenient for the server to understand and execute, and further improve the efficiency of network operation and maintenance troubleshooting.

For example: the natural language information input by the user obtained by the server and used to represent the network troubleshooting intention is: "On June 15, 2021, the signal coverage of cell A is poor, optimize the traffic of cell A". After the server inputs the natural language information into the preset NLP model, the obtained keyword information corresponding to the network failure may include: coverage area, June 15, 2021, cell A, optimization, and traffic.

In an example, the keyword information acquired by the server may be stored in the form of data pairs. For example: the output of the NLP model is "coverage area, June 15, 2021, cell A, optimization, traffic", and the server generates a data pair "<coverage, June 15, 2021, A cell, optimization, traffic>", and use this data pair as the keyword information corresponding to the network fault.

Step 102, calling an application related to keyword information to optimize the network.

In a specific implementation, after obtaining the keyword information corresponding to the natural language information, the server can generate an execution strategy instruction based on the keyword information, and invoke an application related to the keyword information according to the execution strategy instruction to optimize the network. Among them, the server can store a number of applications for operation, maintenance and troubleshooting in various fields, and directly call the stored applications. The server can also communicate with a device that stores a number of applications for operation, maintenance, and troubleshooting in various fields, and indirectly call these applications.

In an example, the keyword information corresponding to the natural language information obtained by the server is "<coverage, June 15, 2021, cell A, optimization, traffic>", and the server can call "Service Optimization" related applications to optimize the network.

Step 103, if it is detected that the optimized network does not have a network fault corresponding to the troubleshooting intention, first feedback information is generated.

In a specific implementation, after the server optimizes the network by invoking the application related to the keyword information, it can detect whether the optimized network still exists according to the logical object and/or geographic object corresponding to the network fault in the keyword information. Network failure, if it is detected that the network failure does not exist in the optimized network, first feedback information indicating that the network failure has been eliminated is generated, and the user is notified that the network optimization is successful and the network failure has been eliminated.

In an example, the server can judge whether the optimized network still has a network failure corresponding to the troubleshooting intention according to the preset optimization standard. The preset optimization standard can be a certain key performance indicator (Key Performance Indicator, referred to as : KPI), or any combination of several KPIs, which is not specifically limited in the embodiments of the present application. If the server judges that the optimized network meets the preset optimization standard, the server can generate the first feedback information.

In this embodiment, if the server receives the natural language information used to represent the network troubleshooting intention, it will input the natural language information into the preset NLP model, obtain the keyword information corresponding to the natural language information, and then automatically Invoke the application related to the keyword information to optimize the network, and when it is detected that the optimized network does not have a network fault corresponding to the troubleshooting intention, generate first feedback information for indicating that the network fault has been eliminated, for The user knows that the network fault has been eliminated, and the process of obtaining keyword information, invoking various applications to optimize the network, and giving feedback based on the network optimization results are all automatically completed by the server without manual participation, which greatly shortens the time. In the time of network operation and maintenance troubleshooting, users only need to input natural language information used to represent the network troubleshooting intention, and then automatic troubleshooting can be completed, minimizing the number of human-computer interactions, thereby improving network operation and maintenance troubleshooting At the same time, the obtained execution strategy and the process of calling the application are highly visualized, which is convenient for users to track.

In one embodiment, the preset NLP model can be pre-trained by the server and stored in the internal memory of the server. The server can iteratively train the NLP model based on the training corpus. The training corpus can include but not limited to: the system specification of the network, User manuals and traditional troubleshooting interaction data for several O&M troubleshooting applications. Training the NLP model based on manuals and traditional troubleshooting interaction data can make the trained NLP model more scientific and reasonable, so that the obtained operation strategy can better complete network operation and maintenance troubleshooting.

In an example, when the server iteratively trains the NLP model based on the training corpus, it may first perform data preprocessing on the training corpus in the training corpus. The preprocessing includes data enhancement, word segmentation, and vectorization.

Another embodiment of the present application relates to a troubleshooting method. The implementation details of the troubleshooting method in this embodiment are described in detail below. The following content is only the implementation details provided for the convenience of understanding, and is not necessary for implementing this solution. Figure 2 It is a flowchart of the troubleshooting method described in this embodiment, including:

Step 201, if the natural language information representing the intention to troubleshoot the network is received, input the natural language information into a preset NLP model to obtain keyword information corresponding to the natural language information.

Step 202, calling an application related to keyword information to optimize the network.

Wherein, Step 201 to Step 202 are substantially the same as Step 101 to Step 102, and will not be repeated here.

Step 203, if it is detected that the network fault corresponding to the troubleshooting intention still exists in the optimized network, second feedback information is generated.

In a specific implementation, after the server optimizes the network by invoking the application related to the keyword information, it can detect whether the optimized network still has a fault based on the logical object and/or geographic object corresponding to the network fault in the keyword information. If it is detected that the network fault still exists in the optimized network, the second feedback information indicating that the network fault has not been eliminated will be generated to inform the user that the network optimization has not been successful and the network fault has not been eliminated. , need to continue to optimize.

In an example, the server can judge whether the optimized network still has a network failure corresponding to the troubleshooting intention according to the preset optimization standard. The preset optimization standard can be a certain KPI, or any combination of several KPIs. The embodiment of the present application does not specifically limit this. If the server determines that the optimized network does not meet the preset optimization standard, the server may generate second feedback information.

In step 204, the application related to the keyword information is invoked again to optimize the network.

In a specific implementation, after the server generates the second feedback information, it can call the application related to the keyword information again, continue to optimize the network, and continue to judge whether there is a network fault corresponding to the troubleshooting intention in the optimized network after optimization , until the optimized network does not have the network fault.

In one example, the keyword information obtained by the server corresponding to the network failure is "<coverage, June 15, 2021, cell A, optimization, traffic>", and the server detects that the optimized network still has "coverage After the failure of "poor", you can call the applications related to "coverage problem" and "traffic optimization" again to continue optimizing the network.

In this embodiment, after the network is optimized by invoking the application related to the keyword information, it further includes: if it is detected that the network failure corresponding to the troubleshooting intention still exists in the optimized network, The second feedback information is generated; wherein, the second feedback information is used to indicate that the network failure has not been eliminated; the application related to the keyword information is called again to optimize the network, the embodiment of the present application , if the optimized network still has the network fault corresponding to the troubleshooting intention, it means that the optimization is not up to standard and the network fault has not been eliminated. The server can generate the second feedback information to indicate that the fault has not been eliminated, and inform the user that it needs to continue to optimize. , and call the application related to the keyword information again to optimize the network until the network fault does not exist in the network, and the troubleshooting is completed.

In one embodiment, the application for operation, maintenance and troubleshooting may include a first type of application for query and a second type of application for network optimization. The server invokes an application related to keyword information to optimize the network. The realization of each sub-step as shown in Figure 3 specifically includes:

Step 301, call the first type of application related to keyword information, detect whether the network fault corresponding to the troubleshooting intention exists, if yes, execute step 302, otherwise, directly execute step 303.

In a specific implementation, after the server obtains the keyword information corresponding to the network fault, it can first call the first type of application for query related to the keyword information, query the network, and detect the network fault corresponding to the troubleshooting intention Whether it exists, before optimizing the network, first detect whether the network fault really exists, which can effectively save network operation and maintenance troubleshooting resources.

In an example, the first type of application used for query may be a KPI data monitoring and calculation application, that is, it is judged whether the network fault exists according to the KPI data in the network.

Step 302, calling a second type of application related to keyword information to optimize the network.

In a specific implementation, if the server detects that the network failure does exist, it may invoke the second type of application for network optimization related to keyword information to optimize the network.

In one example, after the server invokes the second type of application related to the keyword information to optimize the network, it can call the first type of application related to the keyword information again to detect whether the network failure exists: if the network failure If it does not exist, it is determined that the network optimization is successful, and the first feedback information used to represent that the network fault has been eliminated is generated; if the network fault still exists, it is determined that the network optimization has failed, and the second feedback information used to represent that the network fault has not been eliminated is generated. Information, continue to optimize the network. In the embodiment of the present application, one detection can be performed before optimization and one after optimization, that is, double detection, which can effectively improve the accuracy of troubleshooting.

Step 303, generating third feedback information.

In a specific implementation, if the server detects that the network fault does not exist, there is no need to optimize the network, and the server can generate third feedback information that indicates that the network fault does not exist in the network, and inform the user that the network fault does not exist in the network , or the natural language information entered by the user to represent the intent to troubleshoot the network is incorrect.

In one embodiment, the server inputs the natural language description information into the preset NLP model to obtain the keyword information corresponding to the network fault, which can be implemented by the steps shown in Figure 4, specifically including:

Step 401, input natural language information into a preset NLP model, and obtain keyword information output by the NLP model.

Specifically, after the server receives the natural language information used to represent the network troubleshooting intention, it can input the natural language description information into the preset NLP model, and obtain the keyword information output by the NLP model.

Step 402 , compare the keyword information output by the NLP model with the pre-stored standard keyword information, and judge whether the keyword information output by the NLP model is complete, if yes, execute step 403 , otherwise, execute step 404 .

In a specific implementation, after the server obtains the keyword information output by the NLP model, it can compare the keyword information output by the NLP model with the pre-stored standard keyword information to determine whether the keyword information output by the NLP model is complete. Wherein, the preset standard keyword information may be set by those skilled in the art according to actual needs, which is not specifically limited in this embodiment of the present application. The server cannot optimize the network based on incomplete keyword information. The preset standard keyword information is complete keyword information. Use the preset standard keyword information as a standard to check whether the keyword information output by the NLP model is complete. It can ensure that network optimization can be carried out normally, and prevent errors in network operation and maintenance troubleshooting.

In an example, the preset standard keyword information may be "<the business field to which the network fault belongs, the time range of the network fault, the logical object and/or geographic object corresponding to the network fault, the operation to be performed, the required The target corresponding to the operation>", the keyword information output by the NLP model acquired by the server is "<coverage, cell A, optimization, traffic>", after comparison, the server confirms that the keyword information output by the NLP model lacks the occurrence of network failures Time range, the keyword information output by the NLP model is incomplete.

In step 403, the keyword information output by the NLP model is used as the keyword information corresponding to the natural language information.

In a specific implementation, if the server confirms that the keyword information output by the NLP model is complete, then it is determined that the network optimization can be performed normally, and the server uses the keyword information output by the NLP model as keyword information corresponding to the natural language information.

Step 404, complete the keyword information output by the NLP model according to the pre-stored empirical rules for operation, maintenance and troubleshooting.

In the specific implementation, if the server confirms that the keyword information output by the NLP model is incomplete, it can complete the keyword information output by the NLP model according to the pre-stored empirical rules for operation, maintenance and troubleshooting. Those skilled in the art make settings according to actual needs, which is not specifically limited in the embodiments of the present application. Complementing the keyword information output by the incomplete NLP model can ensure the normal progress of network optimization and prevent errors in network operation and maintenance.

In one example, the keyword information output by the NLP model is "<coverage, June 15th, cell A, optimization>", and the server determines that the keyword information output by the NLP model is incomplete and lacks the target corresponding to the operation to be performed. The server can determine according to the pre-stored operation, maintenance and troubleshooting empirical rules. For the coverage problem, it is usually solved by optimizing the traffic. The server completes the keyword information. The complete keyword information is "<coverage, June 15th, A Cell, Optimization, Traffic>".

In step 405, the keyword information output by the supplementary and complete NLP model is used as the keyword information corresponding to the natural language information.

Specifically, after the server completes the keyword information output by the NLP model, the keyword information output by the completed NLP model can be used as the keyword information corresponding to the natural language information to perform a subsequent network optimization process.

In one embodiment, the empirical rules for operation, maintenance and troubleshooting can be obtained through the steps shown in Figure 5, specifically including:

Step 501 , after acquiring the keyword information corresponding to the natural language information each time, saving the keyword information.

In a specific implementation, the server can obtain empirical rules for operation, maintenance and troubleshooting based on historical optimization records, and after obtaining the keyword information corresponding to the natural language information each time, it can save the keyword information corresponding to the natural language information in the server in the memory.

In an example, the keyword information corresponding to the natural language information acquired by the server is "<coverage, June 15th, cell A, optimization, traffic>", and the server can convert "<coverage, June 15th, A cell, optimization, traffic>" is saved to the internal memory of the server.

Step 502: Determine the probability of each keyword combination according to the saved keyword information.

Specifically, the server may calculate the probability of each keyword combination among several stored keyword information.

In one example, the server has saved 100 pieces of keyword information, among which 70 pieces of keyword information include both "A cell" and "traffic", and the keyword information including "A cell" and "traffic" is 30, then the server determines that the probability of the combination of "A cell" and "traffic" is 70%, and the probability of the combination of "A cell" and "traffic" is 30%.

In step 503, the probability of each keyword combination is used as a pre-stored empirical rule for operation, maintenance and troubleshooting.

In a specific implementation, after the server determines the probability of each keyword combination, it can use the probability of each keyword combination as a pre-stored empirical rule for operation, maintenance and troubleshooting, and output the incomplete NLP model according to the pre-stored empirical rules for operation, maintenance and troubleshooting. When supplementing the keyword information of the NLP model, according to the existing keywords in the keyword information output by the NLP model, the keywords with the highest combination probability with these keywords are added to the keyword information output by the NLP model. Based on the complete keyword information obtained in the previous troubleshooting process, the operation and maintenance troubleshooting empirical rules are true, scientific, and reliable, and can be friendly to complete incomplete keyword information for subsequent network operation and maintenance troubleshooting .

In one example, the keyword information output by the NLP model is "<coverage, June 15th, cell A, optimization>", which lacks the target corresponding to the operation that needs to be performed. The server determines the The keyword with the highest combination probability with existing keywords in the keyword information is "traffic", and the server adds "traffic" to the keyword information.

In one embodiment, the server receives the natural language information used to characterize the network troubleshooting intention, which may be implemented by the steps shown in Figure 6, specifically including:

Step 601 , judging whether the natural language information used to characterize the network troubleshooting intention meets the preset semantic standard, if yes, execute step 602 , otherwise, execute step 603 .

In a specific implementation, after receiving the natural language information input by the user to represent the network troubleshooting intention, the server can judge whether the natural language information input by the user is a legal input according to the preset semantic standard. The set semantic standard includes inputting at least one business field to which a network fault belongs, and performing semantic recognition on the natural language information used to represent the troubleshooting intention of the network, which can ensure that what enters the server is legal natural language information, and illegally input information Abandon to avoid wasting network O&M and troubleshooting resources.

In one example, the preset semantic standard is to input at least one business field to which a network fault belongs, and the natural language information input by the user to represent the intention of troubleshooting the network is "On June 15th, a fault occurred in cell A", The server confirms that the natural language information entered by the user to represent the intention to troubleshoot the network does not include the business domain to which the network fault belongs, and judges that the natural language information entered by the user to represent the intention to troubleshoot the network does not conform to the preset semantics standard.

Step 602, confirming that the natural language information used to characterize the network troubleshooting intention is received.

In a specific implementation, if the server judges that the natural language information input by the user to represent the network troubleshooting intention meets the preset semantic standard, the server confirms receipt of the natural language information used to represent the network troubleshooting intention.

Step 603, discarding the natural language information used to characterize the network troubleshooting intention.

In a specific implementation, if the server judges that the natural language information input by the user to represent the intention to troubleshoot the network does not meet the preset semantic standards, the server discards the natural language information input by the user to represent the intention to troubleshoot the network information.

Another embodiment of the present application relates to a troubleshooting system. The details of the troubleshooting system in this embodiment are described in detail below. The following content is only the implementation details provided for the convenience of understanding, and is not necessary for the implementation of this example. FIG. 7 is The schematic diagram of the results of the troubleshooting system in this embodiment includes: an acquisition module 701 , an intent translation module 702 , an execution module 703 , an application module 704 and a feedback module 705 .

The acquisition module 701 is connected with the intention translation module 702, and the intention translation module 702 is also connected with the execution module 703, and the execution module 703 is also connected with the application module 704 and the feedback module 705 respectively, and the application module 704 is also connected with the feedback module 705, and the intention translation module 702 Includes preset NLP models.

The acquiring module 701 is configured to acquire natural language information used to characterize the network troubleshooting intention, and send the acquired natural language information to the intention translation module 702;

The intent translation module 702 is used to input the natural language information into the preset NLP model, obtain keyword information corresponding to the natural language information, and send the keyword information to the execution module 703;

Execution module 703 is used for invoking the application relevant with keyword information, optimizes network;

The application module 704 is used to accept calls from the execution module 703;

The feedback module 705 is configured to generate first feedback information when it is detected that there is no network fault corresponding to the troubleshooting intention in the optimized network; wherein the first feedback information is used to indicate that the network fault has been rectified.

It is not difficult to find that this embodiment is a system embodiment corresponding to the above method embodiment, and this embodiment can be implemented in cooperation with the above method embodiment. The relevant technical details and technical effects mentioned in the above embodiments are still valid in this embodiment, and will not be repeated here to reduce repetition. Correspondingly, the relevant technical details mentioned in this embodiment can also be applied in the above embodiments.

It is worth mentioning that all the modules involved in this embodiment are logical modules. In practical applications, a logical unit can be a physical unit, or a part of a physical unit, or multiple physical units. Combination of units. In addition, in order to highlight the innovative part of the present application, units that are not closely related to solving the technical problem proposed in the present application are not introduced in this embodiment, but this does not mean that there are no other units in this embodiment.

Another embodiment of the present application relates to an electronic device, as shown in FIG. 8 , including: at least one processor 801; and a memory 802 communicatively connected to the at least one processor 801; wherein, the memory 802 stores There are instructions executable by the at least one processor 801, and the instructions are executed by the at least one processor 801, so that the at least one processor 801 can execute the troubleshooting methods in the foregoing embodiments.

Wherein, the memory and the processor are connected by a bus, and the bus may include any number of interconnected buses and bridges, and the bus connects one or more processors and various circuits of the memory together. The bus may also connect together various other circuits such as peripherals, voltage regulators, and power management circuits, all of which are well known in the art and therefore will not be further described herein. The bus interface provides an interface between the bus and the transceivers. A transceiver may be a single element or multiple elements, such as multiple receivers and transmitters, providing means for communicating with various other devices over a transmission medium. The data processed by the processor is transmitted on the wireless medium through the antenna, further, the antenna also receives the data and transmits the data to the processor.

The processor is responsible for managing the bus and general processing, and can also provide various functions, including timing, peripheral interface, voltage regulation, power management, and other control functions. Instead, memory may be used to store data that the processor uses when performing operations.

Another embodiment of the present application relates to a computer-readable storage medium storing a computer program. The above method embodiments are implemented when the computer program is executed by the processor.

That is, those skilled in the art can understand that all or part of the steps in the method of the above-mentioned embodiments can be completed by instructing related hardware through a program, the program is stored in a storage medium, and includes several instructions to make a device ( It may be a single-chip microcomputer, a chip, etc.) or a processor (processor) to execute all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage media include: U disk, mobile hard disk, read-only memory (Read-Only Memory, abbreviated: ROM), random access memory (Random Access Memory, abbreviated: RAM), magnetic disk or optical disc, etc. medium for program code.

Those of ordinary skill in the art can understand that the above-mentioned embodiments are specific embodiments for realizing the present application, and in practical applications, various changes can be made to it in form and details without departing from the spirit and spirit of the present application. scope.

Claims

A troubleshooting method, including:

If the natural language information used to characterize the troubleshooting intention of the network is received, the natural language information is input into a preset natural language processing NLP model to obtain keyword information corresponding to the natural language information;

Invoking an application related to the keyword information to optimize the network;

If it is detected that the optimized network does not have a network fault corresponding to the troubleshooting intention, first feedback information is generated; wherein the first feedback information is used to indicate that the network fault has been eliminated.
The troubleshooting method according to claim 1, wherein, after said calling an application related to said keyword information and optimizing said network, further comprising:

If it is detected that the optimized network still has a network failure corresponding to the troubleshooting intention, then generating second feedback information; wherein the second feedback information is used to indicate that the network failure has not been eliminated;

The application related to the keyword information is invoked again to optimize the network.
The troubleshooting method according to any one of claims 1 to 2, wherein the applications include a first type of application used for query and a second type of application used for network optimization;

The calling of applications related to the keyword information to optimize the network includes:

Invoking the first type of application related to the keyword information to detect whether the network fault corresponding to the troubleshooting intention exists;

If the network fault exists, call a second type of application related to the keyword information to optimize the network;

If the network failure does not exist, third feedback information is generated; wherein the third feedback information is used to indicate that the network failure does not exist in the network.
The troubleshooting method according to any one of claims 1 to 3, wherein the keyword information includes at least: the business field to which the network fault belongs, the time range in which the network fault occurs, and the network fault corresponding to The logical object and/or geographic object of , the operation to be performed and the target corresponding to the operation to be performed.
The troubleshooting method according to claim 4, wherein said inputting said natural language information into a preset NLP model and obtaining keyword information corresponding to said natural language information comprises:

Inputting the natural language information into a preset NLP model, and obtaining keyword information output by the NLP model;

Comparing the keyword information output by the NLP model with pre-stored standard keyword information, and judging whether the keyword information output by the NLP model is complete;

If the keyword information output by the NLP model is complete, then use the keyword information output by the NLP model as the keyword information corresponding to the natural language information;

If the keyword information output by the NLP model is incomplete, then complete the keyword information output by the NLP model according to the pre-stored operation and maintenance troubleshooting empirical rules;

The keyword information output by the supplementary and complete NLP model is used as the keyword information corresponding to the natural language information.
The troubleshooting method according to claim 5, wherein the pre-stored operation and maintenance troubleshooting empirical rules are acquired through the following steps:

After acquiring keyword information corresponding to the natural language information each time, saving the keyword information;

Determine the probability of each keyword combination according to several stored keyword information;

The probability of each keyword combination is used as a pre-stored rule of thumb for operation, maintenance and troubleshooting.
The troubleshooting method according to any one of claims 1 to 6, wherein the received natural language information used to characterize the troubleshooting intention of the network includes:

judging whether the natural language information used to characterize the troubleshooting intention of the network conforms to a preset semantic standard; wherein, the preset semantic standard includes inputting at least one business domain to which a network fault belongs;

If the natural language information used to characterize the intention to troubleshoot the network meets the preset semantic standard, confirming the receipt of the natural language information used to characterize the intention to troubleshoot the network;

If the natural language information used to characterize the network troubleshooting intention does not meet the preset semantic standard, the natural language information used to characterize the network troubleshooting intention is discarded.
The troubleshooting method according to any one of claims 1 to 7, wherein the preset NLP model is obtained through iterative training based on a training corpus, and the training corpus includes any combination of the following: a system specification of the network, Instruction manuals and traditional troubleshooting interaction data for several of said applications.
A troubleshooting system, comprising: an acquisition module, an intent translation module, an execution module, an application module and a feedback module, wherein the intent translation module includes a preset natural language processing NLP model;

The obtaining module is used to obtain natural language information used to characterize the network troubleshooting intention, and send the natural language information to the intention translation module;

The intent translation module is used to input the natural language information into a preset natural language processing NLP model, obtain keyword information corresponding to the natural language information, and send the keyword information to the execution module ;

The execution module is used to call an application related to the keyword information to optimize the network;

The application module is used to accept calls from the execution module;

The feedback module is configured to generate first feedback information when it is detected that the optimized network does not have a network fault corresponding to the troubleshooting intention; wherein the first feedback information is used to indicate that the network fault has to be excluded.
An electronic device comprising:

at least one processor; and,

a memory communicatively coupled to the at least one processor; wherein,

The memory stores instructions executable by the at least one processor, the instructions are executed by the at least one processor, so that the at least one processor can perform the operation described in any one of claims 1 to 8 described troubleshooting method.
A computer-readable storage medium storing a computer program, which implements the troubleshooting method according to any one of claims 1 to 8 when the computer program is executed by a processor.