Information processing method and device, storage medium and program product

Abstract

The invention provides an information processing method and device, a storage medium and a program product, which can be applied to the technical field of artificial intelligence. The method comprises the steps that in response to a to-be-processed problem input by a user through target equipment, current equipment parameters of the target equipment and a target object and a processing intention of the to-be-processed problem are determined, the target object comprises components of the target equipment, and the current equipment parameters comprise current parameters of the target object in the target equipment; determining a standard parameter for the target object from a database corresponding to the target object based on the processing intention; determining an initial solution for the to-be-processed problem based on the current parameter and the standard parameter of the target object; performing structured processing on the initial solution, and determining processing flow information; and outputting processing flow information according to the identity information of the user.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and particularly to an information processing method, device, storage medium, and program product. Background Art

[0002] An expert system is a computer program system based on artificial intelligence technology. It aims to simulate the thinking process and professional knowledge of experts in a specific field to solve complex problems in that field or provide professional advice and decision-making support. A question-and-answer system implemented through keyword matching is a relatively simple expert system. The implementation logic of this kind of expert system is simple and the deployment cost is low. With the increasing demand for professionalism in different fields, a manually coded rule base can be set up and used to drive the reasoning process of the expert system.

[0003] In the process of implementing the present invention, it is found that there are at least the following problems in the related art. In the process of question-and-answer and reasoning using an expert system, the expert system is usually driven only by the question itself. The results output by the expert system are usually general but lack pertinence. Therefore, for some users who are not very familiar with the professional knowledge in this field, the output results are difficult to understand and apply, resulting in a poor user experience. Summary of the Invention

[0004] In view of the above problems, the present invention provides an information processing method, device, storage medium, and program product.

[0005] According to a first aspect of the present invention, there is provided an information processing method, including: in response to a problem to be processed input by a user via a target device, determining the current device parameters of the target device, the target object, and the processing intention of the problem to be processed, where the target object includes components of the target device, and the current device parameters include the current parameters of the target object in the target device; based on the processing intention, determining, from a database corresponding to the target object, standard parameters for the target object; based on the current parameters and the standard parameters of the target object, determining an initial solution to the problem to be processed; performing a structured process on the initial solution to determine process flow information; and outputting the process flow information according to the identity information of the user.

[0006] The second aspect of the present invention provides an information processing apparatus, including: a problem analysis module, configured to determine the current device parameters of the target device, the target object, and the processing intention of the problem to be processed in response to a problem to be processed input by a user via the target device, where the target object includes components of the target device, and the current device parameters include the current parameters of the target object in the target device; a parameter determination module, configured to determine standard parameters for the target object from a database corresponding to the target object based on the processing intention; a solution determination module, configured to determine an initial solution to the problem to be processed based on the current parameters and the standard parameters of the target object; a structured processing module, configured to perform structured processing on the initial solution to determine processing flow information; and an information output module, configured to output the processing flow information according to the identity information of the user.

[0007] The third aspect of the present invention provides an electronic device, including: one or more processors; a memory, configured to store one or more computer programs, where the above-mentioned one or more processors execute the above-mentioned one or more computer programs to implement the steps of the above-mentioned method.

[0008] The fourth aspect of the present invention further provides a computer-readable storage medium, on which a computer program or instruction is stored, and when the computer program or instruction is executed by a processor, the steps of the above-mentioned method are implemented.

[0009] The fifth aspect of the present invention further provides a computer program product, including a computer program or instruction, and when the computer program or instruction is executed by a processor, the steps of the above-mentioned method are implemented. Description of the Drawings

[0010] Through the following description of the embodiments of the present invention with reference to the drawings, the above-mentioned content and other objects, features, and advantages of the present invention will become clearer. In the drawings:

[0011] Figure 1 The application scenario diagram of the information processing method, device, storage medium, and program product according to the embodiment of the present invention is shown.

[0012] Figure 2 The flowchart of the information processing method according to the embodiment of the present invention is shown.

[0013] Figure 3 The flowchart of constructing a database corresponding to a certain object according to the embodiment of the present invention is shown.

[0014] Figure 4 The flowchart of the information processing method according to another embodiment of the present invention is shown.

[0015] Figure 5 The structural block diagram of the information processing apparatus according to the embodiment of the present invention is shown.

[0016] Figure 6 A block diagram of an electronic device suitable for implementing an information processing method according to an embodiment of the present invention is shown. Detailed implementation manners

[0017] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are merely exemplary and are not intended to limit the scope of the present invention. In the following detailed description, for the sake of explanation, many specific details are set forth in order to provide a comprehensive understanding of the embodiments of the present invention. However, obviously, one or more embodiments can also be implemented without these specific details. In addition, in the following description, descriptions of well-known structures and technologies are omitted to avoid unnecessarily confusing the concepts of the present invention.

[0018] The terms used herein are only for describing specific embodiments and are not intended to limit the present invention. The terms "including", "comprising", etc. used herein indicate the presence of the described features, steps, operations, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, or components.

[0019] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification and should not be interpreted in an idealized or overly rigid manner.

[0020] In the case of using expressions such as "at least one of A, B, and C", generally, it should be interpreted according to the meaning commonly understood by those skilled in the art (for example, "a system having at least one of A, B, and C" should include, but is not limited to, a system having only A, only B, only C, having A and B, having A and C, having B and C, and / or having A, B, and C, etc.).

[0021] In the technical solution of the present invention, the user information involved (including but not limited to user personal information, user image information, user device information, such as location information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) are all information and data authorized by the user or fully authorized by all parties. And the processing of relevant data, such as collection, storage, use, processing, transmission, provision, invention, and application, etc., all comply with relevant laws, regulations, and standards, take necessary confidentiality measures, do not violate public order and good customs, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0022] An embodiment of the present invention provides an information processing method, including: in response to a problem to be processed input by a user via a target device, determining the current device parameters of the target device, the target object of the problem to be processed, and the processing intention, where the target object includes components of the target device, and the current device parameters include the current parameters of the target object in the target device; based on the processing intention, determining, from a database corresponding to the target object, standard parameters for the target object; based on the current parameters and the standard parameters of the target object, determining an initial solution to the problem to be processed; performing a structured process on the initial solution to determine process flow information; and outputting the process flow information according to the identity information of the user.

[0023] Figure 1 FIG. shows an application scenario diagram of an information processing method, device, storage medium, and program product according to an embodiment of the present invention.

[0024] As Figure 1 shown, the application scenario 100 according to this embodiment may include a first terminal device 101, a second terminal device 102, a third terminal device 103, a network 104, and a server 105. The network 104 is used to provide a medium for communication links between the first terminal device 101, the second terminal device 102, the third terminal device 103, and the server 105. The network 104 may include various connection types, such as wired, wireless communication links, or fiber optic cables, etc.

[0025] Users can use the first terminal device 101, the second terminal device 102, and the third terminal device 103 to interact with the server 105 through the network 104 to receive or send messages, etc. Various communication client applications, such as Q&A application software (only for example), may be installed on the first terminal device 101, the second terminal device 102, and the third terminal device 103.

[0026] The first terminal device 101, the second terminal device 102, and the third terminal device 103 may be various electronic devices with a display screen and supporting web browsing, including but not limited to smartphones, tablets, laptop computers, and desktop computers, etc.

[0027] The server 105 may be a server providing various services, such as a background management server (only for example) that supports websites browsed by users using the first terminal device 101, the second terminal device 102, and the third terminal device 103. The background management server may analyze and process data such as received user requests, and feedback the processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal device.

[0028] It should be noted that the information processing method provided by the embodiments of the present invention can generally be executed by the server 105. Correspondingly, the information processing device provided by the embodiments of the present invention can generally be set in the server 105. The information processing method provided by the embodiments of the present invention can also be executed by a server or a server cluster different from the server 105 and capable of communicating with the first terminal device 101, the second terminal device 102, the third terminal device 103, and / or the server 105. Correspondingly, the information processing device provided by the embodiments of the present invention can also be set in a server or a server cluster different from the server 105 and capable of communicating with the first terminal device 101, the second terminal device 102, the third terminal device 103, and / or the server 105.

[0029] It should be understood that Figure 1 the numbers of the terminal devices, networks, and servers in

[0030] are merely illustrative. According to the implementation requirements, there can be any number of terminal devices, networks, and servers. Figure 1 are merely illustrative. According to the implementation requirements, there can be any number of terminal devices, networks, and servers. Figures 2 to 4 The following will be based on the

[0031] Figure 2 scenario described below, and the information processing method of the embodiments of the invention will be described in detail through

[0032] As Figure 2 shown, the information processing of this embodiment includes operation S210 to operation S250.

[0033] In operation S210, in response to a problem to be processed input by the user via the target device, determine the current device parameters of the target device, the target object, and the processing intention of the problem to be processed.

[0034] According to the embodiments of the present invention, the target device can include various types of terminal devices or servers, and the problem to be processed input by the user can be directed to the target object in the target device. Among them, the target object can include components of the target device, and the components can include, for example, a Graphics Processing Unit (GPU), a hard disk, etc.

[0035] For example, when the user needs to configure the hardware of the user's personal computer to ensure that the hardware conditions of the personal computer meet the requirements for executing certain specific tasks, the problem to be processed submitted by the user can be: I want to develop using programming language A on this machine. Help me determine whether the GPU of this machine meets the requirements.

[0036] Users can submit multi-modal problems to be processed. For example, through text, voice, pictures including the text of the problem to be processed, etc. After receiving different modal problems to be processed, the problems to be processed can be converted into text form and subsequent processing can be performed on the problems to be processed in text form. For example, when the problem to be processed is voice, speech recognition methods can be used to preprocess the voice and convert it into a problem to be processed in text form. When the problem to be processed is a picture including the text of the problem to be processed, the text in the picture can be accurately extracted through Optical Character Recognition (OCR) to determine the problem to be processed in text form.

[0037] According to an embodiment of the present invention, after receiving a problem to be processed input by a user, the problem to be processed can be parsed to determine the target object the problem to be processed is directed to and the processing intention of the problem to be processed. Among them, the target object can be determined by keyword matching of the problem to be processed, and the processing intention can be determined by natural language processing.

[0038] Taking the problem to be processed "I want to develop using programming language A on this machine. Help me determine whether the GPU of this machine meets the requirements" as an example, the problem to be processed can be matched with the words of multiple components respectively. If it is determined that the keyword "GPU" exists in the problem to be processed, it can be determined that the target object of this problem to be processed is the GPU of the target device. Performing natural language processing on the problem to be processed, it can be determined that the user needs to develop using programming language A on the target device. Therefore, it is necessary to set up a running environment for programming language A. Therefore, the processing intention can be to determine whether the GPU of the target device can meet the requirements for setting up the running environment of programming language A.

[0039] After determining the target component, the current device parameters of the target device can be obtained. Among them, the current device parameters include the current parameters of the target object in the target device, and the current parameters of the target object can be obtained through means such as command line and device manager. For example, when the target object is a GPU, the current values of configuration items such as the video memory capacity, number of stream processors, video memory bit width, bandwidth, brand, and model of the GPU in the target device can be obtained, that is, the current parameters.

[0040] In operation S220, based on the processing intention, the standard parameters for the target object are determined from the database corresponding to the target object.

[0041] According to an embodiment of the present invention, the database corresponding to the target object may include the conditions that the target object needs to meet to satisfy different processing intentions, that is, the standard parameters. The database corresponding to the target object can be a key-value type database, where the key is the processing intention and the value is the conditions that need to be met to satisfy this processing intention.

[0042] Since the database corresponds to the target object and the types of parameters involved in the target object are limited, the type of the database can be a structured database. Each column in this database corresponds to a parameter of the target object, and the value of each column is the standard parameter of a configuration item of the target object.

[0043] For example, if the target object is a GPU, the commonly used parameter types of the GPU include six configuration items: video memory capacity, number of stream processors, video memory bit width, bandwidth, brand, and model. Then the database can include 7 columns. For each row of records, the first column is the processing intention of the current record, and the subsequent six columns respectively correspond to a configuration item.

[0044] In operation S230, based on the current parameters and standard parameters of the target object, an initial solution to the problem to be processed is determined.

[0045] According to an embodiment of the present invention, by comparing the current parameters and the standard parameters, it can be determined whether the target object meets the processing intention according to the comparison result.

[0046] In the case where the standard parameter only includes one parameter, the current parameter corresponding to the target object can be directly compared with the standard parameter, and it can be determined whether the target object meets the processing intention according to the comparison result. In the case where the standard parameter includes multiple parameters, multiple parameters of the target object can be respectively compared with their corresponding standard parameters. In the case where multiple parameters all meet the conditions of the standard parameters, it can be determined that the target object meets the processing intention. In the case where at least one of the multiple parameters does not meet the conditions of the standard parameters, it can be determined that the target object does not meet the processing intention.

[0047] In the case where it is determined that the target object does not meet the processing intention, an initial solution can be determined according to the current parameters and the standard parameters. Among them, the initial solution can be used to adjust the current parameters so that the adjusted current parameters meet the standard parameters, that is, so that the adjusted target object can meet the processing intention.

[0048] In the case where it is determined that the target object meets the processing intention, a prompt can be directly output to inform the user that the target object can meet the processing intention without re-adjustment or configuration.

[0049] In operation S240, the initial solution is structurally processed to determine the processing flow information.

[0050] Since the hardware conditions of the target object not meeting the standards or the configuration error of the target object will both result in the target object not meeting the processing intention, it is necessary to make a preliminary judgment on the initial solution. When it is determined that the initial solution indicates a configuration error of the target object, the initial solution can be structured to obtain at least one operation step, and the processing flow information can be determined according to the at least one operation step.

[0051] According to an embodiment of the present invention, when making a preliminary judgment on the initial solution and determining that the initial solution indicates that the hardware conditions of the target object do not meet the standards, a hardware replacement suggestion can be directly output. The hardware replacement suggestion can include the parameter suggested values of the object to be replaced, and can also include information such as the model and brand of the object recommended to be replaced.

[0052] In operation S250, according to the user's identity information, the processing flow information is output.

[0053] According to an embodiment of the present invention, the user's identity information can include information such as user permissions. According to the identity information of different users, personalized processing flow information is output.

[0054] In the case where the user has a higher permission and has the modification permission for all configurations of the target object, all the processing flow information can be output, providing the processing objectives and processing methods of each operation step. While in the case where the user has a lower permission and does not have the modification permission for some configurations, the operation steps in the processing flow for which the user does not have the modification permission can be desensitized, and only the processing objective of the operation step is output, without providing the processing method.

[0055] According to an embodiment of the present invention, when the problem to be processed input by the user via the target device is obtained, the problem to be processed is parsed to accurately determine the user's processing intention and the target object targeted by the problem to be processed, and the current parameters of the target object in the target device are obtained, providing data support for subsequent comparison. Based on the processing intention, the standard parameters are determined from the target object database, and the standard parameters are compared with the current parameters, which can accurately judge the target device and the target object to determine whether they meet the user's processing intention. In the case of not meeting the processing intention, the direction of optimization and adjustment can be accurately determined according to the difference between the current parameters and the standard parameters, and the initial solution can be determined. Structuring the initial solution can disassemble the initial solution to obtain more specific and highly executable processing flow information. According to the user's identity information, realizing the personalized service output of the processing flow information can improve the user experience and ensure the stable operation of the target device and the target object.

[0056] According to an embodiment of the present invention, the database includes standard parameters of multiple objects in the device, and the standard parameters are obtained by the following methods: obtaining structured knowledge of multiple objects from a structured data source, where the structured data source includes object documents of multiple objects; obtaining unstructured knowledge of multiple objects from an unstructured data source, where the unstructured data source includes maintenance logs of multiple objects and third-party documents for multiple objects; and processing at least one piece of structured knowledge and at least one piece of unstructured knowledge to obtain standard parameters.

[0057] According to an embodiment of the present invention, the structured data source may include official documents such as operation manuals of multiple objects and object documents provided by object manufacturers. The unstructured data source may include maintenance logs of multiple objects and third-party documents for multiple objects, and the third-party documents may include documents about multiple objects in online communities and online forums.

[0058] Since the description methods and terms in official documents are usually relatively rigorous and standardized, while third-party documents usually use a more colloquial and easy-to-understand way to describe, there will be a situation where the actual referents in the structured data source and the unstructured data source are the same but the description methods are different. In this case, storing multiple description methods will result in data redundancy in the database and waste storage space.

[0059] Therefore, the structured knowledge obtained from the structured data source can be used as a processing standard to unify the description methods and terms in the unstructured knowledge, thereby avoiding having more than one record with the same semantics in the database and improving the utilization rate of the database storage space. In addition, constructing the database using heterogeneous data obtained from the structured data source and the unstructured data source increases the richness of the database and facilitates dealing with various problems to be processed.

[0060] According to an embodiment of the present invention, processing at least one piece of structured knowledge and at least one piece of unstructured knowledge to obtain standard parameters includes: respectively performing semantic recognition on the structured knowledge and the unstructured knowledge to obtain structured semantics and unstructured semantics; modifying the description method of the target unstructured knowledge based on the description method of the target structured knowledge to obtain target knowledge, where the structured semantics of the target structured knowledge is the same as the unstructured semantics of the target unstructured knowledge, but the description methods are different; performing natural language processing on the target knowledge to obtain configuration items of the target knowledge and corresponding standard values; and determining standard parameters based on the configuration items and the corresponding standard values.

[0061] According to an embodiment of the present invention, at least one structured knowledge and at least one unstructured knowledge are respectively subjected to semantic recognition to obtain the structured semantics of each of the at least one structured knowledge and at least one unstructured semantics. The structured semantics and the unstructured semantics are matched. For each unstructured semantics, from the at least one structured semantics, the structured semantics with the same semantics as it is determined, and the structured semantics and the unstructured semantics with the same semantics are determined as the target structured semantics and the target unstructured semantics.

[0062] According to multiple pairs of target structured semantics and target unstructured semantics with the same semantics, it is determined whether the description methods in the target structured knowledge corresponding to the target structured semantics and the target unstructured knowledge corresponding to the target unstructured semantics are the same.

[0063] Among them, by parsing the target structured knowledge and the target unstructured knowledge with the same semantics, the terms with the same semantics between the two can be determined, and it is compared whether the structured terms in the target structured knowledge are the same as the unstructured terms in the target unstructured knowledge. If it is determined that the structured terms are the same as the unstructured terms, it can be determined that the description methods of the target structured knowledge and the target unstructured knowledge are the same. If it is determined that the structured terms are different from the unstructured terms, it can be determined that the description methods of the target structured knowledge and the target unstructured knowledge are different.

[0064] In the case where it is determined that the description methods of the target structured knowledge and the target unstructured knowledge are different, the description method of the target unstructured knowledge is modified by using the description method of the target structured knowledge, so as to ensure that the same semantics uses the same description method.

[0065] Although after the description method is modified, it can be determined that the obtained target knowledge has unified the description methods in the structured knowledge and the unstructured knowledge, but since the target knowledge is natural language, it is difficult to directly convert it into a form that can be stored in a structured database.

[0066] For a target knowledge of each object, multiple preset configuration items of the object can be respectively used as keywords to perform keyword matching on the target knowledge, and the standard value corresponding to the configuration item in the target knowledge can be determined, and then the standard parameters can be determined.

[0067] According to an embodiment of the present invention, by unifying the description methods in the structured knowledge and the unstructured knowledge, the standardized representation of knowledge can be realized, so as to ensure that in the output content of the subsequent expert model, it is more unified, more in line with the description method of the structured knowledge, and more professional.

[0068] According to an embodiment of the present invention, determining an initial solution to a problem to be processed based on the current parameters of a target object and the standard parameters of the target object includes: inputting the current parameters of the target object and the standard parameters of the target object into an expert model to obtain an initial solution; wherein, the expert model is obtained in the following manner: inputting sample data into an initial expert model to obtain a sample output result, the sample data including a sample problem from a sample device, the parameters of the sample device, and a sample solution as a sample label; determining a reward value for the sample output result based on the first solution situation of the sample problem by the sample device using the sample output result, the processing time of the sample problem, and the presence of sensitive operations in the sample output result; adjusting the model parameters of the initial expert model when it is determined that the reward value is lower than a reward threshold or the loss value obtained based on the sample output result and the sample label is greater than a first loss threshold; and obtaining an expert model when it is determined that the reward value is higher than the reward threshold and the loss value is less than the first loss threshold.

[0069] According to an embodiment of the present invention, after receiving the current parameters of the target object and the standard parameters of the target object, the expert model can analyze the current parameters and the standard parameters, determine the reason why the current target object does not meet the processing intention, and further determine an adjustment strategy for the target object according to the reason, that is, the initial solution.

[0070] According to an embodiment of the present invention, the expert model can be obtained through pre-training. During the training process, an initial expert model is set, and the initial expert model is used to process sample data to obtain a sample output result. The sample data includes a sample problem input by a user via a sample device and the parameters of the sample device. The sample solution is the label of the sample data, and the sample solution can include an operation plan for solving the sample problem, and a better solution for the sample problem.

[0071] After obtaining the sample output result, configure and adjust the sample device according to the sample output result, and determine the first solution situation of the adjusted sample device for the sample problem, that is, determine whether the adjusted sample device meets the sample processing intention determined according to the sample problem.

[0072] The processing time of the sample problem can be determined according to the time interval between inputting the sample data into the initial expert model and the initial expert model outputting the sample output result.

[0073] When the sample output result indicates that the sample device needs to be reconfigured, determine the presence of sensitive operations in the sample output result.

[0074] According to the above first solution situation, the processing time of the sample problem, and the existence of sensitive operations in the sample output result, the reward value of the sample output result can be determined, and whether to further adjust the model can be determined according to the reward value and the loss value between the sample output result and the sample label.

[0075] The reward value can reflect the solution situation and efficiency of the current initial expert model for the sample problem. Therefore, the higher the reward value, the better the solution situation and the higher the solution efficiency. Thus, when the reward value is lower than the reward threshold, it indicates that the initial expert model still needs to be adjusted. The loss value determined based on the sample output result and the corresponding sample label can be used to represent the solution ability of the initial expert model for the sample problem. Therefore, the lower the loss value, the stronger the solution ability, and the closer the sample output result obtained using the initial expert model is to the sample label.

[0076] According to an embodiment of the present invention, when it is determined that the reward value is lower than the reward threshold or the loss value obtained based on the sample output result and the sample label is greater than the first loss threshold, it can be determined that the current initial expert model has a poor solution ability or low solution efficiency for the sample problem. Therefore, the model parameters of the initial expert model can be adjusted. When it is determined that the reward value is higher than the reward threshold and the loss value is less than the first loss threshold, it can be determined that the current initial expert model has a strong solution ability and high solution efficiency for the sample problem. Therefore, the expert model can be determined according to the model parameters of the current initial expert model.

[0077] According to an embodiment of the present invention, using the expert model to generate an initial solution can quickly and accurately provide an initial solution for the problem to be processed by virtue of the learning ability and summarization ability of the expert model for a large amount of knowledge and experience. During the training process of the expert model, determining the reward value based on multiple factors can comprehensively and objectively evaluate the output quality of the initial expert model. Adjusting the model parameters based on the reward value and the loss value can optimize the model according to the actual output effect, thereby improving the performance of the expert model.

[0078] Figure 3 The flowchart of constructing a database corresponding to a certain object according to an embodiment of the present invention is shown.

[0079] As Figure 3As shown, after modifying the description method of the unstructured data obtained from the unstructured data source according to the structured data obtained from the structured data source, and obtaining the target knowledge, knowledge extraction can be performed from the target knowledge, and the extracted target knowledge can be subjected to data cleaning and transformation to meet the field requirements of the databases corresponding to different objects. After the transformation is completed, the transformed data can be directly loaded into the database corresponding to the object by directly connecting to the database, thereby completing the construction of the database corresponding to the object. Among them, during the processes of data extraction, data cleaning and transformation, and data loading, breakpoints can be set arbitrarily, and the knowledge data after the breakpoint can be stored in the temporary database. After the subsequent processing of the knowledge data before the breakpoint is completed, the method of breakpoint resumption is used to obtain the knowledge data after the breakpoint from the temporary database and continue the data processing.

[0080] According to an embodiment of the present invention, the reward value includes an operation type reward value and a processing time reward value; the operation type reward value is determined according to whether the sample output result includes sensitive operations, and the processing time reward value is determined according to the average processing time of historical problems with the same problem type as the sample problem and the processing time of the sample output result.

[0081] According to an embodiment of the present invention, in the case where it is determined that the sample output result includes sensitive operations, it can be determined that the operation type reward value is low, and in the case where it is determined that the sample output result does not include sensitive operations, it can be determined that the operation type reward value is high. Among them, sensitive operations may include high-risk commands such as deleting configurations.

[0082] For example, in the case where it is determined that the sample output result includes sensitive operations, the operation type reward value can be taken as -10, and in the case where it is determined that the sample output result does not include sensitive operations, the operation type reward value can be taken as 0.

[0083] According to an embodiment of the present invention, determine the average processing time of historical problems with the same problem type as the sample problem , and in the case where it is determined that the processing time of the sample output result meets the preset conditions, it can be determined that the processing time reward value is high, and in the case where it is determined that the processing time of the sample output result does not meet the preset conditions, it can be determined that the processing time reward value is low.

[0084] For example, the preset condition may be , that is, in the case where the processing time of the sample output result is less than 80% of the average processing time of historical problems, it can be determined that the processing time reward value is taken as 0.5, and in the case where the processing time of the sample output result is not less than 80% of the average processing time of historical problems, it can be determined that the processing time reward value is taken as 0.

[0085] According to an embodiment of the present invention, the reward value may further include a processing result reward value, which is determined according to the solution situation of the sample problem after reconfiguring and adjusting the sample device according to the sample output result.

[0086] For example, after reconfiguring and adjusting the sample device according to the sample output result, when it is determined that the adjusted sample device can meet the sample processing intention determined according to the sample problem, the processing result reward value can be taken as 1; when it is determined that the adjusted sample device still cannot meet the sample processing intention determined according to the sample problem, the processing result reward value can be taken as 0.

[0087] According to an embodiment of the present invention, splitting the reward value into an operation type reward value and a processing time reward value can evaluate the sample output result from different key perspectives, making the reward value more comprehensive and more usable. Through the operation type reward value, reflecting whether the sample output result includes sensitive operations in the reward value can make the adjusted model obtained after adjusting the model parameters using the reward value more inclined to output sample output results that do not include sensitive operations, making the subsequent adjustment and optimization safer and more controllable. Through the processing time reward value, reflecting the output time of the sample output result in the reward value can prompt the model to optimize in the direction of higher output efficiency, thereby reducing the waiting time of users, optimizing the user experience, and improving the efficiency of information processing.

[0088] According to an embodiment of the present invention, the information processing method further includes: discarding the sample data when it is determined that the loss value is greater than the second loss threshold, where the second loss threshold is greater than the first loss threshold.

[0089] According to an embodiment of the present invention, when it is determined that the loss value is large, the gradient determined according to the loss value is large, and when the initial expert model is updated according to this gradient, the update amplitude of the model parameters is large, which may lead to the problem of unstable model functions. Therefore, a second loss threshold can be set. When it is determined that the loss value is greater than the second loss threshold, the current sample data is discarded to avoid excessive updates to the model parameters, thereby ensuring the stability of the model parameters and model functions. Preferably, the second loss threshold can be taken as 0.2.

[0090] According to an embodiment of the present invention, the information processing method further includes: determining the reward value of the initial solution according to the user's evaluation of the processing flow information, the second solution situation of the target device using the processing flow information for the problem to be processed, the processing time of the problem to be processed, and the presence of sensitive operations in the initial solution; and adjusting the model parameters of the expert model when it is determined that the reward value is lower than the reward threshold.

[0091] According to an embodiment of the present invention, during the use of the expert model, the reward value of the initial solution can be determined based on the processing result of the expert model for the problem to be processed, and whether the model parameters of the expert model need to be further adjusted can be determined based on the reward value. Among them, the determination process of the reward value during the application of the expert model is similar to the determination process of the reward value during the training of the expert model.

[0092] After obtaining the processing flow information, configure and adjust the target device according to the processing flow information, and determine the second solution situation of the adjusted target device for the problem to be processed, that is, determine whether the adjusted target device meets the processing intention determined according to the problem to be processed.

[0093] The processing time of the problem to be processed can be determined according to the time interval between when the problem to be processed is input to the expert model and when the expert model outputs the processing flow information.

[0094] In addition to the second solution situation, the processing time of the problem to be processed, and the existence of sensitive operations in the initial solution, the user feedback reward value can also be determined according to the user's evaluation of the processing flow information, and the reward value of the initial solution can be determined.

[0095] According to an embodiment of the present invention, after providing the processing flow information to the user, the score of the user for the processing flow information can be obtained, and based on the score of the processing flow information, the user feedback reward value R is determined 1 , as shown in formula (1):

[0096] (1)

[0097] Where rating is the score of the user for the processing flow information, and rating ∈ {1, 2, 3, 4, 5}.

[0098] After determining the user feedback reward value, the operation type reward value, the processing time reward value, and the processing result reward value, the reward value R of the initial solution can be determined using formula (2) total :

[0099] (2)

[0100] Where α, β, γ are hyperparameters, R 2 is the processing result reward value, R 3 is the processing time reward value, R 4 is the operation type reward value. Preferably, α = 0.4, β = 0.3, γ = 0.3.

[0101] According to an embodiment of the present invention, when it is determined that the reward value is lower than the reward threshold, it can be determined that the problem-solving efficiency of the expert model for the problem to be processed is low, or the user experience is poor, or there are potential safety hazards. Therefore, the model parameters of the expert model can be adjusted.

[0102] According to an embodiment of the present invention, in the process of applying the expert model, introducing the user feedback reward value can obtain the user's satisfaction from the perspective of the user experience. Since it is difficult to directly quantify the satisfaction of the user with the processing flow information, which can reflect various aspects of information such as the complexity, convenience, and whether it conforms to the user's personal operation habits of the processing flow information, by obtaining the user score, a quantitative evaluation of the user with respect to the processing flow information can be obtained, and thus combined with the reward values in other dimensions, it is possible to more comprehensively evaluate the processing flow information and the expert model and determine the subsequent adjustment plan.

[0103] According to another embodiment of the present invention, when it is determined that the reward value is lower than the reward threshold, it can be determined that the currently output processing flow information fails to meet the user's usage requirements. Therefore, the expert model can be further optimized to obtain optimized processing flow information by using the optimized expert model.

[0104] According to an embodiment of the present invention, the Generalized Advantage Estimation (GAE) method can be used to determine the advantage function, where the advantage function can be used to represent the degree of advantage of an action relative to the average policy in a certain state. In GAE, the information of multiple time steps is combined to estimate the advantage. Specifically, by performing a weighted sum of the rewards obtained after executing different processing flow information, the advantages and effects of different processing flow information can be determined more accurately. The advantage function can be determined by formula (3):

[0105] (3)

[0106] where t is the position of the current processing flow information among multiple processing flow information, θ is the discount factor, used to measure the importance of the execution result of the processing flow information, λ is a hyperparameter, used to control the balance between bias and variance, l is used to represent the offset of the time step length. Preferably, θ = 0.99 and λ = 0.95. is the TD error after the execution of the (t + l)-th processing flow information, where as shown in formula (4):

[0107] (4)

[0108] where r t+lThe reward value R of the initial solution corresponding to the (t + l)-th processing flow information total , and are the estimated values of the value functions of states s t and state s t+l respectively.

[0109] After determining the advantage value of the advantage function using the above formula (3), the advantage value can be multiplied by the gradient determined according to the loss value of the initial expert model to determine the update direction of the model parameters, thereby improving the optimization efficiency of the model.

[0110] According to an embodiment of the present invention, inputting the current parameters of the target object and the standard parameters of the target object into the expert model to obtain an initial solution includes: obtaining a judgment result according to whether the current parameters of the target object satisfy a judgment condition determined based on the target object and the standard parameters of the target object; and determining the initial solution based on the judgment result.

[0111] According to an embodiment of the present invention, based on the judgment condition determined based on the target object and the standard parameters of the target object, a judgment result of the current parameters of the target object can be obtained, where the judgment result is used to indicate whether the current parameters of the target object satisfy the judgment condition.

[0112] According to an embodiment of the present invention, after obtaining the judgment result, according to the judgment result, it can be determined how the current parameters of the target object need to be adjusted to meet the standard parameters, thereby obtaining the initial solution.

[0113] In one example, after completing the construction of the database corresponding to the target object, the standard parameters of the target object under at least one processing intention can be determined respectively, and a conditional judgment statement can be determined. During the information processing process, according to the processing intention of the problem to be processed, the corresponding conditional judgment statement can be selected from at least one conditional judgment statement, and the current parameters of the target object can be written into the conditional judgment statement to obtain the judgment result and the initial solution.

[0114] For example, for a Redundant Array of Independent Disks (RAID) configuration scenario, according to structured knowledge and unstructured knowledge, it can be determined that RAID5 requires at least 3 hard disks. Then the conditional judgment statement can be, for example, "if (*RAIDConfigRequest(type == "RAID5", disks.size() ≥ 3))", that is, to judge whether the number of hard disks is greater than or equal to 3.

[0115] Upon obtaining the current parameters of the target object of the target device, it is found that there are 2 hard disks in the current target device. The number of hard disks is used to assign a value to disks.size(), that is, disks.size() == 2. Then, a conditional judgment statement is executed. The output of the conditional judgment statement is false, and the branch corresponding to false is entered, and the judgment result can be obtained:

[0116] “$request : RAIDConfigRequest(type == "RAID5", disks.size() < 3)”. For this judgment result, subsequent actions can be triggered, that is:

[0117] “when $request : RAIDConfigRequest(type == "RAID5", disks.size() < 3) then throw new ValidationException("RAID5 requires at least 3 hard disks");”.

[0118] At this time, the initial solution obtained is the text content of the thrown exception "RAID5 requires at least 3 hard disks". According to the above knowledge of RAID, it can be determined that the initial solution output by the expert model for the current target device can solve the problem to be processed.

[0119] According to an embodiment of the present invention, when presenting the initial solution, it may further include the configuration items with exceptions and their corresponding current parameters in the current state of the target object. For example, the initial solution may be "RAID5 requires at least 3 hard disks, and there are 2 hard disks configured in the current host". The configuration items, current parameters, and standard parameters in the initial solution can also be highlighted to facilitate the user to more quickly judge the problems existing in the current target device.

[0120] According to an embodiment of the present invention, according to the database corresponding to the target object, the judgment conditions of the target object can be set. Based on the current parameters of the target object and the judgment conditions, the judgment result can be obtained intuitively and accurately. Based on the judgment result, the initial solution can be determined specifically according to the judgment result indicating an exception, thereby improving the determination efficiency of the initial solution and improving the accuracy and effectiveness of the initial solution.

[0121] According to an embodiment of the present invention, the initial solution is structurally processed to determine the processing flow information, including: decomposing the initial solution using the decomposition method for the initial solution to obtain at least one processing step, where the decomposition method is determined based on the object type of the target object; and determining the processing flow information based on the at least one processing step.

[0122] According to an embodiment of the present invention, according to the object category of different target objects, a decomposition method for the initial solution can be determined, where the decomposition method includes the decomposition granularity. When the object category of the target object indicates that the encapsulation of the target object is relatively strong, the decomposition method for this target object can be to decompose it according to a larger granularity. When the object category of the target object indicates that the encapsulation of the target object is relatively poor, the decomposition method for this target object can be to decompose it according to a smaller granularity.

[0123] In one example, when the target object is a hard disk, the encapsulation of the target object is relatively strong, so the decomposition granularity for the initial solution can be larger, that is, the processing steps in the provided processing flow information are relatively more macroscopic. For example, "increasing the number of hard disks", this operation usually needs to be achieved by directly adding hardware.

[0124] In another example, when the target object is a GPU, the encapsulation of the target object is relatively poor, so the decomposition granularity for the initial solution can be smaller, that is, the processing steps in the provided processing flow information are relatively more microscopic and underlying. For example, "changing the power management mode of the GPU", "changing the fan speed of the GPU", etc., and this operation can usually be achieved in the form of a command line.

[0125] According to an embodiment of the present invention, using the decomposition method for the initial solution, the initial solution can be decomposed to obtain at least one processing step, and the processing flow information can be determined.

[0126] According to an embodiment of the present invention, decomposing the initial solution to obtain at least one processing step and determining the processing flow information based on the processing steps can processify the initial solution, making the obtained processing flow information more operable and optimizing the user experience.

[0127] According to an embodiment of the present invention, the information processing method further includes: determining the average session depth of at least one historical question according to at least one historical question of the user, where the average session depth is determined according to the session depth of each of the at least one historical question, and the session depth is used to represent the number of follow-up questions of the user for the historical question; and determining an evaluation value of the user's proficiency in the target object based on the word frequency of the target keyword and the average session depth in at least one historical question as the user's identity information.

[0128] According to an embodiment of the present invention, since different users have different levels of understanding and knowledge backgrounds regarding the same target object, professional users with more professional knowledge can refer to and optimize the processing flow information to obtain a better processing method than the processing flow information. Such users expect to reduce basic operations in the early stage but retain the possibility of personalized operations. However, users with less professional knowledge usually have difficulty improving the processing flow information. Generally, such users expect to be able to directly complete the operations indicated by the processing flow information without having to operate themselves.

[0129] For users with different proficiency evaluation values, different output strategies can be applied to output the processing flow information. Therefore, the user's identity information can also include the user's proficiency evaluation value for the target object.

[0130] Under normal circumstances, when the user has a high proficiency in the target object, the processing flow information is relatively easy for the user to understand. Therefore, the user usually does not ask the same question multiple times in one conversation. However, when the user has a low proficiency in the target object, the processing flow information is relatively difficult for the user to understand, and there may be terms in the processing flow information that the user does not understand. Therefore, the user usually asks the same question multiple times.

[0131] Therefore, based on the number of times the user asks the same question, it is possible to judge the user's understanding and proficiency in this field and the target object. Since large language models such as expert models are usually in a question-and-answer interaction where the user inputs a question and the large language model outputs an answer, in the large language model, the session depth of the interaction between the user and the large language model for the same question can be used to represent the number of times the user asks the same question.

[0132] Based on at least one historical question of the user, the average session depth of the user for at least one historical question can be determined, where the historical question can be a question that the user has asked the expert model regarding the target object.

[0133] During the interaction between the user and the expert model, when it is necessary to use professional terms to describe the question, users with a higher proficiency in the target object are more inclined to use more professional terms, while users with a lower proficiency in the target object are more likely to use natural language to describe the term. Therefore, it is also possible to judge the user's proficiency in the target object based on the proportion of terms related to the target object in the questions submitted by the user to the expert model.

[0134] According to an embodiment of the present invention, related terms of a target object can be determined as target keywords according to the target object, the word frequency of the target keywords in at least one historical question of a user and the average conversation depth of at least one historical question of the user are obtained, and a proficiency evaluation value of the user for the target object is determined by a logistic regression method.

[0135] According to an embodiment of the present invention, based on the average conversation depth of the user for historical questions and the word frequency of the user using the target keywords, the understanding degree and proficiency of the user for the target object can be evaluated more accurately and comprehensively, making the proficiency evaluation value more reliable.

[0136] According to an embodiment of the present invention, according to the identity information of the user, process flow information is output, including: in the case where it is determined that the proficiency evaluation value represented by the identity information is greater than or equal to the proficiency threshold, editable commands corresponding to at least one processing step included in the process flow information are respectively generated, and the process flow information and the editable commands are sent to the user, so that the user can perform the operations indicated by the process flow information based on the editable commands; and in the case where it is determined that the proficiency evaluation value represented by the identity information is less than the proficiency threshold, an automated processing script generated based on the process flow information and the process flow information are sent to the user, so that the user can perform the operations indicated by the process flow information based on the automated processing script.

[0137] According to an embodiment of the present invention, a proficiency threshold can be set, users with a proficiency evaluation value higher than the proficiency threshold are determined as proficient users for the target object, and users with a proficiency evaluation value lower than the proficiency threshold are determined as novice users for the target object.

[0138] According to an embodiment of the present invention, in the case where it is determined that the user is a proficient user for the target object, editable commands corresponding to the processing steps can be respectively generated based on at least one processing step in the process flow information, at least one editable command is combined in the order of the corresponding processing steps, and the combined editable command group and the text content of the process flow information are sent to the user, so that the user can analyze and re-edit the editable command group with reference to the text content.

[0139] According to an embodiment of the present invention, in the case where it is determined that the user is a novice user for the target object, the combined editable command group can be encapsulated into an automated processing script, and the automated processing script and the process flow information are sent to the user, so that the user can perform the operations indicated by the process flow information automatically and without code based on the automated processing script.

[0140] According to an embodiment of the present invention, different strategies are adopted to output process information for users with different proficiency evaluation values. For proficient users with a higher proficiency evaluation value, an editable command group can be provided so that when the user has a need to re-edit or modify the editable command group, the user can directly edit the editable command group, improving the flexibility of processing the target object according to the process information. For novice users with a lower proficiency evaluation value, a relatively simple one-key processing solution can be provided through an automated processing script, reducing the operation threshold and improving the processing efficiency. Therefore, for users with different proficiency evaluation values, the processing efficiency can be improved and the user experience can be optimized.

[0141] In the case of determining that the user is a proficient user for the target object, providing the editable command group to the user can facilitate the user's re-editing and optimization. However, after the editable command group is provided to the user, if the user uses a command corresponding to a sensitive operation during the editing process, it will cause irreversible changes to the target object after running the edited editable command group.

[0142] Therefore, in the case of determining that the user is a proficient user for the target object, an automated restoration script can be generated based on the current configuration of the target object, and the automated restoration script, the editable command group, and the text content of the process information are sent to the user. After the user edits the editable command group and runs the edited editable command group, the target object can be restored to the state before running the edited editable command group by running the automated restoration script.

[0143] In addition, in the case of determining that the edited editable command group run by the user includes a sensitive operation and the user uses the automated restoration script to restore the target object, the proficiency evaluation value of the user for the target object can be updated to reduce its proficiency evaluation value, thereby avoiding the same situation from occurring again and ensuring the normal operation of the target device and target components and the efficiency and stability of information processing.

[0144] According to an embodiment of the present invention, in response to a problem to be processed input by the user via the target device, the current device parameters of the target device and the target object and processing intention of the problem to be processed are determined, including: performing natural language processing on the problem to be processed to determine the target object and processing intention of the problem to be processed; and collecting information on the target object in the target device to determine the current parameters of the target object.

[0145] According to an embodiment of the present invention, by performing natural language processing on the problem to be processed, the target object targeted by the problem to be processed can be determined through keyword matching, and the processing intention of the user submitting the problem to be processed can be determined through semantic recognition. According to the target object, targeted information collection can be performed on the target object in the target device, so as to determine the current parameters of the target object.

[0146] Figure 4 The flowchart of an information processing method according to another embodiment of the present invention is shown.

[0147] As Figure 4 shown, the method includes operation S401 to operation S412.

[0148] In operation S401, collect the problem to be processed of the user.

[0149] In operation S402, parse the problem to be processed to determine the target object of the problem to be processed and the processing intention of the user.

[0150] In operation S403, collect the current parameters of the target object from the target device.

[0151] In operation S404, obtain the standard parameters of the target object from the database corresponding to the target object.

[0152] In operation S405, compare the current parameters with the standard parameters to determine the initial solution.

[0153] In operation S406, perform structured processing on the initial solution to obtain the processing flow information.

[0154] In operation S407, determine whether the identity information of the user has been determined. If so, execute operation S408; if not, execute operation S409.

[0155] In operation S408, obtain the identity information of the user.

[0156] In operation S409, determine the identity information of the user based on the user's historical problems.

[0157] In operation S410, output the processing flow information to the user according to the identity information of the user.

[0158] In operation S411, calculate the reward value in response to the user's completion of executing the processing flow information.

[0159] In operation S412, determine whether to update the database corresponding to the target object according to the reward value.

[0160] Taking the user's submitted problem to be processed "How to configure RAID5 for the local hard drive?" as an example, first perform natural language processing on the problem to be processed, extract the user's processing intention as RAID configuration, and the target object as the hard drive. Then obtain the current parameters of the hard drive from the target device, that is, the host where the user submitted the problem to be processed, and determine that the number of hard drives is 4 and the health status of the hard drive has a health warning.

[0161] Using RAID configuration as a keyword or index, perform a search in the database corresponding to the hard drive to obtain the standard parameters of the hard drive during RAID configuration, including the number of hard drives ≥ 3 and no health warning for the hard drive. Compare the current parameters with the standard parameters, and it can be determined that the number of local hard drives meets the requirements of the standard parameters, but the health status of the hard drive does not meet the requirements of the standard parameters.

[0162] In this case, an initial solution can be obtained, such as "After detecting and repairing the health status of the hard drive, RAID5 can be configured". Structurally process the initial solution, and the obtained processing flow information can include: detecting the health status of the hard drive; repairing the hard drive according to the detection result; configuring RAID5.

[0163] Package the command for detecting the health status of the hard drive in the first step into an editable command and provide it to the user. After the user executes it, the hard drive failure can be detected. When performing the second step of repairing the hard drive according to the detection result, generate a corresponding editable command based on the detected hard drive failure and provide it to the user. After the user executes it, the hard drive failure can be repaired. For the third step of configuring RAID5, since it has been determined through detection that the number of hard drives meets the requirements of the standard parameters, and the health status of the hard drive has returned to normal after the first two steps of processing, therefore, the configuration process of RAID5 can be directly packaged into an editable command and provided to the user so that the user can complete the RAID5 configuration after execution.

[0164] Before sending the editable command to the user, the user's identity information can be judged. In the case of determining that the user is a proficient user, send the editable command to the user. In the case of determining that the user is a novice user, package the editable command into an executable automated processing script and provide the script to the user.

[0165] After completing the configuration process, calculate the reward value during the user's current processing process, and determine whether it is necessary to update the database corresponding to the target object according to the reward value. For example, during the user's execution of the configuration, the data in the hard drive was not backed up, resulting in data loss. Then the data in the hard drive can be restored using the automated restoration script, and the reward value R can be determined. 2=-10. Update the database corresponding to the target object so that in the case of subsequent RAID5 configuration, a prompt is added to the initial solution or processing flow information output to remind the user to back up the hard disk data.

[0166] Based on the above information processing method, the present invention also provides an information processing device. The following will be combined with Figure 5 to describe this device in detail.

[0167] Figure 5 The structural block diagram of the information processing device according to an embodiment of the present invention is shown.

[0168] As Figure 5 shown, the information processing device 500 of this embodiment includes a problem analysis module 510, a parameter determination module 520, a solution determination module 530, a structured processing module 540, and an information output module 550.

[0169] The problem analysis module 510 is used to determine the current device parameters of the target device, the target object and the processing intention of the problem to be processed in response to the problem to be processed input by the user via the target device. Among them, the target object includes the components of the target device, and the current device parameters include the current parameters of the target object in the target device. In one embodiment, the problem analysis module 510 can be used to execute the operation S210 described above, which will not be elaborated here.

[0170] The parameter determination module 520 is used to determine the standard parameters for the target object from the database corresponding to the target object based on the processing intention. In one embodiment, the parameter determination module 520 can be used to execute the operation S220 described above, which will not be elaborated here.

[0171] The solution determination module 530 is used to determine the initial solution to the problem to be processed based on the current parameters and standard parameters of the target object. In one embodiment, the solution determination module 530 can be used to execute the operation S230 described above, which will not be elaborated here.

[0172] The structured processing module 540 is used to perform structured processing on the initial solution to determine the processing flow information. In one embodiment, the structured processing module 540 can be used to execute the operation S240 described above, which will not be elaborated here.

[0173] The information output module 550 is used to output the processing flow information according to the user's identity information. In one embodiment, the information output module 550 can be used to execute the operation S250 described above, which will not be elaborated here.

[0174] According to an embodiment of the present invention, the information processing apparatus 500 further includes a first knowledge acquisition module, a second knowledge acquisition module, and a knowledge processing module.

[0175] The first knowledge acquisition module is configured to acquire structured knowledge of multiple objects from a structured data source, where the structured data source includes object documents of multiple objects.

[0176] The second knowledge acquisition module is configured to acquire unstructured knowledge of multiple objects from an unstructured data source, where the unstructured data source includes maintenance logs of multiple objects and third-party documents for multiple objects.

[0177] The knowledge processing module is configured to process at least one piece of structured knowledge and at least one piece of unstructured knowledge to obtain standard parameters.

[0178] According to an embodiment of the present invention, the knowledge processing module includes a semantic recognition sub-module, a knowledge processing sub-module, a knowledge normalization sub-module, and a parameter determination sub-module.

[0179] The semantic recognition sub-module is configured to perform semantic recognition on the structured knowledge and the unstructured knowledge respectively to obtain structured semantics and unstructured semantics.

[0180] The knowledge processing sub-module is configured to modify the description manner of the target unstructured knowledge based on the description manner of the target structured knowledge to obtain target knowledge, where the structured semantics of the target structured knowledge is the same as the unstructured semantics of the target unstructured knowledge, and the description manners are different.

[0181] The knowledge normalization sub-module is configured to perform natural language processing on the target knowledge to obtain configuration items of the target knowledge and corresponding standard values.

[0182] The parameter determination sub-module is configured to determine standard parameters based on the configuration items and the corresponding standard values.

[0183] According to an embodiment of the present invention, the solution determination module 530 includes a solution determination sub-module.

[0184] The solution determination sub-module is configured to input the current parameters of the target object and the standard parameters of the target object into an expert model to obtain an initial solution.

[0185] According to an embodiment of the present invention, the information processing apparatus 500 further includes a sample input module, a first reward determination module, a first parameter adjustment module, and a model determination module.

[0186] The sample input module is configured to input sample data into the initial expert model to obtain a sample output result, where the sample data includes sample problems from sample devices, parameters of the sample devices, and sample solutions as sample labels.

[0187] The first reward determination module is configured to determine the reward value of the sample output result based on the first solution situation of the sample problem by the sample device using the sample output result, the processing time of the sample problem, and the presence of sensitive operations in the sample output result.

[0188] The first parameter adjustment module is configured to adjust the model parameters of the initial expert model when it is determined that the reward value is lower than the reward threshold or the loss value obtained based on the sample output result and the sample label is greater than the first loss threshold.

[0189] The model determination module is configured to obtain the expert model when it is determined that the reward value is higher than the reward threshold and the loss value is less than the first loss threshold.

[0190] According to an embodiment of the present invention, the information processing device 500 further includes a sample discarding module.

[0191] The sample discarding module is configured to discard the sample data when it is determined that the loss value is greater than the second loss threshold, and the second loss threshold is greater than the first loss threshold.

[0192] According to an embodiment of the present invention, the information processing device 500 further includes a second reward determination module and a second parameter adjustment module.

[0193] The second reward determination module is configured to determine the reward value of the initial solution based on the user's evaluation of the processing flow information, the second solution situation of the problem to be processed by the target device using the processing flow information, the processing time of the problem to be processed, and the presence of sensitive operations in the initial solution.

[0194] The second parameter adjustment module is configured to adjust the model parameters of the expert model when it is determined that the reward value is lower than the reward threshold.

[0195] According to an embodiment of the present invention, the solution determination sub-module includes a result determination unit and a solution determination unit.

[0196] The result determination unit is configured to obtain a judgment result according to whether the current parameters of the target object meet the judgment conditions determined based on the target object and the standard parameters of the target object.

[0197] The solution determination unit is configured to determine an initial solution based on the judgment result.

[0198] According to an embodiment of the present invention, the structured processing module 540 includes a solution decomposition sub-module and a process determination sub-module.

[0199] The solution decomposition sub-module is configured to decompose the initial solution using a decomposition method for the initial solution to obtain at least one processing step, and the decomposition method is determined based on the object type of the target object.

[0200] A process determination sub-module, configured to determine process information based on at least one processing step.

[0201] According to an embodiment of the present invention, the information processing device 500 further includes a depth determination module and an identity determination module.

[0202] The depth determination module is configured to determine an average session depth of at least one historical question according to at least one historical question of a user, wherein the average session depth is determined according to the session depth of each of the at least one historical question, and the session depth is used to represent the number of follow-up questions of the user for the historical question.

[0203] The identity determination module is configured to determine a proficiency evaluation value of the user for the target object based on the word frequency of the target keyword and the average session depth in at least one historical question, and use it as the identity information of the user.

[0204] According to an embodiment of the present invention, the information output module 550 includes a first information output sub-module and a second information output sub-module.

[0205] The first information output sub-module is configured to, when it is determined that the proficiency evaluation value represented by the identity information is greater than or equal to the proficiency threshold, generate editable commands corresponding to at least one processing step included in the process information respectively, and send the process information and the editable commands to the user, so that the user can perform operations indicated by the process information based on the editable commands.

[0206] The second information output sub-module is configured to, when it is determined that the proficiency evaluation value represented by the identity information is less than the proficiency threshold, send an automated processing script generated based on the process information and the process information to the user, so that the user can perform operations indicated by the process information based on the automated processing script.

[0207] According to an embodiment of the present invention, the problem parsing module 510 includes a problem parsing sub-module and an information collection sub-module.

[0208] The problem parsing sub-module is configured to perform natural language processing on the problem to be processed, and determine the target object and the processing intention of the problem to be processed.

[0209] The information collection sub-module is configured to collect information about the target object in the target device and determine the current parameters of the target object.

[0210] According to an embodiment of the present invention, any plurality of modules among the problem analysis module 510, the parameter determination module 520, the solution determination module 530, the structured processing module 540, and the information output module 550 may be combined and implemented in one module, or any one of them may be split into multiple modules. Alternatively, at least part of the functions of one or more of these modules may be combined with at least part of the functions of other modules and implemented in one module. According to an embodiment of the present invention, at least one of the problem analysis module 510, the parameter determination module 520, the solution determination module 530, the structured processing module 540, and the information output module 550 may be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on chip, a system on a substrate, a system in a package, an application specific integrated circuit (ASIC), or any other reasonable manner that can be achieved by integrating or packaging circuits, etc., in hardware or firmware, or implemented in any one of the three implementation manners of software, hardware, and firmware, or in an appropriate combination of any several of them. Alternatively, at least one of the problem analysis module 510, the parameter determination module 520, the solution determination module 530, the structured processing module 540, and the information output module 550 may be at least partially implemented as a computer program module, and when the computer program module is run, it can execute the corresponding functions.

[0211] Figure 6 FIG. shows a block diagram of an electronic device suitable for implementing the information processing method according to an embodiment of the present invention.

[0212] As Figure 6 shown, the electronic device 600 according to an embodiment of the present invention includes a processor 601, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 602 or the program loaded from the storage section 608 into the random access memory (RAM) 603. The processor 601 may include, for example, a general microprocessor (such as a CPU), an instruction set processor, and / or a related chipset, and / or a dedicated microprocessor (such as an application specific integrated circuit (ASIC)), etc. The processor 601 may also include on-board memory for caching purposes. The processor 601 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of the present invention.

[0213] In the RAM 603, various programs and data required for the operation of the electronic device 600 are stored. The processor 601, the ROM 602, and the RAM 603 are connected to each other via the bus 604. The processor 601 performs various operations of the method flow according to the embodiments of the present invention by executing the programs in the ROM 602 and / or the RAM 603. It should be noted that the programs may also be stored in one or more memories other than the ROM 602 and the RAM 603. The processor 601 may also perform various operations of the method flow according to the embodiments of the present invention by executing the programs stored in the one or more memories.

[0214] According to an embodiment of the present invention, the electronic device 600 may further include an input / output (I / O) interface 605, and the input / output (I / O) interface 605 is also connected to the bus 604. The electronic device 600 may further include one or more of the following components connected to the input / output (I / O) interface 605: an input part 606 including a keyboard, a mouse, etc.; an output part 607 including, for example, a cathode ray tube (CRT), a liquid crystal display (LCD), etc. and a speaker, etc.; a storage part 608 including a hard disk, etc.; and a communication part 609 including a network interface card such as a LAN card, a modem, etc. The communication part 609 performs communication processing via a network such as the Internet. The drive 610 is also connected to the input / output (I / O) interface 605 as needed. A removable medium 611, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 610 as needed so that a computer program read from it can be installed into the storage part 608 as needed.

[0215] The present invention also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments; or may exist separately without being assembled into the device / apparatus / system. The above computer-readable storage medium carries one or more programs, and when the one or more programs are executed, the method according to the embodiments of the present invention is implemented.

[0216] According to an embodiment of the present invention, the computer-readable storage medium may be a non-volatile computer-readable storage medium, for example, it may include but is not limited to: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the above. In the present invention, the computer-readable storage medium may be any tangible medium that contains or stores a program, and this program can be used by or in conjunction with an instruction execution system, apparatus, or device. For example, according to an embodiment of the present invention, the computer-readable storage medium may include the ROM 602 and / or RAM 603 described above and / or one or more memories other than the ROM 602 and RAM 603.

[0217] An embodiment of the present invention further includes a computer program product, which includes a computer program that contains program code for executing the method shown in the flowchart. When the computer program product runs in a computer system, the program code is used to enable the computer system to implement the method provided by the embodiment of the present invention.

[0218] When the computer program is executed by the processor 601, it executes the above functions defined in the system / apparatus of the embodiment of the present invention. According to an embodiment of the present invention, the above-described systems, apparatuses, modules, units, etc. can be implemented by computer program modules.

[0219] In one embodiment, the computer program can rely on tangible storage media such as optical storage devices and magnetic storage devices. In another embodiment, the computer program can also be transmitted and distributed in the form of a signal on a network medium, and be downloaded and installed through the communication part 609, and / or be installed from the removable medium 611. The program code contained in the computer program can be transmitted by any suitable network medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0220] In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 609, and / or be installed from the removable medium 611. When the computer program is executed by the processor 601, it executes the above functions defined in the system of the embodiment of the present invention. According to an embodiment of the present invention, the above-described systems, devices, apparatuses, modules, units, etc. can be implemented by computer program modules.

[0221] According to embodiments of the present invention, program code for executing the computer programs provided by the embodiments of the present invention can be written in any combination of one or more programming languages. Specifically, these computing programs can be implemented using high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages include, but are not limited to, such as Java, C++, Python, the "C" language, or similar programming languages. The program code can be executed entirely on the user's computing device, partially on the user's device, partially on a remote computing device, or entirely on a remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the user's computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., by connecting through the Internet using an Internet service provider).

[0222] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagram can represent a module, a program segment, or a part of code, and the above-mentioned module, program segment, or part of code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than that marked in the accompanying drawings. For example, two consecutive blocks shown may actually be executed substantially in parallel, and they can sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram or flowchart, and the combination of blocks in the block diagram or flowchart, can be implemented by a dedicated hardware-based system for performing the specified functions or operations, or can be implemented by a combination of dedicated hardware and computer instructions.

[0223] Those skilled in the art can understand that the features described in the various embodiments of the present invention can be combined and / or combined in various ways, even if such combinations or combinations are not explicitly described in the present invention. In particular, without departing from the spirit and teachings of the present invention, the features described in the various embodiments of the present invention can be combined and / or combined in various ways. All such combinations and / or combinations fall within the scope of the present invention.

[0224] The above describes the embodiments of the present invention. However, these embodiments are for illustrative purposes only and are not intended to limit the scope of the present invention. Although the embodiments are described separately above, this does not mean that the measures in each embodiment cannot be used advantageously in combination. Without departing from the scope of the present invention, those skilled in the art can make various substitutions and modifications, and all such substitutions and modifications should fall within the scope of the present invention.

Claims

1. An information processing method, characterized in that: The method comprises: In response to a problem to be processed input by a user via a target device, determine current device parameters of the target device and a target object and processing intent of the problem to be processed, wherein the target object includes a component of the target device, and the current device parameters include current parameters of the target object in the target device; Based on the processing intention, determining standard parameters for the target object from a database corresponding to the target object; Determining an initial solution to the problem to be solved based on the current parameters of the target object and the standard parameters; Structuring the initial solution to determine processing flow information; and The processing flow information is output according to the identity information of the user.

2. The method according to claim 1, characterized in that: The database includes standard parameters of each of a plurality of objects in the device, and the standard parameters are obtained by: Acquiring structured knowledge of multiple objects from a structured data source, wherein the structured data source includes object documents of the multiple objects; Acquire unstructured knowledge of the plurality of objects from an unstructured data source, wherein the unstructured data source includes maintenance logs of the plurality of objects and third-party documents for the plurality of objects; and The at least one structured knowledge and the at least one unstructured knowledge are processed to obtain the standard parameters.

3. The method according to claim 2, characterized in that The processing of the at least one structured knowledge and the at least one unstructured knowledge to obtain standard parameters includes: Performing semantic recognition on the structured knowledge and the unstructured knowledge respectively to obtain structured semantics and unstructured semantics; Based on the description method of the target structured knowledge, the description method of the target unstructured knowledge is modified to obtain the target knowledge, wherein the structured semantics of the target structured knowledge is the same as the unstructured semantics of the target unstructured knowledge, but the description methods are different; Performing natural language processing on the target knowledge to obtain configuration items of the target knowledge and corresponding standard values; and Based on the configuration items and corresponding standard values, the standard parameters are determined.

4. The method according to claim 1, characterized in that: The determining an initial solution to the problem to be processed based on the current parameters of the target object and the standard parameters of the target object includes: Inputting the current parameters of the target object and the standard parameters of the target object into the expert model to obtain the initial solution; The expert model is obtained in the following manner: Inputting sample data into the initial expert model to obtain a sample output result, wherein the sample data includes a sample problem from a sample device, parameters of the sample device, and a sample solution as a sample label; Determine a reward value of the sample output result based on a first solution of the sample problem by the sample device using the sample output result, a processing time of the sample problem, and the presence of sensitive operations in the sample output result; When it is determined that the reward value is lower than the reward threshold or the loss value obtained based on the sample output result and the sample label is greater than the first loss threshold, adjusting the model parameters of the initial expert model; and In the case where it is determined that the reward value is higher than the reward threshold and the loss value is lower than the first loss threshold, the expert model is obtained.

5. The method according to claim 4, characterized in that The reward value includes an operation type reward value and a processing time reward value; The operation type reward value is determined according to whether the sample output result includes the sensitive operation, and the processing time reward value is determined according to the average processing time of historical questions of the same question type as the sample question and the processing time of the sample output result.

6. The method according to claim 4, characterized in that The method further comprises: In the event that it is determined that the loss value is greater than a second loss threshold, the sample data is discarded, and the second loss threshold is greater than the first loss threshold.

7. The method according to any one of claims 4 to 6, characterized in that: The method further comprises: determining a reward value of the initial solution based on a user's evaluation of the processing flow information, a second solution of the problem to be processed by the target device using the processing flow information, a processing time of the problem to be processed, and the presence of sensitive operations in the initial solution; and When it is determined that the reward value is lower than the reward threshold, the model parameters of the expert model are adjusted.

8. The method according to claim 7, characterized in that The step of inputting the current parameters of the target object and the standard parameters of the target object into the expert model to obtain the initial solution comprises: Obtaining a judgment result according to whether the current parameters of the target object meet a judgment condition determined based on the target object and the standard parameters of the target object; and Based on the determination result, the initial solution is determined.

9. The method according to claim 1, characterized in that: The structural processing of the initial solution to determine the processing flow information includes: Decomposing the initial solution using a decomposition method for the initial solution to obtain at least one processing step, wherein the decomposition method is determined based on an object type of the target object; and Based on the at least one processing step, the processing flow information is determined.

10. The method according to claim 1, characterized in that The method further comprises: Determine, based on at least one historical question of the user, an average conversation depth of the at least one historical question, wherein the average conversation depth is determined based on the conversation depth of each of the at least one historical question, and the conversation depth is used to represent the number of follow-up questions of the user regarding the historical question; and Based on the word frequency of the target keyword in the at least one historical question and the average conversation depth, a proficiency evaluation value of the user for the target object is determined as the identity information of the user.

11. The method according to claim 10, characterized in that The step of outputting the processing flow information according to the identity information of the user includes: In the case where it is determined that the proficiency evaluation value represented by the identity information is greater than or equal to the proficiency threshold, respectively generating editable commands corresponding to at least one processing step included in the processing flow information, and sending the processing flow information and the editable commands to the user, so that the user performs the operation indicated by the processing flow information based on the editable commands; and When it is determined that the identity information represents that the proficiency assessment value is less than the proficiency threshold, an automated processing script generated based on the processing flow information and the processing flow information are sent to the user so that the user performs the operation indicated by the processing flow information based on the automated processing script.

12. The method according to claim 1, characterized in that The step of determining, in response to the problem to be processed input by the user via the target device, the current device parameters of the target device and the target object and processing intention of the problem to be processed includes: Performing natural language processing on the problem to be processed to determine the target object and the processing intention of the problem to be processed; and Information is collected from the target object in the target device to determine current parameters of the target object.

13. An electronic device, comprising: one or more processors; a memory for storing one or more computer programs, It is characterized in that the one or more processors execute the one or more computer programs to implement the steps of the method according to any one of claims 1 to 12.

14. A computer-readable storage medium having a computer program or instruction stored thereon, characterized in that: When the computer program or instruction is executed by a processor, the steps of the method according to any one of claims 1 to 12 are implemented.

15. A computer program product comprising a computer program or instructions, characterized in that When the computer program or instruction is executed by a processor, the steps of the method according to any one of claims 1 to 12 are implemented.

Images