Service request distribution method and device, electronic equipment and storage medium

Abstract

The invention discloses a service request distribution method and device, electronic equipment and a storage medium, and relates to the technical field of data processing, and the main technical scheme comprises the following steps: obtaining a first precise harmony degree of a historical service request; if the first accurate harmonic degree is smaller than a preset second threshold value, adjusting model parameters of a preset distribution model until the first accurate harmonic degree is larger than or equal to the preset second threshold value, and obtaining an adjusted preset distribution model; acquiring a current service request; and respectively inputting the service request text and the service request image into the adjusted preset distribution model to obtain a distribution result corresponding to the current service request. Compared with the prior art, the embodiment of the invention has the advantages that the preset distribution model is adjusted by utilizing the first precise harmony degree of the historical service request, the adaptability to the novel service request is improved, the adjusted preset distribution model is obtained, the distribution result of the current service request is predicted by utilizing the adjusted preset distribution model, and the service request distribution accuracy is improved. And the distribution accuracy of the service request is improved.

Description

Technical Field

[0001] The present disclosure relates to the field of data processing technology, and in particular to a method and device for dispatching a service request, an electronic device, and a storage medium. Background Art

[0002] Service request dispatching refers to assigning service requests submitted by users to appropriate processing personnel or teams so that they can process the service requests. This ensures that service requests are assigned to the most appropriate processing personnel or teams, optimizing the processing efficiency and quality of service requests.

[0003] In the related technologies of dispatching service requests, service requests are usually dispatched according to predefined rules (such as keyword matching, request source, etc.).

[0004] In real life, users may submit new service requests that are not within the predefined rules. Since the new service requests are not within the scope of the predefined rules, the predefined rules are less adaptable to the new service requests, resulting in a low accuracy rate in dispatching service requests. Summary of the invention

[0005] The present disclosure provides a method and device for dispatching a service request, an electronic device and a storage medium, the main purpose of which is to solve the problem of low accuracy in dispatching service requests.

[0006] According to a first aspect of the present disclosure, a method for dispatching a service request is provided, comprising:

[0007] When it is determined that the total number of historical service requests processed by the preset dispatch model within a preset period reaches a preset first threshold, obtaining a first precise reconciliation degree of the historical service requests;

[0008] If the first precise harmony degree is less than a preset second threshold, adjusting the model parameters of the preset distribution model until the first precise harmony degree is greater than or equal to the preset second threshold, thereby obtaining an adjusted preset distribution model;

[0009] When it is determined that the first precise harmony degree is greater than or equal to the preset second threshold, obtaining a current service request; the current service request includes a service request text and a service request image;

[0010] The service request text and the service request image are respectively input into the adjusted preset dispatch model to obtain a dispatch result corresponding to the current service request.

[0011] Optionally, the training method of the preset distribution model includes:

[0012] Acquire a training service request; the training service request includes a training service request text and a training service request image;

[0013] The initial dispatch model is trained using the training service request text and the training service request image to obtain the preset dispatch model.

[0014] Optionally, the using the training service request text and the training service request image to train the initial dispatch model to obtain the preset dispatch model includes:

[0015] Obtaining initial text weights and initial image weights in the initial distribution model, and obtaining a reference distribution result of the training service request;

[0016] Inputting the training service request text and the training service request image into the initial dispatch model respectively to obtain a predicted dispatch result of the training service request;

[0017] Based on the reference delivery result, calculating a second precise harmony of the predicted delivery result;

[0018] The initial text weight and the initial image weight are adjusted respectively until the second precise harmony is greater than or equal to the preset second threshold value, thereby obtaining the preset distribution model; the model parameters include the adjusted initial text weight and the adjusted initial image weight.

[0019] Optionally, the preset distribution model includes:

[0020] Input layer, encoder, fusion layer, fully connected layer, output layer; among them,

[0021] The input layer is used to input the current service request text and the current service request image into the encoder respectively;

[0022] The encoder is used to convert the current service request text into a first vector, and convert the current service request image into a second vector;

[0023] The fusion layer is used to perform a product calculation on the first vector and the adjusted initial text weight to obtain a first product value result, and to perform a product calculation on the second vector and the adjusted initial image weight to obtain a second product value result;

[0024] The fusion layer is further used to add the first product value result and the second product value result to obtain a fusion vector;

[0025] The fully connected layer is used to perform category prediction on the fusion vector to obtain a target category result;

[0026] The output layer is used to find the distribution result corresponding to the target category result according to the pre-established mapping relationship between the category result and the distribution result.

[0027] Optionally, after respectively inputting the service request text and the service request image into the adjusted preset dispatching model to obtain a dispatching result corresponding to the current service request, the method further includes:

[0028] Search the interface name in the dispatch result to obtain the processing interface name;

[0029] The processing interface corresponding to the processing interface name is called to process the current service request to obtain a processing result, so as to adjust the preset dispatch model based on the processing result.

[0030] Optionally, the obtaining the first precise reconciliation degree of the historical service request includes:

[0031] Acquire a first historical service request that needs to be dispatched from the historical service requests, and acquire a second historical service request that does not need to be dispatched from the historical service requests;

[0032] Searching the first historical service requests for a first number of historical service requests whose dispatch results predicted by the preset dispatch model are correct;

[0033] Searching the second historical service requests for a second number of historical service requests for which the preset dispatch model predicts that the second number of historical service requests needs to be dispatched;

[0034] Searching the first historical service requests for a third number of historical service requests for which the preset dispatch model predicts that the historical service requests do not need to be dispatched;

[0035] The first quantity, the second quantity, and the third quantity are calculated to obtain the first precise degree of harmony.

[0036] According to a second aspect of the present disclosure, a device for dispatching a service request is provided, comprising:

[0037] A first acquisition unit, configured to acquire a first precise reconciliation degree of the historical service requests when it is determined that the total number of historical service requests processed by a preset dispatch model within a preset period reaches a preset first threshold;

[0038] an adjusting unit, configured to adjust the model parameters of the preset distribution model when the first precise harmony degree is less than a preset second threshold value, until the first precise harmony degree is greater than or equal to the preset second threshold value, thereby obtaining an adjusted preset distribution model;

[0039] A second acquisition unit is used to acquire a current service request when it is determined that the first precise harmony degree is greater than or equal to the preset second threshold; the current service request includes a service request text and a service request image;

[0040] The input unit is used to input the service request text and the service request image into the adjusted preset dispatch model respectively to obtain the dispatch result corresponding to the current service request.

[0041] Optionally, the device further comprises:

[0042] A third acquisition unit is used to acquire a training service request; the training service request includes a training service request text and a training service request image;

[0043] The training unit is used to train the initial dispatch model using the training service request text and the training service request image to obtain the preset dispatch model.

[0044] Optionally, the training unit includes:

[0045] A first acquisition module, used to acquire an initial text weight and an initial image weight in the initial distribution model, and to acquire a reference distribution result of the training service request;

[0046] An input module, used to input the training service request text and the training service request image into the initial dispatch model respectively, to obtain a predicted dispatch result of the training service request;

[0047] A first calculation module, configured to calculate a second precision harmony of the predicted distribution result based on the reference distribution result;

[0048] An adjustment module is used to adjust the initial text weight and the initial image weight respectively until the second precise harmony is greater than or equal to the preset second threshold value, thereby obtaining the preset distribution model; the model parameters include the adjusted initial text weight and the adjusted initial image weight.

[0049] Optionally, the preset distribution model includes:

[0050] Input layer, encoder, fusion layer, fully connected layer, output layer; among them,

[0051] The input layer is used to input the current service request text and the current service request image into the encoder respectively;

[0052] The encoder is used to convert the current service request text into a first vector, and convert the current service request image into a second vector;

[0053] The fusion layer is used to perform a product calculation on the first vector and the adjusted initial text weight to obtain a first product value result, and to perform a product calculation on the second vector and the adjusted initial image weight to obtain a second product value result;

[0054] The fusion layer is further used to add the first product value result and the second product value result to obtain a fusion vector;

[0055] The fully connected layer is used to perform category prediction on the fusion vector to obtain a target category result;

[0056] The output layer is used to find the distribution result corresponding to the target category result according to the pre-established mapping relationship between the category result and the distribution result.

[0057] Optionally, the device further comprises:

[0058] A search unit, configured to input the service request text and the service request image into the adjusted preset dispatch model respectively, obtain a dispatch result corresponding to the current service request, and then search for an interface name in the dispatch result to obtain a processing interface name;

[0059] The processing unit is used to call the processing interface corresponding to the processing interface name, process the current service request, and obtain a processing result, so as to adjust the preset dispatch model based on the processing result.

[0060] Optionally, the first acquiring unit includes:

[0061] A second acquisition module, used for acquiring a first historical service request that needs to be dispatched among the historical service requests, and acquiring a second historical service request that does not need to be dispatched among the historical service requests;

[0062] A first search module, configured to search the first historical service requests for a first number of historical service requests for which the dispatch results predicted by the preset dispatch model are correct;

[0063] A second search module, configured to search the second historical service requests for a second number of historical service requests that need to be dispatched as predicted by the preset dispatch model;

[0064] A third search module, configured to search the first historical service requests for a third number of historical service requests for which the preset dispatch model predicts that the historical service requests do not need to be dispatched;

[0065] The second calculation module is used to calculate the first quantity, the second quantity and the third quantity to obtain the first precise harmony.

[0066] According to a third aspect of the present disclosure, there is provided an electronic device, including:

[0067] at least one processor; and

[0068] a memory communicatively connected to the at least one processor; wherein,

[0069] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method described in the first aspect.

[0070] According to a fourth aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to enable the computer to execute the method described in the first aspect.

[0071] According to a fifth aspect of the present disclosure, a computer program product is provided, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements the method as described in the first aspect above.

[0072] The method and device for dispatching service requests provided by the present disclosure, the electronic device and the storage medium, when it is determined that the total number of historical service requests processed by the preset dispatch model within a preset period reaches a preset first threshold, obtain the first precision harmony of the historical service requests; if the first precision harmony is less than the preset second threshold, adjust the model parameters of the preset dispatch model until the first precision harmony is greater than or equal to the preset second threshold, and obtain the adjusted preset dispatch model; when it is determined that the first precision harmony is greater than or equal to the preset second threshold, obtain the current service request; the current service request includes a service request text and a service request image; the service request text and the service request image are respectively input into the adjusted preset dispatch model to obtain the dispatch result corresponding to the current service request. Compared with the related art, the embodiment of the present disclosure adjusts the preset dispatch model by using the first precision harmony of the historical service requests, improves the adaptability to new service requests, obtains the adjusted preset dispatch model, and uses the adjusted preset dispatch model to predict the dispatch result of the current service request, thereby improving the accuracy of the dispatch of the service request.

[0073] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure.

[0075] Figure 1 A flowchart of a method for dispatching a service request provided by an embodiment of the present disclosure;

[0076] Figure 2 A flowchart of a method for training a preset distribution model provided in an embodiment of the present disclosure;

[0077] Figure 3 A schematic diagram of the structure of a preset dispatch model provided in an embodiment of the present disclosure;

[0078] Figure 4 A schematic diagram of the structure of a device for dispatching a service request provided by an embodiment of the present disclosure;

[0079] Figure 5 A schematic diagram of the structure of another device for dispatching service requests provided by an embodiment of the present disclosure;

[0080] Figure 6 A schematic block diagram of an exemplary electronic device provided for an embodiment of the present disclosure. DETAILED DESCRIPTION

[0081] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, the description of well-known functions and structures is omitted in the following description.

[0082] The following describes a method and apparatus for dispatching a service request, an electronic device, and a storage medium according to an embodiment of the present disclosure with reference to the accompanying drawings.

[0083] Figure 1 A flowchart of a method for dispatching a service request provided in an embodiment of the present disclosure.

[0084] like Figure 1 As shown, the method is applied in a server, and the method comprises the following steps:

[0085] Step 101: when it is determined that the total number of historical service requests processed by a preset dispatch model within a preset period reaches a preset first threshold, a first precise coordination degree of the historical service requests is obtained.

[0086] The preset dispatch model is a pre-set or trained model. The preset dispatch model is used to predict or decide which processing personnel or team to dispatch the service request to for processing based on the input service request (including service request text and service request image). The preset dispatch model is a model built on a multimodal fusion network. A multimodal fusion network refers to a network that can integrate information from different modalities (such as text, images, videos, voice, etc.) for more efficient information processing, understanding, and recognition. By fusing information from multiple modalities, the multimodal network can make full use of the complementary information between the modalities and improve the generalization ability and accuracy of the network.

[0087] Historical service requests are service requests whose dispatch results have been predicted by the preset dispatch model. Historical service requests are stored in the preset storage database. By calling the preset storage database, historical service requests can be obtained. Among them, a service request refers to a request initiated by a user or a user's authorized representative to the server to obtain a service action that has been agreed to be a normal part of service delivery. Service requests include service request text and service request images. Service requests usually cover a series of user needs. Service requests include but are not limited to software updates, hardware replacements, problem consultations, account settings, password resets, device replacements, information requests, and access requests.

[0088] The first precision reconciliation is an indicator for measuring the performance of the preset dispatch model. The first precision reconciliation is used to evaluate the accuracy and efficiency of the preset dispatch model in processing historical service requests. The higher the first precision reconciliation is, the higher the accuracy of the preset dispatch model in processing historical service requests is.

[0089] The calculation formula of the first precise harmonic degree can be realized by formula (1):

[0090]

[0091] Among them, F1 is the first precise reconciliation degree, Precision is the precision of the preset dispatch model in processing historical service requests, and Recall is the recall rate of the preset dispatch model in processing instance service requests.

[0092] The calculation formula of the accuracy can be realized by formula (2):

[0093]

[0094] Among them, Precision is the accuracy rate, TP is the first number of historical service requests whose dispatch results predicted by the preset dispatch model are correct among the first historical service requests that need to be dispatched in the historical service requests, and FP is the second number of historical service requests whose dispatch results predicted by the preset dispatch model are correct among the second historical service requests that do not need to be dispatched in the historical service requests.

[0095] The calculation formula of recall rate can be realized by formula (3):

[0096]

[0097] Among them, Recall is the recall rate, TP is the first number of historical service requests whose dispatch results predicted by the preset dispatch model are correct in the first historical service requests that need to be dispatched in the historical service requests, and FN is the third number of historical service requests whose dispatch results predicted by the preset dispatch model are not required to be dispatched in the first historical service requests that need to be dispatched.

[0098] Exemplarily, the preset period and the preset first threshold are arbitrary values ​​and are not limited in the embodiments of the present disclosure.

[0099] For ease of understanding, an example is provided. Assuming that the preset period is 3 hours and the preset first threshold is 30, if the total number of historical service requests processed by the preset dispatch model within 3 hours is 40, the first precise coordination degree of the historical service requests is obtained.

[0100] The purpose of obtaining the first precise reconciliation of historical service requests is to detect the accuracy of the preset dispatch model in processing service requests, provide a basis for subsequent adjustment and optimization of the preset dispatch model, and ultimately improve the accuracy of service request dispatch.

[0101] Step 102: If the first precise harmony degree is less than a preset second threshold, adjust the model parameters of the preset distribution model until the first precise harmony degree is greater than or equal to the preset second threshold, thereby obtaining an adjusted preset distribution model.

[0102] Model parameters include but are not limited to text weight and image weight. Text weight refers to the importance or influence that the preset dispatch model gives to the text part of the service request when processing the service request. A higher text weight means that the preset dispatch model will rely more on the service request text to make dispatch decisions, while a relatively lower text weight means that the preset dispatch model will consider the service request text less when making dispatch decisions. Image weight refers to the importance or influence that the preset dispatch model gives to the image part of the service request when processing the service request. A higher image weight means that the preset dispatch model will rely more on the service request image to make dispatch decisions, while a relatively lower image weight means that the preset dispatch model will consider the service request image less when making decisions.

[0103] Different proportions of text weights and image weights will affect the accuracy of the preset dispatch model in processing service requests, and thus affect the precise coordination of service requests.

[0104] Exemplarily, the preset second threshold may be any value, and the embodiment of the present disclosure does not limit the preset second threshold.

[0105] For ease of understanding, an example is provided. Assuming that the preset second threshold is 0.6, when the text weight is 0.5 and the image weight is 0.5, the first precise harmony degree is 0.4. When the text weight is adjusted to 0.7 and the image weight is adjusted to 0.3, the first precise harmony degree is 0.8. Then the text weight is adjusted from 0.5 to 0.7, and the image weight is adjusted from 0.5 to 0.3.

[0106] By adjusting the model parameters, the performance of the preset dispatch model in processing service requests can be improved, so that its predicted results are closer to the actual results. When the first precision reconciliation is less than the preset second threshold, it means that the prediction accuracy of the preset dispatch model needs to be improved, so the model parameters need to be adjusted.

[0107] Step 103: when it is determined that the first precise harmony degree is greater than or equal to the preset second threshold, a current service request is obtained; the current service request includes a service request text and a service request image.

[0108] After the model parameters are adjusted, it is ensured that the first precision reconciliation reaches or exceeds the preset second threshold value, and the performance of the adjusted preset dispatch model has been optimized. At this time, the current service request is obtained in order to use the adjusted preset dispatch model with better performance for prediction. The current service request is a newly generated service request that has not been predicted by the preset dispatch model or the adjusted preset dispatch model for dispatch results.

[0109] For example, the current service request can be obtained through an application programming interface (API), and the current service request can be stored in a preset storage database for subsequent processing. However, it should be clear that this statement is not intended to limit the method of obtaining the current service request to only be achieved through the API, and the current service request can also be obtained through other methods. The embodiment of the present disclosure does not limit the method of obtaining the current service request.

[0110] An example is provided for ease of understanding. Assuming that the preset second threshold is 0.6 and the first accuracy is 0.7, the current service request is obtained.

[0111] Step 104 , inputting the service request text and the service request image into the adjusted preset dispatching model respectively, to obtain a dispatching result corresponding to the current service request.

[0112] The dispatch result indicates the target processing person or team to which the current service request should be assigned. The dispatch result is used to dispatch the current service request to the target processing person or team.

[0113] The adjusted preset dispatch model has been optimized to process the current service request more accurately. Therefore, by inputting the current service request into the adjusted preset dispatch model, a more accurate dispatch result can be expected.

[0114] For ease of understanding, an example is provided. Assume that there are three persons handling the current service request, namely a, b, and c. The dispatch result indicates that the current service request is assigned to c, then the dispatch result is assigned to c for processing.

[0115] The method for dispatching service requests provided by the present disclosure, when it is determined that the sum of the number of historical service requests processed by the preset dispatch model within a preset period reaches a preset first threshold, obtains the first precision reconciliation of the historical service requests; if the first precision reconciliation is less than the preset second threshold, adjusts the model parameters of the preset dispatch model until the first precision reconciliation is greater than or equal to the preset second threshold, and obtains the adjusted preset dispatch model; when it is determined that the first precision reconciliation is greater than or equal to the preset second threshold, obtains the current service request; the current service request includes a service request text and a service request image; the service request text and the service request image are respectively input into the adjusted preset dispatch model to obtain the dispatch result corresponding to the current service request. Compared with the related art, the embodiment of the present disclosure improves the adaptability to new service requests by adjusting the preset dispatch model using the first precision reconciliation of the historical service requests, obtains the adjusted preset dispatch model, and uses the adjusted preset dispatch model to predict the dispatch result of the current service request, thereby improving the accuracy of the dispatch of the service request.

[0116] In practical applications, in order to improve the accuracy of the preset dispatch model in predicting the dispatch result of the service request, the preset dispatch model needs to be trained. The training method of the preset dispatch model can be implemented in the following ways, but is not limited to: Figure 2 As shown, Figure 2 A flowchart of a training method for a preset distribution model provided in an embodiment of the present disclosure includes:

[0117] Step 201, obtaining a training service request; the training service request includes a training service request text and a training service request image.

[0118] Continuously obtain service requests generated by users through the API, and store the service requests in a preset storage database. Obtain service requests by calling the preset storage database, and preprocess the service requests to obtain training service requests, wherein the preprocessing includes cleaning and formatting data, and extracting key features from unstructured data, such as extracting keywords from text or images.

[0119] The diversity and richness of training service requests are crucial to improving the generalization ability of the preset dispatch model. Training service requests involve various scenarios, problems, and situations. By collecting training service requests containing different training service request texts and training service request images, the preset dispatch model can still make reasonable dispatch decisions when facing new and unseen service requests.

[0120] Step 202: train the initial dispatch model using the training service request text and the training service request image to obtain the preset dispatch model.

[0121] The initial distribution model refers to a model that has been built before training but has not yet been optimized with actual data. The initial distribution model is built based on a multimodal attention fusion network.

[0122] The initial dispatch model is the starting point of the training process. By training with training service request text and training service request images, the model parameters and structure of the initial dispatch model can be gradually adjusted and optimized, so that it can better understand and process service requests and make accurate dispatch decisions.

[0123] Since the training service request contains both training service request text and training service request image information, the training process will teach the initial dispatch model how to process data in these two modalities at the same time. This multimodal processing capability enables the initial dispatch model to understand and analyze the training service request more comprehensively, thereby making more accurate dispatch decisions.

[0124] As a refinement of step 202, when executing the training of the initial dispatch model using the training service request text and the training service request image to obtain the preset dispatch model, it can be implemented in but not limited to the following ways: obtaining the initial text weight and the initial image weight in the initial dispatch model, and obtaining the reference dispatch result of the training service request; inputting the training service request text and the training service request image into the initial dispatch model respectively to obtain the predicted dispatch result of the training service request; calculating the second precise harmony of the predicted dispatch result based on the reference dispatch result; adjusting the initial text weight and the initial image weight respectively until the second precise harmony is greater than or equal to the preset second threshold value to obtain the preset dispatch model; the model parameters include the adjusted initial text weight and the adjusted initial image weight.

[0125] Exemplarily, the initial text weight refers to the weight assigned to the text data portion of the service request in the initial dispatch model. The size of the weight reflects the degree of importance or reliance of the initial dispatch model on the service request text when processing the service request. The initial text weight is a set value before the initial dispatch model is trained, which may be determined based on experience, domain knowledge or preliminary experiments. The initial image weight refers to the weight assigned to the image data portion of the service request in the initial dispatch model. It reflects the degree of importance or reliance of the initial dispatch model on the image data. The initial image weight is also a set value before the initial dispatch model is trained, and its determination method is similar to that of the initial text weight.

[0126] The reference dispatch result refers to the known or preset correct or expected dispatch result for each training service request. During the training of the initial dispatch model, the reference dispatch result is used as a benchmark to evaluate the accuracy of the model's predicted dispatch results. Usually, the reference dispatch result is manually annotated or determined based on historical data.

[0127] The calculation method of the second precise harmony is the same as that of the first precise harmony, and the second precise harmony can be calculated by formula (1).

[0128] In the process of training the initial dispatch model, by continuously adjusting the initial text weights and the initial image weights and observing how these adjustments affect the second precision reconciliation, a set of weight values ​​can be found so that the initial dispatch model can achieve a higher accuracy (i.e., the second precision reconciliation is greater than or equal to the preset second threshold) when predicting the dispatch result of the service request. This set of weight values ​​and corresponding model parameters constitute the final preset dispatch model.

[0129] By building and optimizing the preset dispatch model, we can realize the automatic processing of service requests, reduce manual intervention, and improve processing efficiency. At the same time, since the preset dispatch model has high accuracy, it can ensure that the dispatch results of service requests are more reasonable and reliable, thereby improving the overall service quality and customer satisfaction.

[0130] As a refinement of the above embodiment, in order to facilitate a better understanding of the structure of the preset dispatch model, as shown in FIG. Figure 3 As shown, Figure 3 A structural schematic diagram of a preset dispatch model provided in an embodiment of the present disclosure, the preset dispatch model includes: an input layer, an encoder, a fusion layer, a fully connected layer, and an output layer; wherein the input layer is used to input the current service request text and the current service request image into the encoder respectively; the encoder is used to convert the current service request text into a first vector, and convert the current service request image into a second vector; the fusion layer is used to perform a product calculation on the first vector and the adjusted initial text weight to obtain a first product value result, and to perform a product calculation on the second vector and the adjusted initial image weight to obtain a second product value result; the fusion layer is also used to perform a sum calculation on the first product value result and the second product value result to obtain a fused vector; the fully connected layer is used to perform a category prediction on the fused vector to obtain a target category result; the output layer is used to search for the dispatch result corresponding to the target category result according to a pre-established mapping relationship between the category result and the dispatch result.

[0131] In actual applications, after the service request text and the service request image are respectively input into the adjusted preset dispatch model to obtain the dispatch result corresponding to the current service request, the current service request needs to be dispatched to the corresponding processing personnel or team for processing according to the dispatch result. This can be achieved by but is not limited to the following methods: performing an interface name search in the dispatch result to obtain a processing interface name; calling a processing interface corresponding to the processing interface name to process the current service request and obtain a processing result, so as to adjust the preset dispatch model based on the processing result.

[0132] Exemplarily, the processing interface name refers to the unique identifier or name of the processing interface. Each processing interface has a unique corresponding processing interface name. The processing interface is an interface specifically used to process one or more types of service requests. When a system or application needs to perform a processing task, it will do so by calling the processing interface. The processing result refers to the data or status returned by the processing interface after the processing interface executes its internal logic after calling the processing interface. This result is the output of the processing interface after processing the input data. The processing result may contain a variety of information, such as a sign of success or failure, processed data, error information, etc. Based on this processing result, the system or application can make further decisions, such as updating the database, displaying information to the user, or adjusting subsequent processing procedures based on the results.

[0133] The processing results are an important basis for evaluating the performance of the preset distribution model. By comparing the processing results with the expected results, the system can find deficiencies and errors in the model during the distribution process. Based on these feedbacks, the system can adjust and optimize the preset distribution model, such as adjusting the distribution rules, weight parameters, etc., to improve the accuracy and efficiency of the preset distribution model. This adjustment method based on actual processing results helps to achieve continuous improvement and adaptive capabilities of the preset distribution model.

[0134] As a refinement of step 101, when executing the method of obtaining the first precise degree of harmony of the historical service requests, it can be implemented in but not limited to the following manner: obtaining the first historical service request that needs to be dispatched among the historical service requests, and obtaining the second historical service request that does not need to be dispatched among the historical service requests; searching the first historical service requests for a first number of historical service requests whose dispatch results predicted by the preset dispatch model are correct; searching the second historical service requests for a second number of historical service requests whose dispatch results predicted by the preset dispatch model are required to be dispatched; searching the first historical service requests for a third number of historical service requests whose dispatch results predicted by the preset dispatch model are not required to be dispatched; and calculating the first number, the second number and the third number to obtain the first precise degree of harmony.

[0135] The implementation process of this embodiment is a textual description of formula (1).

[0136] In summary, the embodiments of the present disclosure can achieve the following effects:

[0137] The disclosed embodiment adjusts the preset dispatch model by utilizing the first precise degree of reconciliation of historical service requests, thereby improving adaptability to new service requests, obtaining an adjusted preset dispatch model, and utilizing the adjusted preset dispatch model to predict the dispatch result of the current service request, thereby improving the accuracy of the dispatch of service requests.

[0138] Corresponding to the above service request dispatching method, the present invention also provides a service request dispatching device. Since the device embodiment of the present invention corresponds to the above method embodiment, the details not disclosed in the device embodiment can be referred to the above method embodiment, and will not be repeated in the present invention.

[0139] Figure 4 A schematic diagram of a structure of a device for dispatching a service request provided by an embodiment of the present disclosure, such as Figure 4 As shown, the device is applied to a server and includes:

[0140] A first acquisition unit 31 is used to acquire a first precise reconciliation degree of the historical service requests when it is determined that the total number of historical service requests processed by a preset dispatch model within a preset period reaches a preset first threshold;

[0141] An adjusting unit 32, configured to adjust the model parameters of the preset distribution model when the first precise harmony degree is less than a preset second threshold value, until the first precise harmony degree is greater than or equal to the preset second threshold value, thereby obtaining an adjusted preset distribution model;

[0142] The second acquisition unit 33 is used to acquire a current service request when it is determined that the first precise harmony degree is greater than or equal to the preset second threshold; the current service request includes a service request text and a service request image;

[0143] The input unit 34 is used to input the service request text and the service request image into the adjusted preset dispatch model respectively to obtain the dispatch result corresponding to the current service request.

[0144] The service request dispatching device provided by the present disclosure obtains the first precision reconciliation degree of the historical service requests when it is determined that the total number of historical service requests processed by the preset dispatching model within a preset period reaches a preset first threshold; if the first precision reconciliation degree is less than the preset second threshold, adjusts the model parameters of the preset dispatching model until the first precision reconciliation degree is greater than or equal to the preset second threshold, and obtains the adjusted preset dispatching model; when it is determined that the first precision reconciliation degree is greater than or equal to the preset second threshold, obtains the current service request; the current service request includes a service request text and a service request image; the service request text and the service request image are respectively input into the adjusted preset dispatching model to obtain the dispatching result corresponding to the current service request. Compared with the related art, the embodiment of the present disclosure adjusts the preset dispatching model by using the first precision reconciliation degree of the historical service requests, thereby improving the adaptability to new service requests, obtaining the adjusted preset dispatching model, and using the adjusted preset dispatching model to predict the dispatching result of the current service request, thereby improving the accuracy of the dispatching of the service request.

[0145] Furthermore, in a possible implementation of this embodiment, as Figure 5 As shown, the device also includes:

[0146] The third acquisition unit 35 is used to acquire a training service request; the training service request includes a training service request text and a training service request image;

[0147] The training unit 36 ​​is used to train the initial dispatch model using the training service request text and the training service request image to obtain the preset dispatch model.

[0148] Furthermore, in a possible implementation of this embodiment, as Figure 5 As shown, the training unit 36 ​​includes:

[0149] A first acquisition module 361 is used to acquire an initial text weight and an initial image weight in the initial distribution model, and to acquire a reference distribution result of the training service request;

[0150] An input module 362, configured to input the training service request text and the training service request image into the initial dispatch model respectively, to obtain a predicted dispatch result of the training service request;

[0151] A first calculation module 363, configured to calculate a second accurate harmony degree of the predicted distribution result based on the reference distribution result;

[0152] The adjustment module 364 is used to adjust the initial text weight and the initial image weight respectively until the second precise harmony is greater than or equal to the preset second threshold value, thereby obtaining the preset distribution model; the model parameters include the adjusted initial text weight and the adjusted initial image weight.

[0153] Furthermore, in a possible implementation of this embodiment, the preset distribution model includes:

[0154] Input layer, encoder, fusion layer, fully connected layer, output layer; among them,

[0155] The input layer is used to input the current service request text and the current service request image into the encoder respectively;

[0156] The encoder is used to convert the current service request text into a first vector, and convert the current service request image into a second vector;

[0157] The fusion layer is used to perform a product calculation on the first vector and the adjusted initial text weight to obtain a first product value result, and to perform a product calculation on the second vector and the adjusted initial image weight to obtain a second product value result;

[0158] The fusion layer is further used to add the first product value result and the second product value result to obtain a fusion vector;

[0159] The fully connected layer is used to perform category prediction on the fusion vector to obtain a target category result;

[0160] The output layer is used to find the distribution result corresponding to the target category result according to the pre-established mapping relationship between the category result and the distribution result.

[0161] Furthermore, in a possible implementation of this embodiment, as Figure 5 As shown, the device also includes:

[0162] A search unit 37 is used for inputting the service request text and the service request image into the adjusted preset dispatch model respectively to obtain the dispatch result corresponding to the current service request, and then searching for the interface name in the dispatch result to obtain the processing interface name;

[0163] The processing unit 38 is used to call the processing interface corresponding to the processing interface name to process the current service request and obtain a processing result so as to adjust the preset dispatch model based on the processing result.

[0164] Furthermore, in a possible implementation of this embodiment, as Figure 5 As shown, the first acquisition unit 31 includes:

[0165] A second acquisition module 311 is used to acquire a first historical service request that needs to be dispatched among the historical service requests, and acquire a second historical service request that does not need to be dispatched among the historical service requests;

[0166] A first search module 312, configured to search the first historical service requests for a first number of historical service requests for which the dispatch results predicted by the preset dispatch model are correct;

[0167] A second search module 313 is used to search the second historical service requests for a second number of historical service requests that need to be dispatched as predicted by the preset dispatch model;

[0168] A third search module 314 is used to search the first historical service requests for a third number of historical service requests for which the preset dispatch model predicts that the historical service requests do not need to be dispatched;

[0169] The second calculation module 315 is used to calculate the first number, the second number and the third number to obtain the first precise harmony.

[0170] It should be noted that the above explanation of the method embodiment is also applicable to the device of the embodiment of the present disclosure, and the principle is the same, which is no longer limited in the embodiment of the present disclosure.

[0171] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

[0172] Figure 6 A schematic block diagram of an example electronic device 400 that can be used to implement an embodiment of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.

[0173] like Figure 6 As shown, the device 400 includes a computing unit 401, which can perform various appropriate actions and processes according to a computer program stored in a ROM (Read-Only Memory) 402 or a computer program loaded from a storage unit 408 to a RAM (Random Access Memory) 403. In the RAM 403, various programs and data required for the operation of the device 400 can also be stored. The computing unit 401, the ROM 402, and the RAM 403 are connected to each other via a bus 404. An I / O (Input / Output) interface 405 is also connected to the bus 404.

[0174] A number of components in the device 400 are connected to the I / O interface 405, including: an input unit 406, such as a keyboard, a mouse, etc.; an output unit 407, such as various types of displays, speakers, etc.; a storage unit 408, such as a disk, an optical disk, etc.; and a communication unit 409, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 409 allows the device 400 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0175] The computing unit 401 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 401 include, but are not limited to, a CPU (Central Processing Unit), a GPU (Graphic Processing Units), various dedicated AI (Artificial Intelligence) computing chips, various computing units running machine learning model algorithms, a DSP (Digital Signal Processor), and any appropriate processor, controller, microcontroller, etc. The computing unit 401 performs the various methods and processes described above, such as a method for dispatching a service request. For example, in some embodiments, the method for dispatching a service request may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 408. In some embodiments, part or all of the computer program may be loaded and / or installed on the device 400 via the ROM 402 and / or the communication unit 409. When the computer program is loaded into the RAM 403 and executed by the computing unit 401, one or more steps of the method described above may be performed. Alternatively, in other embodiments, the computing unit 401 may be configured to execute the aforementioned service request dispatching method in any other appropriate manner (for example, by means of firmware).

[0176] Various embodiments of the systems and techniques described above herein may be implemented in digital electronic circuit systems, integrated circuit systems, FPGAs (Field Programmable Gate Arrays), ASICs (Application-Specific Integrated Circuits), ASSPs (Application Specific Standard Products), SOCs (System On Chips), CPLDs (Complex Programmable Logic Devices), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include: being implemented in one or more computer programs that may be executed and / or interpreted on a programmable system including at least one programmable processor that may be a dedicated or general-purpose programmable processor that may receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0177] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0178] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or device, or any suitable combination of the foregoing. More specific examples of machine-readable storage media may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a RAM, a ROM, an EPROM (Electrically Programmable Read-Only-Memory) or a flash memory, an optical fiber, a CD-ROM (Compact Dis sc Read-Only Memory), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0179] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (Cathode-Ray Tube) or LCD (Liquid Crystal Display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0180] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: LAN (Local Area Network), WAN (Wide Area Network), the Internet, and blockchain networks.

[0181] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship between the client and the server is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services ("Virtual Private Server", or "VPS" for short). The server may also be a server of a distributed system, or a server combined with a blockchain.

[0182] It should be noted that artificial intelligence is a discipline that studies how computers can simulate certain human thought processes and intelligent behaviors (such as learning, reasoning, thinking, planning, etc.), and includes both hardware-level and software-level technologies. Artificial intelligence hardware technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, and big data processing; artificial intelligence software technologies mainly include computer vision technology, speech recognition technology, natural language processing technology, as well as machine learning / deep learning, big data processing technology, knowledge graph technology, and other major directions.

[0183] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.

[0184] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A method for dispatching a service request, characterized in that: include: When it is determined that the total number of historical service requests processed by the preset dispatch model within a preset period reaches a preset first threshold, obtaining a first precise reconciliation degree of the historical service requests; If the first precise harmony degree is less than a preset second threshold, adjusting the model parameters of the preset distribution model until the first precise harmony degree is greater than or equal to the preset second threshold, thereby obtaining an adjusted preset distribution model; When it is determined that the first precise harmony degree is greater than or equal to the preset second threshold, obtaining a current service request; the current service request includes a service request text and a service request image; The service request text and the service request image are respectively input into the adjusted preset dispatch model to obtain a dispatch result corresponding to the current service request.

2. The method according to claim 1, characterized in that The training method of the preset distribution model includes: Acquire a training service request; the training service request includes a training service request text and a training service request image; The initial dispatch model is trained using the training service request text and the training service request image to obtain the preset dispatch model.

3. The method according to claim 2, characterized in that The using the training service request text and the training service request image to train the initial dispatch model to obtain the preset dispatch model includes: Obtaining initial text weights and initial image weights in the initial distribution model, and obtaining a reference distribution result of the training service request; Inputting the training service request text and the training service request image into the initial dispatch model respectively to obtain a predicted dispatch result of the training service request; Based on the reference delivery result, calculating a second precise harmony of the predicted delivery result; The initial text weight and the initial image weight are adjusted respectively until the second precise harmony is greater than or equal to the preset second threshold value, thereby obtaining the preset distribution model; the model parameters include the adjusted initial text weight and the adjusted initial image weight.

4. The method according to claim 3, characterized in that: The preset distribution model includes: Input layer, encoder, fusion layer, fully connected layer, output layer; among them, The input layer is used to input the current service request text and the current service request image into the encoder respectively; The encoder is used to convert the current service request text into a first vector, and convert the current service request image into a second vector; The fusion layer is used to perform a product calculation on the first vector and the adjusted initial text weight to obtain a first product value result, and to perform a product calculation on the second vector and the adjusted initial image weight to obtain a second product value result; The fusion layer is further used to add the first product value result and the second product value result to obtain a fusion vector; The fully connected layer is used to perform category prediction on the fusion vector to obtain a target category result; The output layer is used to find the distribution result corresponding to the target category result according to the pre-established mapping relationship between the category result and the distribution result.

5. The method according to claim 1, characterized in that After respectively inputting the service request text and the service request image into the adjusted preset dispatching model to obtain a dispatching result corresponding to the current service request, the method further includes: Search the interface name in the dispatch result to obtain the processing interface name; The processing interface corresponding to the processing interface name is called to process the current service request to obtain a processing result, so as to adjust the preset dispatch model based on the processing result.

6. The method according to claim 1, characterized in that The obtaining of the first precise reconciliation degree of the historical service request comprises: Acquire a first historical service request that needs to be dispatched from the historical service requests, and acquire a second historical service request that does not need to be dispatched from the historical service requests; Searching the first historical service requests for a first number of historical service requests whose dispatch results predicted by the preset dispatch model are correct; Searching the second historical service requests for a second number of historical service requests for which the preset dispatch model predicts that the second number of historical service requests needs to be dispatched; Searching the first historical service requests for a third number of historical service requests for which the preset dispatch model predicts that the historical service requests do not need to be dispatched; The first quantity, the second quantity, and the third quantity are calculated to obtain the first precise degree of harmony.

7. A device for dispatching service requests, characterized in that: include: A first acquisition unit, configured to acquire a first precise reconciliation degree of the historical service requests when it is determined that the total number of historical service requests processed by a preset dispatch model within a preset period reaches a preset first threshold; an adjusting unit, configured to adjust the model parameters of the preset distribution model when the first precise harmony degree is less than a preset second threshold value, until the first precise harmony degree is greater than or equal to the preset second threshold value, thereby obtaining an adjusted preset distribution model; A second acquisition unit is used to acquire a current service request when it is determined that the first precise harmony degree is greater than or equal to the preset second threshold; the current service request includes a service request text and a service request image; The input unit is used to input the service request text and the service request image into the adjusted preset dispatch model respectively to obtain the dispatch result corresponding to the current service request.

8. An electronic device, characterized in that: include: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 6.

9. A non-transitory computer-readable storage medium storing computer instructions, characterized in that: The computer instructions are used to cause the computer to execute the method according to any one of claims 1-6.

10. A computer program product, characterized in that The invention comprises a computer program which, when executed by a processor, implements the method according to any one of claims 1 to 6.

Images