A method and device for processing consultation request

Abstract

The present invention discloses a method and device for processing consultation requests, and relates to the field of computer technology. A specific implementation of the method includes: constructing an atomic map based on process information, target information, and fill-in item information; wherein, in the atomic map, atomic nodes with associations are directly connected; receiving a user's consultation request, and matching a target atomic node on the atomic map according to the consultation statement in the consultation request; obtaining a response result corresponding to the consultation statement according to the atomic nodes connected to the target atomic node and / or the atomic knowledge associated with the target atomic node, thereby responding to the consultation request. This implementation can solve the technical problem that it is impossible to provide different degrees of answers to the granularity of the problem.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a method and device for processing a consultation request. Background Art

[0002] Customer service needs to answer user questions in different ways. User questions are all kinds of strange, so different levels of answers are needed according to the granularity of the questions. Currently, the following two methods are mainly used to answer user questions:

[0003] (1) Manual method: For customer inquiries, customer service staff need to check the content of the corresponding questions in the existing customer service knowledge system, and clarify the scope of the user's inquiry based on the questions raised by the user to provide users with different levels of answers.

[0004] (2) Semantic understanding-based approach: This approach mainly matches the results most similar to the user text from the text library, thereby selecting the question-answer pairs most relevant to the user's inquiry content to obtain the answer results.

[0005] Since neither of the above two methods can provide different levels of answers based on the granularity of the questions, it is difficult to accurately answer the user's questions, resulting in a poor user experience. Summary of the invention

[0006] In view of this, an embodiment of the present invention provides a method and apparatus for processing a consultation request to solve the technical problem that it is impossible to provide answers to different degrees according to the granularity of the question.

[0007] To achieve the above object, according to one aspect of an embodiment of the present invention, a method for processing a consultation request is provided, comprising:

[0008] Constructing an atomic map according to the process information, the target information and the filling item information; wherein in the atomic map, the atomic nodes with association are directly connected;

[0009] receiving a consultation request from a user, and matching a target atomic node in the atomic map according to a consultation statement in the consultation request;

[0010] According to the atomic nodes connected to the target atomic node and / or the atomic knowledge associated with the target atomic node, a response result corresponding to the consultation statement is obtained, thereby responding to the consultation request.

[0011] Optionally, an atomic map is constructed based on the process information, target information, and fill-in item information, including:

[0012] Build a process tree based on process information;

[0013] Marking the association between each process node in the process tree and the target information and the fill-in item information;

[0014] The labeled results are used as training samples to train the dual-tower model and construct an atomic map.

[0015] Optionally, after training the dual-tower model using the labeled results as training samples and constructing the atomic map, the following steps are also included:

[0016] For the newly added process nodes, the trained dual-tower model is used to calculate the correlation between the newly added process nodes and the target information and the filling item information, so as to update the atomic nodes in the atomic map;

[0017] For the newly added target information, the trained dual-tower model is used to calculate the correlation between the newly added target information and the process nodes, thereby updating the atomic nodes in the atomic map;

[0018] For the newly added filling item information, the trained dual-tower model is used to calculate the correlation between the newly added filling item information and the process nodes, so as to update the atomic nodes in the atomic map.

[0019] Optionally, after training the dual-tower model using the labeled results as training samples and constructing the atomic map, the following steps are also included:

[0020] Decompose the document information to obtain the atomic knowledge of each slot;

[0021] For each atomic node on the atomic map, the similarity between the atomic node and the atomic knowledge of each slot is calculated respectively, and the atomic knowledge with a similarity greater than a similarity threshold is associated with the atomic node.

[0022] Optionally, matching a target atomic node on the atomic map according to the consultation statement in the consultation request includes:

[0023] identifying consulting keywords from the consulting sentences in the consulting request;

[0024] The similarities between the consulting keywords and the atomic nodes on the atomic map are calculated respectively, so as to select the target atomic nodes with the highest similarity.

[0025] Optionally, obtaining a response result corresponding to the consultation statement according to the atomic node connected to the target atomic node and / or the atomic knowledge associated with the target atomic node, thereby responding to the consultation request, includes:

[0026] According to the atomic nodes connected to the target atomic node, the target information or fill-in item information corresponding to the atomic node is obtained, thereby responding to the consultation request;

[0027] and / or,

[0028] The atomic knowledge associated with the target atomic node is used as the response result corresponding to the consultation statement, thereby responding to the consultation request.

[0029] Optionally, obtaining target information or fill-in item information corresponding to the target atomic node according to the atomic node connected to the target atomic node, thereby responding to the consultation request, includes:

[0030] According to the atomic nodes connected to the target atomic node, the target information corresponding to the atomic node is obtained, and the target information is used as the response result corresponding to the consultation statement, thereby responding to the consultation request;

[0031] or,

[0032] According to the atomic nodes connected to the target atomic node, the filling item information corresponding to the atomic node is obtained, and a service order is generated according to the filling item information to respond to the consultation request.

[0033] In addition, according to another aspect of an embodiment of the present invention, there is provided a device for processing a consultation request, comprising:

[0034] A construction module, used to construct an atomic map according to process information, target information and fill-in item information; wherein, in the atomic map, atomic nodes with association are directly connected;

[0035] A matching module, configured to receive a user's consultation request and match a target atomic node on the atomic map according to a consultation statement in the consultation request;

[0036] The response module is used to obtain a response result corresponding to the consultation statement according to the atomic nodes connected to the target atomic node and / or the atomic knowledge associated with the target atomic node, so as to respond to the consultation request.

[0037] Optionally, the building block is further used to:

[0038] Build a process tree based on process information;

[0039] Marking the association between each process node in the process tree and the target information and the fill-in item information;

[0040] The labeled results are used as training samples to train the dual-tower model and construct an atomic map.

[0041] Optionally, the building block is further used to:

[0042] After training the twin-tower model and constructing the atomic map using the labeled results as training samples,

[0043] For a newly added process node, the trained dual-tower model is used to calculate the correlation between the newly added process node and the target information and the fill-in item information, thereby updating the atomic node on the atomic map;

[0044] For the newly added target information, the trained dual-tower model is used to calculate the correlation between the newly added target information and the process nodes, thereby updating the atomic nodes on the atomic map;

[0045] For the newly added filling item information, the trained dual-tower model is used to calculate the correlation between the newly added filling item information and the process nodes, so as to update the atomic nodes on the atomic map.

[0046] Optionally, the building block is further used to:

[0047] After training the dual-tower model and constructing the atomic map using the annotation results as training samples, the document information is disassembled to obtain the atomic knowledge of each slot;

[0048] For each atomic node on the atomic map, the similarity between the atomic node and the atomic knowledge of each slot is calculated respectively, and the atomic knowledge with a similarity greater than a similarity threshold is associated with the atomic node.

[0049] Optionally, the matching module is further used for:

[0050] identifying consulting keywords from the consulting sentences in the consulting request;

[0051] The similarities between the consulting keywords and the atomic nodes on the atomic map are calculated respectively, so as to select the target atomic nodes with the highest similarity.

[0052] Optionally, the response module is further used for:

[0053] According to the atomic nodes connected to the target atomic node, the target information or fill-in item information corresponding to the atomic node is obtained, thereby responding to the consultation request;

[0054] and / or,

[0055] The atomic knowledge associated with the target atomic node is used as the response result corresponding to the consultation statement, thereby responding to the consultation request.

[0056] Optionally, the response module is further used for:

[0057] According to the atomic nodes connected to the target atomic node, the target information corresponding to the atomic node is obtained, and the target information is used as the response result corresponding to the consultation statement, thereby responding to the consultation request;

[0058] or,

[0059] According to the atomic nodes connected to the target atomic node, the filling item information corresponding to the atomic node is obtained, and a service order is generated according to the filling item information to respond to the consultation request.

[0060] According to another aspect of an embodiment of the present invention, there is further provided an electronic device, including:

[0061] one or more processors;

[0062] a storage device for storing one or more programs,

[0063] When the one or more programs are executed by the one or more processors, the one or more processors implement the method described in any one of the above embodiments.

[0064] According to another aspect of an embodiment of the present invention, a computer-readable medium is provided, on which a computer program is stored. When the program is executed by a processor, the method described in any one of the above embodiments is implemented.

[0065] An embodiment of the above invention has the following advantages or beneficial effects: because it adopts a technical means of constructing an atomic map according to process information, target information and fill-in item information, matching the target atomic node on the atomic map according to the consultation statement, and thus responding to the consultation request, it overcomes the technical problem in the prior art that it is impossible to provide different degrees of answers to the granularity of the question. The embodiment of the present invention accurately locates the user's consultation content on the atomic map, identifies the specific information of the user's consultation based on the located atomic nodes, and obtains the response result based on the association relationship between the atomic nodes, so it can provide different degrees of answers to the granularity of the consultation question, thereby improving the accuracy of the response content and enhancing the user experience.

[0066] The further effects of the above-mentioned non-conventional optional manner will be described below in conjunction with the specific implementation manner. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] The accompanying drawings are used to better understand the present invention and do not constitute an improper limitation of the present invention.

[0068] Figure 1 is a schematic diagram of the main process of the method for processing a consultation request according to an embodiment of the present invention;

[0069] Figure 2is a schematic diagram of the structure of a process tree according to an embodiment of the present invention;

[0070] Figure 3 is a schematic diagram of the structure of an atomic map according to an embodiment of the present invention;

[0071] Figure 4 is a schematic diagram of the main process of a method for processing a consultation request according to a reference embodiment of the present invention;

[0072] Figure 5 is a schematic diagram of main modules of an apparatus for processing a consultation request according to an embodiment of the present invention;

[0073] Figure 6 is an exemplary system architecture diagram to which embodiments of the present invention may be applied;

[0074] Figure 7 It is a schematic diagram of the structure of a computer system of a terminal device or a server suitable for implementing an embodiment of the present invention. DETAILED DESCRIPTION

[0075] The following is a description of exemplary embodiments of the present invention in conjunction with the accompanying drawings, including various details of the embodiments of the present invention 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 invention. Similarly, for clarity and conciseness, the description of well-known functions and structures is omitted in the following description.

[0076] Figure 1 FIG. 1 is a schematic diagram of the main flow of a method for processing a consultation request according to an embodiment of the present invention. As an embodiment of the present invention, Figure 1 As shown, the method for processing a consultation request may include:

[0077] Step 101, constructing an atomic map according to process information, target information and fill-in item information. In the atomic map, atomic nodes with association are directly connected;

[0078] First, an atomic map is constructed based on the existing process information, target information, and filling information, and when new process information, target information, and filling item information appear, the atomic map can also be updated.

[0079] Optionally, step 101 may include: constructing a process tree based on process information; marking the association between each process node in the process tree and the target information and fill-in item information; using the marked results as training samples to train the dual-tower model and construct an atomic map; wherein, in the atomic map, atomic nodes with associations are directly connected. In an embodiment of the present invention, a process tree is first constructed based on process information. Generally speaking, a set of process information can construct a process tree, and the process nodes on the process tree are the process nodes in the process information. Figure 2 As shown in the figure, each process node is extracted from the process information related to return and exchange, so as to construct a process tree related to return and exchange. Then the association between each process node in the process tree and the target information and fill-in item information is annotated. The annotated labels can be associated or unassociated. Finally, the annotated results are vectorized and the annotated results are used as training samples to train the dual-tower model. At the same time, an atomic map is constructed based on the annotated results. Figure 3 As shown, in the atomic map, atomic nodes with associations are directly connected.

[0080] Optionally, the target information refers to conditional information, for example, the compensation process can be triggered only when the following three conditions are met:

[0081] Condition 1: Amount < 3000;

[0082] Condition 2: The shopping time does not exceed 7 days;

[0083] Condition 3: The product is unopened.

[0084] For example, there are three conditions that must be met when buying an item:

[0085] Condition 1: The user's registration duration is greater than 3 years;

[0086] Condition 2: Total consumption is greater than 50,000;

[0087] Condition 3: Consumption times > 20 times.

[0088] Optionally, the fill-in items may include the reason for compensation, refund method, etc.

[0089] In an embodiment of the present invention, label learning of the twin tower model is based on entity spacing, and the calculation formula is as follows:

[0090]

[0091] Wherein, W represents the node positioning parameter to be learned, V represents the vector of target information or filler information, and R represents the vector of process nodes. d and m are constants, the value of d can be 2, and the value of m can be 5, which is not limited in the embodiment of the present invention.

[0092] Optionally, after training the twin-tower model with the annotation results as training samples and constructing the atomic map, it also includes: for newly added process nodes, using the trained twin-tower model to calculate the correlation between the newly added process nodes and the target information and the filling item information, thereby updating the atomic nodes on the atomic map; for newly added target information, using the trained twin-tower model to calculate the correlation between the newly added target information and the process nodes, thereby updating the atomic nodes on the atomic map; for newly added filling item information, using the trained twin-tower model to calculate the correlation between the newly added filling item information and the process nodes, thereby updating the atomic nodes on the atomic map. After constructing the atomic map based on the existing process information, target information, and filling item information, when newly added process information, target information, and filling item information appear, the trained twin-tower model is used to calculate whether the newly added process information, target information, and filling item information are correlated with the existing process information, target information, and filling item information. If so, the correlation is updated to the atomic map, that is, a new atomic node is added to the atomic map.

[0093] Specifically, if a new process node appears, the trained dual-tower model is used to calculate the correlation between the new process node and the existing target information and fill-in item information. If there is a correlation, a new atomic node is added to the atomic map, and the new atomic node is connected to the existing atomic node with correlation, thereby updating the atomic node on the atomic map; if new target information appears, the trained dual-tower model is used to calculate the correlation between the new target information and the existing process node. If there is a correlation, a new atomic node is added to the atomic map, and the new atomic node is connected to the existing atomic node with correlation, thereby updating the atomic node on the atomic map; if new fill-in item information appears, the trained dual-tower model is used to calculate the correlation between the new fill-in item information and the existing process node. If there is a correlation, a new atomic node is added to the atomic map, and the new atomic node is connected to the existing atomic node with correlation, thereby updating the atomic node on the atomic map. Therefore, once new process information, target information and fill-in item information appear, their correlation can be calculated by the trained dual-tower model, thereby ensuring the reliability and accuracy of the atomic map.

[0094] Optionally, after training the double-tower model with the annotation results as training samples and constructing the atomic map, it also includes: disassembling the document information to obtain the atomic knowledge of each slot; for each atomic node on the atomic map, respectively calculating the similarity between the atomic node and the atomic knowledge of each slot, and associating the atomic knowledge with a similarity greater than a similarity threshold with the atomic node. In an embodiment of the present invention, after constructing the atomic map, the atomic nodes can be further associated with knowledge. Specifically, firstly, the content information of the document is disassembled to obtain the atomic knowledge of each slot (such as slots such as title, keyword, background, target and time), and then for each atomic node on the atomic map, respectively calculate the similarity between the atomic node and the atomic knowledge of each slot, if the similarity is greater than a preset similarity threshold, then associate the corresponding atomic knowledge with the atomic node, so that when the atomic node is located, the atomic knowledge associated with the atomic node can be quickly queried.

[0095] The following is an example of document information:

[0096] 1. Background and Purpose

[0097] This standard was formulated to guide front-line customer service staff in handling scenarios involving compensation for third-party (POP) orders and to clarify the operating methods and application paths for compensation.

[0098] II. Scope of application

[0099] Applicable to retail groups - all staff of the central operations department

[0100] 3. Applicable Scenarios

[0101] Compensation for POP orders outside the scope of dispute forms

[0102] IV. Effective Date

[0103] January 16, 2019

[0104] The slots of the document information include background purpose, applicable scope, applicable scenario and effective time. By disassembling the document information, the atomic knowledge corresponding to each slot can be obtained.

[0105] Step 102: receiving a consultation request from a user, and matching a target atomic node on the atomic map according to a consultation statement in the consultation request.

[0106] After receiving the consultation request sent by the terminal user, the server parses it to obtain the consultation statement in the consultation request, and then matches the target atomic node on the atomic map according to the keywords involved in the consultation statement, so as to accurately identify the user's consultation request.

[0107] Optionally, the target atomic node is matched on the atomic map according to the consulting statement in the consulting request, including: identifying consulting keywords from the consulting statement in the consulting request; respectively calculating the similarity between the consulting keywords and each atomic node on the atomic map, thereby filtering out the target atomic node with the highest similarity. When a consulting behavior occurs, whether it is a telephone consultation or an online consultation, the consulting keywords are identified from the consulting statement in the user's consulting request, and then the similarity between each consulting keyword and each atomic node on the atomic map is respectively calculated, thereby filtering out the target atomic node with the highest similarity. When a user calls in, the call time and call pin information are included, and the call pin information is a mobile phone number or a user pin.

[0108] Step 103, obtaining a response result corresponding to the consultation statement according to the atomic nodes connected to the target atomic node and / or the atomic knowledge associated with the target atomic node, thereby responding to the consultation request.

[0109] After matching the target atomic node on the atomic map, the corresponding response result of the consultation statement is obtained according to the atomic nodes connected to the target atomic node and / or the atomic knowledge associated with the target atomic node, and then the user's consultation request is responded to according to the response result.

[0110] Optionally, according to the atomic node connected to the target atomic node and / or the atomic knowledge associated with the target atomic node, the response result corresponding to the consultation statement is obtained, thereby responding to the consultation request, including: according to the atomic node connected to the target atomic node, the target information or fill-in item information corresponding to the atomic node is obtained, thereby responding to the consultation request; and / or, the atomic knowledge associated with the target atomic node is used as the response result corresponding to the consultation statement, thereby responding to the consultation request. If the atomic node directly connected to the target atomic node is the target information or fill-in item information, the user's consultation request is responded to according to the target information or fill-in item information. If the target atomic node is associated with atomic knowledge, the atomic knowledge is used as the response result corresponding to the consultation statement to respond to the consultation request.

[0111] Optionally, according to the atomic node connected to the target atomic node, the target information or fill-in item information corresponding to the atomic node is obtained, so as to respond to the consultation request, including: according to the atomic node connected to the target atomic node, the target information corresponding to the atomic node is obtained, and the target information is used as the response result corresponding to the consultation statement, so as to respond to the consultation request; or, according to the atomic node connected to the target atomic node, the fill-in item information corresponding to the atomic node is obtained, and a service order is generated according to the fill-in item information, so as to respond to the consultation request. Figure 3As shown, for example, if the target atomic node is compensation, and the atomic nodes adjacent to it have conditions 1, 2 and 3, the contents of these three conditions are returned to the user terminal as the response result; for another example, if the atomic node adjacent to the target atomic node is fill-in item information, a service order is automatically generated based on the fill-in item information.

[0112] According to the various embodiments described above, it can be seen that the embodiments of the present invention solve the technical problem that the prior art cannot provide different degrees of answers to questions according to the granularity of the questions by constructing an atomic map based on process information, target information and fill-in item information, matching the target atomic nodes on the atomic map according to the consultation statement, and thus responding to the consultation request. The embodiments of the present invention accurately locate the user's consultation content on the atomic map, identify the specific information of the user's consultation based on the located atomic nodes, and obtain the response result based on the association relationship between the atomic nodes. Therefore, it is possible to provide different degrees of answers to the granularity of the consultation question, thereby improving the accuracy of the response content and enhancing the user experience.

[0113] Figure 4 FIG. 1 is a schematic diagram of the main flow of a method for processing a consultation request according to a reference embodiment of the present invention. As another embodiment of the present invention, Figure 4 As shown, the method for processing a consultation request may include:

[0114] Step 401, constructing a process tree according to the process information.

[0115] First, a process tree is constructed based on the existing process information. The process nodes on the process tree are the process nodes in the process information.

[0116] Step 402, marking the association between each process node in the process tree and the target information and the fill-in item information.

[0117] Mark the association between each process node in the process tree and the target information and fill-in item information. The marked label can be associated or unassociated.

[0118] Step 403, using the labeled results as training samples to train the twin-tower model and construct an atomic map.

[0119] The annotation results are vectorized and used as training samples to train the dual-tower model. At the same time, an atomic map is constructed based on the annotation results.

[0120] Step 404, disassemble the document information to obtain the atomic knowledge of each slot.

[0121] After the atomic map is constructed, the knowledge association of atomic nodes can be further performed by disassembling document information. By disassembling document information, the atomic knowledge of each slot in the document information can be obtained.

[0122] Step 405 , for each atomic node in the atomic map, respectively calculate the similarity between the atomic node and the atomic knowledge of each slot, and associate the atomic knowledge with a similarity greater than a similarity threshold with the atomic node.

[0123] For each atomic node on the atomic map, the similarity between the atomic node and the atomic knowledge of each slot is calculated respectively. If the similarity is greater than the preset similarity threshold, the corresponding atomic knowledge is associated with the atomic node, so that when the atomic node is located, the atomic knowledge associated with the atomic node can be quickly queried.

[0124] Step 406, updating the atomic map according to the newly added process nodes, target information and fill item information.

[0125] Specifically, for newly added process nodes, the trained twin-tower model is used to calculate the correlation between the newly added process nodes and the target information and the fill-in item information, so as to update the atomic nodes on the atomic map; for newly added target information, the trained twin-tower model is used to calculate the correlation between the newly added target information and the process nodes, so as to update the atomic nodes in the atomic map; for newly added fill-in item information, the trained twin-tower model is used to calculate the correlation between the newly added fill-in item information and the process nodes, so as to update the atomic nodes in the atomic map. Therefore, once newly added process information, target information, and fill-in item information appear, their correlation can be calculated by the trained twin-tower model, so as to ensure the reliability and accuracy of the atomic map.

[0126] Step 407: receiving a consultation request from a user, and identifying consultation keywords from the consultation sentences in the consultation request.

[0127] After receiving the consultation request sent by the terminal user, the server parses it to obtain the consultation sentence in the consultation request, and then identifies the consultation keywords from the consultation sentence.

[0128] Step 408 , respectively calculating the similarity between the consulting keyword and each atomic node in the atomic map, thereby selecting a target atomic node with the highest similarity.

[0129] According to the keywords involved in the consultation sentence, the target atomic nodes are matched on the atomic map, so as to accurately identify the user's consultation needs.

[0130] Step 409, obtaining a response result corresponding to the consultation statement according to the atomic nodes connected to the target atomic node and / or the atomic knowledge associated with the target atomic node, thereby responding to the consultation request.

[0131] Optionally, according to the atomic node connected to the target atomic node and / or the atomic knowledge associated with the target atomic node, the response result corresponding to the consultation statement is obtained, thereby responding to the consultation request, including: according to the atomic node connected to the target atomic node, the target information or fill-in item information corresponding to the atomic node is obtained, thereby responding to the consultation request; and / or, the atomic knowledge associated with the target atomic node is used as the response result corresponding to the consultation statement, thereby responding to the consultation request. If the atomic node directly connected to the target atomic node is the target information or fill-in item information, the user's consultation request is responded to according to the target information or fill-in item information. If the target atomic node is associated with atomic knowledge, the atomic knowledge is used as the response result corresponding to the consultation statement to respond to the consultation request.

[0132] In addition, the specific implementation content of the method for processing a consultation request in a reference embodiment of the present invention has been described in detail in the method for processing a consultation request described above, so the repeated content will not be described here.

[0133] Figure 5 is a schematic diagram of main modules of an apparatus for processing a consultation request according to an embodiment of the present invention, such as Figure 5 As shown, the device 500 for processing consultation requests includes a construction module 501, a matching module 502 and a response module 503; wherein the construction module 501 is used to construct an atomic map according to process information, target information and fill-in item information; wherein, in the atomic map, atomic nodes with association are directly connected; the matching module 502 is used to receive a user's consultation request, and match a target atomic node on the atomic map according to the consultation statement in the consultation request; the response module 503 is used to obtain a response result corresponding to the consultation statement according to the atomic node connected to the target atomic node and / or the atomic knowledge associated with the target atomic node, thereby responding to the consultation request.

[0134] Optionally, the construction module 501 is further used for:

[0135] Build a process tree based on process information;

[0136] Marking the association between each process node in the process tree and the target information and the fill-in item information;

[0137] The labeled results are used as training samples to train the dual-tower model and construct an atomic map.

[0138] Optionally, the construction module 501 is further used for:

[0139] After training the twin-tower model and constructing the atomic map using the labeled results as training samples,

[0140] For a newly added process node, the trained dual-tower model is used to calculate the correlation between the newly added process node and the target information and the fill-in item information, thereby updating the atomic node on the atomic map;

[0141] For the newly added target information, the trained dual-tower model is used to calculate the correlation between the newly added target information and the process nodes, thereby updating the atomic nodes on the atomic map;

[0142] For the newly added filling item information, the trained dual-tower model is used to calculate the correlation between the newly added filling item information and the process nodes, so as to update the atomic nodes on the atomic map.

[0143] Optionally, the construction module 501 is further used for:

[0144] After training the dual-tower model and constructing the atomic map using the annotation results as training samples, the document information is disassembled to obtain the atomic knowledge of each slot;

[0145] For each atomic node on the atomic map, the similarity between the atomic node and the atomic knowledge of each slot is calculated respectively, and the atomic knowledge with a similarity greater than a similarity threshold is associated with the atomic node.

[0146] Optionally, the matching module 502 is further configured to:

[0147] identifying consulting keywords from the consulting sentences in the consulting request;

[0148] The similarities between the consulting keywords and the atomic nodes on the atomic map are calculated respectively, so as to select the target atomic nodes with the highest similarity.

[0149] Optionally, the response module 503 is further configured to:

[0150] According to the atomic nodes connected to the target atomic node, the target information or fill-in item information corresponding to the atomic node is obtained, thereby responding to the consultation request;

[0151] and / or,

[0152] The atomic knowledge associated with the target atomic node is used as the response result corresponding to the consultation statement, thereby responding to the consultation request.

[0153] Optionally, the response module 503 is further configured to:

[0154] According to the atomic nodes connected to the target atomic node, the target information corresponding to the atomic node is obtained, and the target information is used as the response result corresponding to the consultation statement, thereby responding to the consultation request;

[0155] or,

[0156] According to the atomic nodes connected to the target atomic node, the filling item information corresponding to the atomic node is obtained, and a service order is generated according to the filling item information to respond to the consultation request.

[0157] According to the various embodiments described above, it can be seen that the embodiments of the present invention solve the technical problem that the prior art cannot provide different degrees of answers to questions according to the granularity of the questions by constructing an atomic map based on process information, target information and fill-in item information, matching the target atomic nodes on the atomic map according to the consultation statement, and thus responding to the consultation request. The embodiments of the present invention accurately locate the user's consultation content on the atomic map, identify the specific information of the user's consultation based on the located atomic nodes, and obtain the response result based on the association relationship between the atomic nodes. Therefore, it is possible to provide different degrees of answers to the granularity of the consultation question, thereby improving the accuracy of the response content and enhancing the user experience.

[0158] It should be noted that the specific implementation content of the device for processing consultation requests in the present invention has been described in detail in the method for processing consultation requests described above, so the repeated content will not be described here.

[0159] Figure 6 An exemplary system architecture 600 is shown to which the method for processing a consultation request or the apparatus for processing a consultation request according to the embodiments of the present invention can be applied.

[0160] like Figure 6 As shown, system architecture 600 may include terminal devices 601, 602, 603, network 604 and server 605. Network 604 is used to provide a medium for communication links between terminal devices 601, 602, 603 and server 605. Network 604 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0161] Users can use terminal devices 601, 602, and 603 to interact with server 605 through network 604 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 601, 602, and 603, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, social platform software, etc. (only examples).

[0162] The terminal devices 601 , 602 , and 603 may be various electronic devices having a display screen and supporting web browsing, including but not limited to smart phones, tablet computers, laptop computers, and desktop computers, etc.

[0163] The server 605 may be a server that provides various services, such as a backend management server (only an example) that provides support for shopping websites browsed by users using the terminal devices 601, 602, and 603. The backend management server may analyze and process the received data such as the item information query request, and feed back the processing results (such as target push information, item information - only an example) to the terminal device.

[0164] It should be noted that the method for processing a consultation request provided in the embodiment of the present invention is generally executed by the server 605 , and accordingly, the apparatus for processing a consultation request is generally disposed in the server 605 .

[0165] It should be understood that Figure 6 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0166] Reference below Figure 7 , which shows a schematic diagram of the structure of a computer system 700 of a terminal device suitable for implementing an embodiment of the present invention. Figure 7 The terminal device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0167] like Figure 7 As shown, the computer system 700 includes a central processing unit (CPU) 701, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 702 or a program loaded from a storage part 708 into a random access memory (RAM) 703. In the RAM 703, various programs and data required for the operation of the system 700 are also stored. The CPU 701, the ROM 702, and the RAM 703 are connected to each other via a bus 704. An input / output (I / O) interface 705 is also connected to the bus 704.

[0168] The following components are connected to the I / O interface 705: an input section 706 including a keyboard, a mouse, etc.; an output section 707 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 708 including a hard disk, etc.; and a communication section 709 including a network interface card such as a LAN card, a modem, etc. The communication section 709 performs communication processing via a network such as the Internet. A drive 710 is also connected to the I / O interface 705 as needed. A removable medium 711, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 710 as needed, so that a computer program read therefrom is installed into the storage section 708 as needed.

[0169] In particular, according to the embodiments disclosed in the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present invention include a computer program, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 709, and / or installed from the removable medium 711. When the computer program is executed by the central processing unit (CPU) 701, the above-mentioned functions defined in the system of the present invention are executed.

[0170] It should be noted that the computer-readable medium shown in the present invention may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present invention, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. In the present invention, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0171] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program according to various embodiments of the present invention. In this regard, each box in the flow chart or block diagram can represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flow chart, and the combination of the boxes in the block diagram or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0172] The modules involved in the embodiments of the present invention may be implemented by software or hardware. The modules described may also be set in a processor, for example, it may be described as: a processor includes a building module, a matching module and a response module, wherein the names of these modules do not constitute a limitation on the modules themselves in some cases.

[0173] As another aspect, the present invention also provides a computer-readable medium, which may be included in the device described in the above embodiment; or it may exist independently and not be assembled into the device. The above computer-readable medium carries one or more programs. When the above one or more programs are executed by a device, the device implements the following method: construct an atomic map according to process information, target information and fill-in item information; wherein, in the atomic map, atomic nodes with associations are directly connected; receive a user's consultation request, and match a target atomic node on the atomic map according to the consultation statement in the consultation request; obtain a response result corresponding to the consultation statement according to the atomic nodes connected to the target atomic node and / or the atomic knowledge associated with the target atomic node, thereby responding to the consultation request.

[0174] According to the technical solution of the embodiment of the present invention, because it adopts a technical means of building an atomic map according to process information, target information and fill-in item information, matching the target atomic node on the atomic map according to the consultation statement, and thus responding to the consultation request, it overcomes the technical problem in the prior art that it is impossible to provide different degrees of answers to the granularity of the question. The embodiment of the present invention accurately locates the user's consultation content on the atomic map, identifies the specific information of the user's consultation based on the located atomic nodes, and obtains the response result based on the correlation between the atomic nodes. Therefore, it is possible to provide different degrees of answers to the granularity of the consultation question, thereby improving the accuracy of the response content and enhancing the user experience.

[0175] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions may occur depending on design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for processing a consultation request, characterized in that: include: Constructing an atomic map according to the process information, the target information and the filling item information; wherein in the atomic map, the atomic nodes with association are directly connected; receiving a consultation request from a user, and matching a target atomic node on the atomic map according to a consultation statement in the consultation request; Obtaining a response result corresponding to the consultation statement according to the atomic node connected to the target atomic node and / or the atomic knowledge associated with the target atomic node, thereby responding to the consultation request; Build an atomic map based on process information, target information, and fill item information, including: Build a process tree based on process information; Marking the association between each process node in the process tree and the target information and the fill-in item information; Use the labeled results as training samples to train the dual-tower model and construct an atomic map; For a newly added process node, the trained dual-tower model is used to calculate the correlation between the newly added process node and the target information and the fill-in item information, thereby updating the atomic node on the atomic map; For the newly added target information, the trained dual-tower model is used to calculate the correlation between the newly added target information and the process nodes, thereby updating the atomic nodes in the atomic map; For the newly added filling item information, the trained dual-tower model is used to calculate the correlation between the newly added filling item information and the process nodes, so as to update the atomic nodes in the atomic map.

2. The method according to claim 1, characterized in that After training the dual-tower model using the labeled results as training samples and constructing the atomic map, it also includes: Decompose the document information to obtain the atomic knowledge of each slot; For each atomic node in the atomic map, the similarity between the atomic node and the atomic knowledge of each slot is calculated respectively, and the atomic knowledge with a similarity greater than a similarity threshold is associated with the atomic node.

3. The method according to claim 1, characterized in that Matching a target atomic node in the atomic map according to the consulting statement in the consulting request includes: identifying consulting keywords from the consulting sentences in the consulting request; The similarities between the consulting keywords and the atomic nodes in the atomic map are calculated respectively, so as to select the target atomic nodes with the highest similarity.

4. The method according to claim 1, characterized in that: Obtaining a response result corresponding to the consultation statement according to the atomic node connected to the target atomic node and / or the atomic knowledge associated with the target atomic node, thereby responding to the consultation request, including: According to the atomic nodes connected to the target atomic node, the target information or the filling item information corresponding to the atomic node is obtained, so as to respond to the consultation request; and / or, The atomic knowledge associated with the target atomic node is used as the response result corresponding to the consultation statement, thereby responding to the consultation request.

5. The method according to claim 4, characterized in that According to the atomic node connected to the target atomic node, the target information or the filling item information corresponding to the atomic node is obtained, so as to respond to the consultation request, including: According to the atomic nodes connected to the target atomic node, the target information corresponding to the atomic node is obtained, and the target information is used as the response result corresponding to the consultation statement, thereby responding to the consultation request; or, According to the atomic nodes connected to the target atomic node, the filling item information corresponding to the atomic node is obtained, and a service order is generated according to the filling item information to respond to the consultation request.

6. A device for processing a consultation request, characterized in that: include: A construction module, used to construct an atomic map according to process information, target information and fill-in item information; wherein, in the atomic map, atomic nodes with association are directly connected; A matching module, configured to receive a user's consultation request and match a target atomic node in the atomic map according to a consultation statement in the consultation request; A response module, configured to obtain a response result corresponding to the consultation statement according to the atomic node connected to the target atomic node and / or the atomic knowledge associated with the target atomic node, thereby responding to the consultation request; The building blocks are also used to: Build a process tree based on process information; Marking the association between each process node in the process tree and the target information and the fill-in item information; Use the labeled results as training samples to train the dual-tower model and construct an atomic map; For a newly added process node, the trained dual-tower model is used to calculate the correlation between the newly added process node and the target information and the fill-in item information, thereby updating the atomic node on the atomic map; For the newly added target information, the trained dual-tower model is used to calculate the correlation between the newly added target information and the process nodes, thereby updating the atomic nodes on the atomic map; For the newly added filling item information, the trained dual-tower model is used to calculate the correlation between the newly added filling item information and the process nodes, so as to update the atomic nodes on the atomic map.

7. An electronic device, characterized in that: include: one or more processors; a storage device for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 5.

8. A computer readable medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.

Images