Dialogue flow mining method, device, electronic device and computer storage medium

Abstract

An embodiment of the present application provides a method, device, electronic device, and computer storage medium for dialogue flow mining, which provides a visual interface; in response to a triggering operation on a first virtual node in the interface, determine a first type of cluster intention that has a mapping relationship with the node intention of the first virtual node; perform intention clustering on the corresponding first dialogue file to obtain a second dialogue file and a second type of cluster intention; add an initial node intention corresponding to the second type of cluster intention and a second initial virtual node corresponding to the initial node intention, and form a topological connection between virtual nodes by using the second initial virtual node as a downstream node of the first virtual node; in response to an editing operation on the selected second initial virtual node, adjust the second initial virtual node to a second updated virtual node; establish a mapping relationship between the node intention of the second updated virtual node and the second type of cluster intention, and obtain and display the node intention of the second updated virtual node based on the mapping relationship. The mining efficiency is improved.

Description

Technical Field

[0001] Embodiments of the present application relate to the field of Internet technologies, and in particular, to a method, an apparatus, an electronic device, and a computer storage medium for mining a conversation flow. Background Art

[0002] At present, intelligent chatbots have been widely used in various scenarios. An intelligent chatbot can communicate with a customer through natural language to complete a specified task of the customer. In the actual operation process, the intelligent chatbot operates based on a pre-mined conversation flow. Specifically: first, obtain a conversation statement input by the customer, and perform intent recognition on the statement. Then, locate the node corresponding to the above intent in the pre-constructed conversation flow, and further determine the next node (target node) of the node in the entire process. Finally, perform corresponding operations based on the intent corresponding to the target node and output a corresponding response statement.

[0003] The process of mining a conversation flow is a process of clustering the intents of multi-round conversation data. When a specific mining round is triggered, a set of conversation data will be clustered by intent to obtain multiple intent clusters, and each intent cluster corresponds to a set of sub-conversation data with the same intent.

[0004] To improve the mining efficiency, usually, parallel mining is independently triggered for multiple sets of conversation data. However, in each mining round, the user may perform manual editing operations such as merging and deleting on the mining results. The above operations will affect the mining results. To ensure the accuracy of the mining results, it is necessary to perform intent clustering on the conversation data that has performed the above editing operations again. In this way, the mining efficiency of the entire conversation flow will be drastically reduced. Summary of the Invention

[0005] In view of this, embodiments of the present application provide a method for mining a conversation flow to at least partially solve the above problems.

[0006] According to a first aspect of the embodiments of the present application, there is provided a method for mining a conversation flow for mining an original conversation file, the method including:

[0007] Provide a visual interface for displaying at least one virtual node and node intent information of the at least one virtual node;

[0008] In response to a trigger operation on a first virtual node, determine first cluster intent information having a mapping relationship with the node intent information of the first virtual node;

[0009] Perform intent clustering on a first conversation file corresponding to the first cluster intent information to obtain multiple second conversation files and second cluster intent information corresponding to each second conversation file;

[0010] Add initial node intention information corresponding to each second - type cluster intention information respectively in the visualization interface, and second initial virtual nodes corresponding to each initial node intention information respectively, and use the second initial virtual nodes as downstream nodes of the first virtual node to form a topological connection relationship among the virtual nodes;

[0011] Receive an editing operation input from the user for one or more selected second initial virtual nodes;

[0012] In response to the editing operation on the selected second initial virtual node, adjust each second initial virtual node to a second updated virtual node to update the topological connection relationship;

[0013] Establish a mapping relationship between the node intention information of each second updated virtual node and the second - type cluster intention information, and obtain and display the node intention information of each second updated virtual node based on the mapping relationship.

[0014] According to the second aspect of the embodiments of the present application, there is provided a dialogue flow mining device for mining an original dialogue file, including:

[0015] A visualization interface providing module for providing a visualization interface, which is used to display at least one virtual node and the node intention information of the at least one virtual node;

[0016] A first - type cluster intention information determining module for determining, in response to a triggering operation on a first virtual node, first - type cluster intention information having a mapping relationship with the node intention information of the first virtual node;

[0017] An intention clustering module for performing intention clustering on the first dialogue file corresponding to the first - type cluster intention information to obtain a plurality of second dialogue files and second - type cluster intention information corresponding to each second dialogue file;

[0018] A topological connection relationship obtaining module for adding initial node intention information corresponding to each second - type cluster intention information respectively in the visualization interface, and second initial virtual nodes corresponding to each initial node intention information respectively, and using the second initial virtual nodes as downstream nodes of the first virtual node to form a topological connection relationship among the virtual nodes;

[0019] A receiving module for receiving an editing operation input from the user for one or more selected second initial virtual nodes;

[0020] A topological connection relationship updating module for adjusting each second initial virtual node to a second updated virtual node in response to the editing operation on the selected second initial virtual node to update the topological connection relationship;

[0021] The mapping relationship establishment and node intention information display module is used to establish the mapping relationship between the node intention information of each second updated virtual node and the second type of cluster intention information, and obtain and display the node intention information of each second updated virtual node based on the mapping relationship.

[0022] According to the third aspect of the embodiments of the present application, an electronic device is provided, including: a processor, a memory, a communication interface, and a communication bus. The processor, the memory, and the communication interface complete mutual communication through the communication bus; the memory is used to store at least one executable instruction, and the executable instruction causes the processor to execute the operations corresponding to the dialogue flow mining method described in the first aspect.

[0023] According to the fourth aspect of the embodiments of the present application, a computer storage medium storing a computer program for live interaction is provided. A computer program is stored thereon, and when the program is executed by a processor, it implements the dialogue flow mining method of the real-time live room described in the first aspect.

[0024] According to the fifth aspect of the embodiments of the present application, a computer program product for live interaction is provided, including computer instructions, and the computer instructions instruct a computing device to execute the operations corresponding to the dialogue flow mining method of the real-time live room described in the first aspect.

[0025] According to the dialogue flow mining method provided by the embodiments of the present application, the data involved in the dialogue flow mining process is divided into 4 levels: the virtual node level, the node intention information level, the cluster intention information level, and the dialogue file level. Among them, the following association relationships exist between the above levels in sequence: the virtual nodes in the virtual node level correspond one-to-one with the node intention information in the node intention information level; the node intention information in the node intention information level has a mapping relationship with the cluster intention information in the cluster intention information level; the cluster intention information in the cluster intention information level corresponds one-to-one with the dialogue files in the dialogue file level.

[0026] Display virtual nodes in the virtual node hierarchy and corresponding node intent information in the node intent information hierarchy in the visualization interface for users to perform an interactive mining process; when the first virtual node is triggered, that is, when mining starts, according to the above-mentioned association relationship, sequentially determine: the first type of cluster intent information that has a mapping relationship with the node intent information of this node, and the corresponding first dialogue file; then perform intent clustering on the first dialogue file to obtain a second dialogue file and the corresponding second type of cluster intent information, and add the initial node intent information corresponding to the second type of cluster intent information and the corresponding second initial virtual node (the downstream node of the first virtual node, or in other words, the child node) to the interface to form a preliminary mining result; when the selected second initial virtual node is triggered for an editing operation, that is, when the user corrects the preliminary mining result, only adjust the virtual nodes in the virtual node hierarchy, the node intent information in the corresponding node intent information hierarchy, and the mapping relationship between the corresponding node intent information and the cluster intent information in the cluster intent information hierarchy, without having to re-perform intent clustering on the dialogue files in the dialogue file hierarchy. Therefore, the efficiency of dialogue flow mining is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings described below are only some embodiments recorded in the embodiments of the present application. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings.

[0028] Figure 1 It is a schematic diagram of the management structure for visual dialogue flow data mining according to Embodiment 1 of the present application;

[0029] Figure 2 It is a schematic diagram of the scenario of a dialogue flow mining method according to Embodiment 1 of the present application;

[0030] Figure 3 It is a schematic diagram of the process for triggering a mining task in visual dialogue flow mining according to Embodiment 1 of the present application;

[0031] Figure 4 It is a schematic diagram of editing an operation on a node during an interactive annotation process according to Embodiment 1 of the present application;

[0032] Figure 5 It is a schematic diagram of the multi-round mining process of visual dialogue flow mining according to Embodiment 1 of the present application;

[0033] Figure 6 It is a flowchart of the steps of a dialogue flow mining method according to Embodiment 1 of the present application;

[0034] Figure 7It is a structural block diagram of a dialogue flow mining device according to Embodiment 2 of the present application;

[0035] Figure 8 It is a schematic structural diagram of an electronic device according to Embodiment 3 of the present application. Specific implementation manners

[0036] In order to enable those skilled in the art to better understand the technical solutions in the embodiments of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the embodiments of the present application, all other embodiments obtained by those of ordinary skill in the art shall fall within the protection scope of the embodiments of the present application.

[0037] The following further illustrates the specific implementation of the embodiments of the present application with reference to the accompanying drawings of the embodiments of the present application.

[0038] Embodiment 1

[0039] See Figure 1 , Figure 1 It is a schematic management structure diagram of visual dialogue flow data mining provided by the embodiments of the present application. To facilitate the understanding of the technical solutions of the present application, first, in combination with Figure 1 the management structure of visual dialogue flow data mining in the embodiments of the present application will be described.

[0040] In the embodiments of the present application, the dialogue flow data mining is abstracted into four layers: a virtual node layer, a node intention information layer, a cluster intention information layer, and a dialogue file layer. Among them, the virtual node layer includes one or more virtual nodes. When there are multiple virtual nodes, the virtual node layer also includes the topological connection relationships between the virtual nodes, which are used to display to the user in the visualization interface. The node intention information layer includes node intention information corresponding one-to-one to each virtual node in the virtual node layer. The node intention information is used to represent the intention of the corresponding virtual node, and this layer is also displayed in the visualization interface. The cluster intention information layer includes cluster intention information, and there is a one-to-one or one-to-many mapping relationship between the node intention information and the cluster intention information. Specifically, a node intention information may have a mapping relationship with only one cluster intention information (in this case, the node intention information may be the same as the cluster intention information), or a node intention information may have a mapping relationship with multiple cluster intention information at the same time (in this case, the node intention information may be the intention information obtained by aggregating various cluster intention information). The dialogue file layer includes each dialogue file corresponding one-to-one to the cluster intention information, that is, after intention clustering, the dialogue files with the same intention are used as a cluster, and their specific dialogue files are stored in the dialogue file layer for use in the dialogue mining task. The intention of the cluster is stored in the cluster intention information layer as the corresponding cluster intention information.

[0041] Next, based on the Figure 1 management structure, a general introduction to the dialogue flow mining process of the embodiments of the present application will be given first. Assume that after the first round of mining on the original virtual node a0, two first virtual nodes, a11 and a12, are formed in the virtual node layer. For the first virtual node a11, there is a node intention information b11 corresponding in the node intention information layer, and the node intention information b11 has a mapping relationship with the first cluster intention information c11, c12, and c13 in the cluster intention information layer. The first cluster intention information c11, c12, and c13 respectively correspond to d11, d12, and d13 in the dialogue file layer. When the second round of mining is triggered for the first virtual node a11, then d11, d12, and d13 in the dialogue file layer will be used as a whole to be clustered, and intention clustering is performed to obtain the corresponding second dialogue file d2. Correspondingly, there is a second cluster intention information c2 corresponding to the second dialogue file d2 in the cluster intention information layer, there is a node intention information b2 corresponding to the second cluster intention information c2 in the node intention information layer, and there is a second virtual node a2 corresponding to the node intention information b2 in the virtual node layer. After that, according to the need, by analogy, the subsequent mining sub-processes of the 3rd round and so on are triggered for the virtual nodes in the virtual node layer until the mining process ends, which will not be elaborated here.

[0042] See Figure 2 , Figure 2 , which is a schematic diagram of the scenario of a dialogue flow mining method according to Embodiment 1 of the present application. For ease of understanding, first, in combination with Figure 2 , the application scenario of the dialogue flow mining method provided in Embodiment 1 of the present application will be explained and described.

[0043] The embodiments of the present application provide a visualization interface as shown in Figure 2 (a). This interface is used to display at least one virtual node (in the figure, a first virtual node is taken as an example for illustration, which does not limit the number of virtual nodes in the embodiments of the present application) and the node intention information of the at least one virtual node. Among them, the first virtual node is in the Figure 1 virtual node layer, and the corresponding node intention information is in the Figure 1 node intention information layer. When the user performs a trigger operation on the first virtual node, the first type of cluster intention information (in the Figure 2 cluster intention information layer) that has a mapping relationship with the node intention information of the first virtual node is determined, and the first dialogue file (in the Figure 2 dialogue file layer) corresponding to the first type of cluster intention information. Then, intention clustering is performed on the first dialogue file to obtain two second dialogue files (in the figure, only two second dialogue files are taken as an example for illustration, which does not limit the number of second dialogue files) and the second type of cluster intention information corresponding to each second dialogue file. Then, as shown in Figure 2 (b), the initial node intention information corresponding to each second type of cluster intention information and the second initial virtual node corresponding to each initial node intention information are added to the visualization interface, and each second initial virtual node is used as the downstream node of the first virtual node to form the topological connection relationship between the virtual nodes. When receiving the user's editing operation (such as a merge operation) input for the two second initial virtual nodes (in the figure, two second initial virtual nodes are taken as an example for illustration, which does not limit the number of selected second initial virtual nodes) in the visualization interface shown in Figure 2 (b), as shown in Figure 2 (c), the second initial virtual node is adjusted to a second updated virtual node to update the topological connection relationship; a mapping relationship is established between the node intention information of the second updated virtual node and the second type of cluster intention information, and the node intention information of each second updated virtual node is obtained and displayed based on the mapping relationship.

[0044] It should be noted that in the embodiments of the present application, in the visualization interface, the specific display form of the topological connection relationship between the virtual nodes is not limited. For example, it can be displayed in the form of a flowchart as shown in Figure 2 , or in other forms such as a table.Figure 2 It is only illustrated by way of a flowchart and does not limit the embodiments of the present application.

[0045] The following introduces the process of visual dialogue flow mining by the user in the embodiments of the present application. Refer to Figures 3 to 5 , the embodiments of the present application can provide a visual interaction interface to display one or more virtual nodes and corresponding node intention information for the user to perform mining or editing operations. In the process of visual dialogue flow mining, it is usually necessary to perform multiple rounds of interactive mining subprocesses on the original dialogue file to generate the dialogue flow of the final task-based dialogue. For a single interactive mining subprocess, it generally includes two steps: the first step is to trigger the mining task; the second step is interactive annotation. Refer to Figure 3 , the first step is: when the user performs a trigger operation on a certain node (such as the start node) through the visual interface, the dialogue file corresponding to the node is automatically clustered by intention, so as to obtain multiple dialogue sub-files and corresponding sub-nodes (such as nodes 1 to 5). In the second step, the user can perform manual intervention annotation on the mining result through the visual interaction interface, that is, perform an editing operation on the node, so as to obtain the annotated mining result, where, refer to Figure 4 , the editing operation on the node can include: merge operation, delete operation, etc.

[0046] Refer to Figure 5 , Figure 5 is a schematic diagram of the multi-round dialogue flow mining process. A total of 2 rounds of interactive mining subprocesses are performed in this figure. In the first round, the user can perform a merge operation on the visual interaction interface, so that the newly generated node 2 and node 3 are merged into node 3', and at the same time perform a delete operation to delete node 5, and finally obtain node 1, node 3', and node 4. In the second round, the user can trigger the mining operations of nodes 1, 3', and 4 respectively, so as to generate node 1A and node 1B based on node 1, generate node 3'A based on node 3', and generate node 4A based on node 4.

[0047] Refer to Figure 6 , Figure 6 is a flowchart of the steps of a dialogue flow mining method according to Embodiment 1 of the present application; specifically, the dialogue flow mining method provided in this embodiment is used to mine the original dialogue file, where the original dialogue file may refer to multiple dialogue statements input by the customer, and each dialogue statement may include multi-round dialogue data.

[0048] The dialogue flow mining method includes the following steps:

[0049] Step 602, provide a visual interface, and the visual interface is used to display at least one virtual node and the node intention information of at least one virtual node.

[0050] In the embodiments of the present application, there is a one-to-one correspondence between the virtual nodes in the visual interface and the node intention information of the virtual nodes. The node intention information of the virtual nodes represents the intention represented by the virtual nodes.

[0051] For example: If the original dialogue file is a dialogue file between a customer service and a customer generated in the scenario of purchasing air tickets, the node intention information may be inquiries about travel time, number of travelers, identity information of travelers, type of air tickets, and so on.

[0052] Step 604: In response to the triggering operation on the first virtual node, determine the first type of cluster intention information that has a mapping relationship with the node intention information of the first virtual node.

[0053] In the embodiments of the present application, there is no limitation on how to perform the triggering operation on the first virtual node. For example: Operations such as clicking and dragging on the first virtual node displayed in the visual interface can be used as triggering operations to trigger the mining process of the dialogue flow.

[0054] As pointed out in the above description for Figure 1 In the embodiments of the present application, there may be a one-to-one or one-to-many mapping relationship between the node intention information of the virtual nodes and the cluster intention information. Specifically, the node intention information of a virtual node may only have a mapping relationship with one piece of cluster intention information, or it may simultaneously have a mapping relationship with multiple pieces of cluster intention information. When the node intention information of a virtual node simultaneously has a mapping relationship with multiple pieces of cluster intention information, the node intention information of the virtual node can be aggregated from the above-mentioned multiple pieces of cluster intention information. For example, the node intention information of the virtual node is: purchasing air tickets, and the cluster intention information having a mapping relationship with it may include 2 pieces, which are respectively: purchasing air tickets of airline A and purchasing air tickets of airline B.

[0055] Step 606: Perform intention clustering on the first dialogue file corresponding to the first type of cluster intention information to obtain multiple second dialogue files and the second type of cluster intention information corresponding to each second dialogue file.

[0056] The second type of cluster intention information corresponds one-to-one with the second dialogue file, that is: one second dialogue file corresponds to one second type of cluster intention information, and the second type of cluster intention information represents the intention corresponding to the second dialogue file.

[0057] Further, in the embodiments of the present application, after obtaining multiple second dialogue files, each second dialogue file can be stored separately.

[0058] Step 608: Add the initial node intention information corresponding to each second - type cluster intention information and the second initial virtual nodes corresponding to each initial node intention information in the visualization interface, and use the second initial virtual nodes as the downstream nodes of the first virtual node to form the topological connection relationship among the virtual nodes.

[0059] In this step, for each second - type cluster intention information obtained in step 606, the corresponding initial node intention information and the corresponding second initial virtual node can be added in the visualization interface. At the same time, the newly added second initial virtual node is used as the child node (downstream node) of the original first virtual node in the visualization interface, so as to form the topological connection relationship between the first virtual node and the second initial virtual node.

[0060] There is a one - to - one correspondence among the initial node intention information displayed in the visualization interface, the second initial virtual nodes displayed in the visualization interface, and the second - type cluster intention information not displayed in the visualization interface.

[0061] Step 610: Receive the editing operation input from the user for one or more selected second initial virtual nodes.

[0062] Step 612: In response to the editing operation on the selected second initial virtual node, adjust each second initial virtual node to a second updated virtual node to update the topological connection relationship.

[0063] Step 614: Establish the mapping relationship between the node intention information of each second updated virtual node and the second - type cluster intention information, and obtain and display the node intention information of each second updated virtual node based on the mapping relationship.

[0064] The following is an explanation of steps 610 - 614:

[0065] After step 608, the second initial virtual nodes formed in the visualization interface are all virtual nodes corresponding to the second - type cluster intention information obtained by the automatic intention clustering operation. That is to say, the above - mentioned mining process is all automatically completed without manual correction. However, according to the different service scopes actually provided by the user, it may be necessary for the user to annotate and correct the mining results, such as performing editing operations such as merging, deleting, and restoring the second initial virtual nodes according to the initial node intention information, so as to obtain the second updated virtual nodes and adjust the mapping relationship between the node intention information of each second updated virtual node and the second - type cluster intention information.

[0066] Taking the air ticket purchase scenario as an example, assume that 2 second initial virtual nodes are newly added in the visual interface after step 608, and the corresponding initial node intention information is: purchasing an air ticket of airline A and purchasing an air ticket of airline B. Considering the actual services provided by the customer, there is no substantial difference in the service processing flow between purchasing an air ticket of airline A and purchasing an air ticket of airline B. At this time, the user can perform a merging operation on the 2 second initial virtual nodes, thereby updating the second initial virtual nodes to second updated virtual nodes in the visual interface and adjusting the mapping relationship between the node intention information of the second updated virtual nodes and the second type of cluster intention information.

[0067] For another example, assume that the initial node intention information corresponding to the above 2 second initial virtual nodes is: purchasing an air ticket and handling check-in service respectively, but the actual services provided by the customer currently do not include the check-in service. At this time, the user can perform a deletion operation on the second initial virtual node corresponding to the check-in service. Subsequently, as the types of services provided by the customer continue to expand, the check-in service may be added. At this time, the user can perform a restoration operation on the above deleted second initial virtual node corresponding to the check-in service and adjust the mapping relationship between the node intention information of the restored second initial virtual node and the second type of cluster intention information.

[0068] Furthermore, in the embodiments of the present application, the editing operation is only used to adjust the mapping relationship between the node intention information of each second updated virtual node and the second type of cluster intention information, without changing the storage state of multiple second dialogue files. For example, for the merging operation, only the mapping relationship between the node intention information of the second updated virtual node and the second type of cluster intention information is adjusted, and the second dialogue files themselves are not merged. For the deletion operation, only the to-be-deleted second initial virtual node and its corresponding initial node intention information are deleted, and none of the second dialogue files themselves are deleted.

[0069] Optionally, in some embodiments, the mapping relationship between the node intention information of each second updated virtual node and the second type of cluster intention information can be adjusted based on the specific type of the editing operation.

[0070] Specifically, for the merging operation, the process of obtaining the second updated virtual node and the process of adjusting the mapping relationship between the node intention information of the second updated virtual node and the second type of cluster intention information may include:

[0071] In response to the merging operation on the selected multiple second initial virtual nodes, delete the selected multiple second initial virtual nodes and add a merged virtual node in the visual interface; use the unselected second initial virtual nodes and the merged virtual node as the second updated virtual nodes;

[0072] If the second updated virtual node is the merged virtual node, determine the second type of cluster intention information corresponding to each selected second initial virtual node as the second type of cluster intention information having a mapping relationship with this second updated virtual node; aggregate the second type of cluster intention information corresponding to each selected second initial virtual node to obtain the node intention information of this second updated virtual node and display it;

[0073] If the second updated virtual node is a non - selected second initial virtual node, use the initial node intention information corresponding to this non - selected second initial virtual node as the node intention information of this second updated virtual node; determine the second type of cluster intention information corresponding to this non - selected second initial virtual node as the second type of cluster intention information having a mapping relationship with the node intention information of this second updated virtual node.

[0074] For the deletion operation, the process of obtaining the second updated virtual node and the process of adjusting the mapping relationship between the node intention information of the second updated virtual node and the second type of cluster intention information may include:

[0075] In response to the deletion operation on the selected second initial virtual node, delete the selected second initial virtual node and the initial node intention information corresponding to the selected second initial virtual node in the visualization interface; retain the second type of cluster intention information corresponding to the selected second initial virtual node as the soft - deleted second type of cluster intention information; retain the second dialogue file corresponding to the soft - deleted second type of cluster intention information; determine the non - selected second initial virtual nodes as the second updated virtual nodes; use the initial node intention information corresponding to each non - selected second initial virtual node as the node intention information of each second updated virtual node; determine the second type of cluster intention information corresponding to each non - selected second initial virtual node as the second type of cluster intention information having a mapping relationship with the node intention information of each second updated virtual node.

[0076] In addition, for the recovery operation, the virtual node update process and the process of adjusting the mapping relationship between the node intention information of the updated virtual node and the second type of cluster intention information may include:

[0077] In response to the recovery operation on the deleted second initial virtual node, determine the soft - deleted second type of cluster intention information corresponding to the deleted second initial virtual node; add the recovered second virtual node in the visualization interface; determine the soft - deleted second type of cluster intention information corresponding to the deleted second initial virtual node as the second type of cluster intention having a mapping relationship with the node intention information of the recovered second virtual node; display the soft - deleted second type of cluster intention information corresponding to the deleted second initial virtual node as the node intention information of the recovered second virtual node.

[0078] In the embodiments of the present application, virtual nodes in the virtual node hierarchy and corresponding node intention information in the node intention information hierarchy can be displayed in a visual interface for users to perform an interactive mining process. When the first virtual node is triggered, that is, when the mining starts, the following are determined in sequence according to the above association relationship: the first type of cluster intention information having a mapping relationship with the node intention information of the node, and the corresponding first dialogue file; then, intention clustering is performed on the first dialogue file to obtain a second dialogue file and the corresponding second type of cluster intention information, and the initial node intention information corresponding to the second type of cluster intention information and the corresponding second initial virtual node (the downstream node of the first virtual node, or in other words, the child node) are added to the interface to form a preliminary mining result. When an edit operation is triggered on the selected second initial virtual node, that is, when the user corrects the preliminary mining result, only the virtual nodes in the virtual node hierarchy, the node intention information in the corresponding node intention information hierarchy, and the mapping relationship between the corresponding node intention information and the cluster intention information in the cluster intention information hierarchy are adjusted, without having to perform intention clustering on the dialogue files in the dialogue file hierarchy again. Therefore, the efficiency of dialogue flow mining is improved.

[0079] Embodiment 2

[0080] See Figure 7 , Figure 7A structural block diagram of a dialogue flow mining device according to Embodiment 2 of the present application. The dialogue flow mining device provided by the embodiments of the present application includes: a visualization interface providing module 702, configured to provide a visualization interface for displaying at least one virtual node and node intention information of at least one virtual node. A first type of cluster intention information determining module 704, configured to determine, in response to a triggering operation on a first virtual node, a first type of cluster intention information having a mapping relationship with the node intention information of the first virtual node. An intention clustering module 706, configured to perform intention clustering on a first dialogue file corresponding to the first type of cluster intention information to obtain a plurality of second dialogue files and second type of cluster intention information corresponding to each second dialogue file. A topological connection relationship obtaining module 708, configured to add initial node intention information corresponding to each second type of cluster intention information and second initial virtual nodes corresponding to each initial node intention information in the visualization interface, and use the second initial virtual nodes as downstream nodes of the first virtual node to form a topological connection relationship between each virtual node. A receiving module 710, configured to receive an editing operation input from a user for one or more selected second initial virtual nodes. A topological connection relationship updating module 712, configured to, in response to an editing operation on a selected second initial virtual node, adjust each second initial virtual node to a second updated virtual node to update the topological connection relationship. A mapping relationship establishing and node intention information displaying module 714, configured to establish a mapping relationship between the node intention information of each second updated virtual node and the second type of cluster intention information, and obtain and display the node intention information of each second updated virtual node based on the mapping relationship.

[0081] Optionally, in some embodiments, when the mapping relationship establishing and node intention information displaying module 714 executes the step of establishing a mapping relationship between the node intention information of each second updated virtual node and the second type of cluster intention information, it is specifically configured to: adjust the mapping relationship between the node intention information of each second updated virtual node and the second type of cluster intention information based on the type of the editing operation.

[0082] Optionally, in some embodiments, the dialogue flow mining device further includes: a storage module, configured to store a plurality of second dialogue files respectively after obtaining the plurality of second dialogue files.

[0083] Optionally, in some embodiments, the editing operation is used to adjust the mapping relationship between the node intention information of each second updated virtual node and the second type of cluster intention information without changing the storage state of the plurality of second dialogue files.

[0084] Optionally, in some embodiments, the types of the editing operation include: a merging operation, a deleting operation, or a restoring operation.

[0085] Optionally, in some embodiments, the topology connection relationship update module 712 is specifically configured to: in response to a merge operation on a selected plurality of second initial virtual nodes, delete the selected plurality of second initial virtual nodes, and add a merged virtual node in the visualization interface; use the unselected second initial virtual nodes and the merged virtual node as second updated virtual nodes.

[0086] Optionally, in some embodiments, the mapping relationship establishment and node intent information display module 714 is specifically configured to: if the second updated virtual node is a merged virtual node, determine the second type of cluster intent information corresponding to each selected second initial virtual node as the second type of cluster intent information having a mapping relationship with the second updated virtual node; aggregate the second type of cluster intent information corresponding to each selected second initial virtual node to obtain the node intent information of the second updated virtual node and display it.

[0087] Optionally, in some embodiments, the topology connection relationship update module 712 is specifically configured to: in response to a deletion operation on a selected second initial virtual node, delete the selected second initial virtual node and the initial node intent information corresponding to the selected second initial virtual node in the visualization interface; retain the second type of cluster intent information corresponding to the selected second initial virtual node as the soft-deleted second type of cluster intent information; retain the second dialogue file corresponding to the soft-deleted second type of cluster intent information; use the unselected second initial virtual nodes as the second updated virtual nodes; the mapping relationship establishment and node intent information display module 714 is specifically configured to use the initial node intent information corresponding to each unselected second initial virtual node as the node intent information of each second updated virtual node; and determine the second type of cluster intent information corresponding to each unselected second initial virtual node as the second type of cluster intent information having a mapping relationship with the node intent information of each second updated virtual node.

[0088] Optionally, in some embodiments, the dialogue flow mining device further includes: a soft-deleted second type of cluster intent information, configured to determine the soft-deleted second type of cluster intent information corresponding to a deleted second initial virtual node in response to a recovery operation on the deleted second initial virtual node; a restored node addition module, configured to add a restored second virtual node in the visualization interface; a mapping relationship determination module, configured to determine the soft-deleted second type of cluster intent information corresponding to the deleted second initial virtual node as the second type of cluster intent having a mapping relationship with the node intent information of the restored second virtual node; and a restored node intent information display module, configured to display the soft-deleted second type of cluster intent information corresponding to the deleted second initial virtual node as the node intent information of the restored second virtual node.

[0089] Optionally, in some embodiments, the topological connection relationships between virtual nodes are presented in a visualization interface in the form of a flowchart or a table.

[0090] The dialogue flow mining device of this embodiment is used to implement the corresponding dialogue flow mining methods in the foregoing multiple method embodiments, and has the beneficial effects of the corresponding method embodiments, which will not be elaborated here. In addition, the function implementation of each module in the dialogue flow mining device of this embodiment can be referred to the description of the corresponding part in the foregoing method embodiments, which will not be elaborated here either.

[0091] Embodiment III

[0092] Referring to Figure 8 , a schematic structural diagram of an electronic device according to Embodiment III of the present application is shown. The specific implementation of the electronic device in the specific embodiments of the present application is not limited.

[0093] As Figure 8 shown, the electronic device may include: a processor 802, a communications interface 804, a memory 806, and a communication bus 808. The processor 802, the communications interface 804, and the memory 806 communicate with each other through the communication bus 808. The communications interface 804 is used to communicate with other electronic devices or servers. The processor 802 is used to execute the program 810, and specifically can execute the relevant steps in the foregoing dialogue flow mining method embodiments. Specifically, the program 810 may include program code, and the program code includes computer operation instructions.

[0094] The processor 802 may be a CPU, or a specific integrated circuit ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement the embodiments of the present application. One or more processors included in the intelligent device may be of the same type of processor, such as one or more CPUs; or may be of different types of processors, such as one or more CPUs and one or more ASICs.

[0095] The memory 806 is used to store the program 810. The memory 806 may include a high-speed RAM memory, and may also include non-volatile memory, such as at least one disk memory.

[0096] The program 810 can specifically be used to cause the processor 802 to perform the following operations: provide a visual interface for displaying at least one virtual node and the node intention information of at least one virtual node; in response to a triggering operation on the first virtual node, determine the first type of cluster intention information having a mapping relationship with the node intention information of the first virtual node; perform intention clustering on the first dialogue file corresponding to the first type of cluster intention information to obtain a plurality of second dialogue files and the second type of cluster intention information corresponding to each second dialogue file; add the initial node intention information corresponding to each second type of cluster intention information and the second initial virtual node corresponding to each initial node intention information to the visual interface, and use the second initial virtual node as the downstream node of the first virtual node to form a topological connection relationship between virtual nodes; receive an editing operation input from the user for one or more selected second initial virtual nodes; in response to the editing operation on the selected second initial virtual node, adjust each second initial virtual node to a second updated virtual node to update the topological connection relationship; establish a mapping relationship between the node intention information of each second updated virtual node and the second type of cluster intention information, and obtain and display the node intention information of each second updated virtual node based on the mapping relationship.

[0097] For the specific implementation of each step in the program 810, reference can be made to the corresponding steps and descriptions in the corresponding units in the foregoing embodiment of the dialogue flow mining method, which will not be elaborated herein. Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the devices and modules described above can refer to the corresponding process descriptions in the foregoing method embodiments, which will not be repeated herein.

[0098] Through the electronic device of this embodiment, virtual nodes in the virtual node level and corresponding node intention information in the node intention information level are displayed in the visual interface for the user to perform an interactive mining process; when the first virtual node is triggered, that is, when the mining starts, according to the above-mentioned association relationship, the following are determined in sequence: the first type of cluster intention information having a mapping relationship with the node intention information of the node, and the corresponding first dialogue file; then intention clustering is performed on the first dialogue file to obtain a second dialogue file and the corresponding second type of cluster intention information, and the initial node intention information corresponding to the second type of cluster intention information and the corresponding second initial virtual node (the downstream node, or in other words, the child node, of the first virtual node) are added to the interface to form a preliminary mining result; when an editing operation is triggered on the selected second initial virtual node, that is, when the user corrects the preliminary mining result, only the virtual nodes in the virtual node level, the node intention information in the corresponding node intention information level, and the mapping relationship between the corresponding node intention information and the cluster intention information in the cluster intention information level are adjusted, without having to perform intention clustering on the dialogue files in the dialogue file level again. Therefore, the efficiency of dialogue flow mining is improved.

[0099] The embodiments of the present application further provide a computer storage medium storing a computer program for dialogue flow mining, on which a computer program is stored. When the program is executed by a processor, any one of the dialogue flow mining methods in the above-mentioned multiple method embodiments is implemented.

[0100] The embodiments of the present application further provide a computer program product for dialogue flow mining, including computer instructions, and the computer instructions direct a computing device to perform operations corresponding to any one of the dialogue flow mining methods in the above-mentioned multiple method embodiments.

[0101] It should be noted that, according to the needs of implementation, each component / step described in the embodiments of the present application can be split into more components / steps, or two or more components / steps or partial operations of components / steps can be combined into new components / steps to achieve the purpose of the embodiments of the present application.

[0102] The method according to the embodiments of the present application can be implemented in hardware, firmware, or be implemented as software or computer code that can be stored in a recording medium (such as a CD ROM, RAM, floppy disk, hard disk, or magneto-optical disk), or be implemented as computer code originally stored in a remote recording medium or a non-transitory machine-readable medium and to be downloaded through a network and stored in a local recording medium, so that the method described herein can be processed by such software stored on a recording medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware (such as an ASIC or FPGA). It can be understood that a computer, a processor, a microprocessor controller, or programmable hardware includes a storage component (such as a RAM, a ROM, a flash memory, etc.) that can store or receive software or computer code. When the software or computer code is accessed and executed by the computer, the processor, or the hardware, the dialogue flow mining method described herein is implemented. In addition, when a general-purpose computer accesses the code for implementing the dialogue flow mining method shown herein, the execution of the code converts the general-purpose computer into a dedicated computer for executing the dialogue flow mining method shown herein.

[0103] Those of ordinary skill in the art can realize that the units and method steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. A professional technician can use different methods for each specific application to implement the described functions, but such implementation should not be considered to exceed the scope of the embodiments of the present application.

[0104] The above embodiments are only used to illustrate the embodiments of the present application, rather than to limit the embodiments of the present application. Those of ordinary skill in the relevant technical field can also make various changes and modifications without departing from the spirit and scope of the embodiments of the present application. Therefore, all equivalent technical solutions also belong to the scope of the embodiments of the present application. The patent protection scope of the embodiments of the present application shall be defined by the claims.

Claims

1. A method for mining a dialogue flow for mining an original dialogue file, including: Provide a visualization interface for displaying at least one virtual node and the node intention information of the at least one virtual node; In response to a trigger operation on a first virtual node, determine a first type of cluster intention information that has a mapping relationship with the node intention information of the first virtual node; Perform intention clustering on the first dialogue file corresponding to the first type of cluster intention information to obtain a plurality of second dialogue files and the second type of cluster intention information corresponding to each second dialogue file; Add initial node intention information corresponding to each second type of cluster intention information and second initial virtual nodes corresponding to each initial node intention information to the visualization interface, and use the second initial virtual nodes as downstream nodes of the first virtual node to form a topological connection relationship between the virtual nodes; Receive user input of an editing operation for one or more selected second initial virtual nodes; In response to the editing operation on the selected second initial virtual node, adjust each second initial virtual node to a second updated virtual node to update the topological connection relationship; Establish a mapping relationship between the node intention information of each second updated virtual node and the second type of cluster intention information, and obtain and display the node intention information of each second updated virtual node based on the mapping relationship, where the editing operation is used to adjust the mapping relationship between the node intention information of each second updated virtual node and the second type of cluster intention information without changing the storage state of the plurality of second dialogue files.

2. The method according to claim 1, wherein, The establishing the mapping relationship between the node intention information of each second updated virtual node and the second type of cluster intention information includes: Adjust the mapping relationship between the node intention information of each second updated virtual node and the second type of cluster intention information based on the type of the editing operation.

3. The method according to claim 2, wherein, After obtaining the plurality of second dialogue files, the method further includes: Store the plurality of second dialogue files separately.

4. The method according to any one of claims 1-3, wherein, The types of the editing operation include: a merge operation, a deletion operation, or a restoration operation.

5. The method according to claim 4, wherein, The adjusting each second initial virtual node to a second updated virtual node in response to the editing operation on the selected second initial virtual node includes: In response to a merge operation on the selected plurality of second initial virtual nodes, delete the selected plurality of second initial virtual nodes and add a merged virtual node to the visualization interface; Use the unselected second initial virtual nodes and the merged virtual node as the second updated virtual nodes.

6. The method according to claim 5, wherein, The adjusting the mapping relationship between the node intention information of each second updated virtual node and the second type of cluster intention information based on the type of the editing operation, and obtaining and displaying the node intention information of each second updated virtual node based on the mapping relationship includes: If the second updated virtual node is a merged virtual node, determine the second type of cluster intention information corresponding to each selected second initial virtual node as the second type of cluster intention information that has a mapping relationship with the second updated virtual node; Aggregate the second type of cluster intention information corresponding to each selected second initial virtual node to obtain and display the node intention information of the second updated virtual node.

7. The method according to claim 4, wherein, In response to an editing operation on a selected second initial virtual node, adjusting each second initial virtual node to a second updated virtual node includes: In response to a deletion operation on a selected second initial virtual node, deleting the selected second initial virtual node and the initial node intention information corresponding to the selected second initial virtual node in the visualization interface; Retaining the second type of cluster intention information corresponding to the selected second initial virtual node as soft-deleted second type of cluster intention information; retaining the second dialogue file corresponding to the soft-deleted second type of cluster intention information; Determining the non-selected second initial virtual nodes as second updated virtual nodes; Based on the type of the editing operation, adjusting the mapping relationship between the node intention information and the second type of cluster intention information of each second updated virtual node, and obtaining and displaying the node intention information of each second updated virtual node based on the mapping relationship, includes: Taking the initial node intention information corresponding to each non-selected second initial virtual node as the node intention information of each second updated virtual node; Respectively determining the second type of cluster intention information corresponding to each non-selected second initial virtual node as the second type of cluster intention information having a mapping relationship with the node intention information of each second updated virtual node.

8. The method according to claim 7, wherein, The method further includes: In response to a recovery operation on a deleted second initial virtual node, determining the soft-deleted second type of cluster intention information corresponding to the deleted second initial virtual node; Adding a recovered second virtual node in the visualization interface; Determining the soft-deleted second type of cluster intention information corresponding to the deleted second initial virtual node as the second type of cluster intention having a mapping relationship with the node intention information of the recovered second virtual node; Displaying the soft-deleted second type of cluster intention information corresponding to the deleted second initial virtual node as the node intention information of the recovered second virtual node.

9. The method according to claim 1, wherein, The topological connection relationship between the virtual nodes is displayed in the visualization interface in the form of a flowchart or a table.

10. A dialogue flow mining device for mining an original dialogue file, comprising: A visualization interface providing module for providing a visualization interface for displaying at least one virtual node and the node intention information of the at least one virtual node; A first type of cluster intention information determining module for, in response to a triggering operation on a first virtual node, determining the first type of cluster intention information having a mapping relationship with the node intention information of the first virtual node; An intention clustering module for clustering the intention of the first dialogue file corresponding to the first type of cluster intention information to obtain a plurality of second dialogue files and the second type of cluster intention information corresponding to each second dialogue file; A topological connection relationship obtaining module for adding, in the visualization interface, the initial node intention information corresponding to each second type of cluster intention information respectively, and the second initial virtual nodes corresponding to each initial node intention information respectively, and taking the second initial virtual nodes as the downstream nodes of the first virtual node to form the topological connection relationship between the virtual nodes; A receiving module for receiving user input of an editing operation on one or more selected second initial virtual nodes; A topological connection relationship update module, configured to, in response to an editing operation on a selected second initial virtual node, adjust each second initial virtual node to a second updated virtual node, so as to update the topological connection relationship; A mapping relationship establishment and node intention information display module, configured to establish a mapping relationship between the node intention information of each second updated virtual node and the second type of cluster intention information, and obtain and display the node intention information of each second updated virtual node based on the mapping relationship, wherein the editing operation is used to adjust the mapping relationship between the node intention information of each second updated virtual node and the second type of cluster intention information, without changing the storage state of the plurality of second dialogue files.

11. An electronic device, comprising: A processor, a memory, a communication interface, and a communication bus, where the processor, the memory, and the communication interface complete communication with each other through the communication bus; The memory is used to store at least one executable instruction, and the executable instruction causes the processor to perform the operations corresponding to the dialogue flow mining method according to any one of claims 1-9.

12. A computer storage medium storing a computer program for dialogue flow mining, having a computer program stored thereon, and when the program is executed by a processor, implementing the dialogue flow mining method according to any one of claims 1-9.

13. A computer program product for dialogue flow mining, comprising computer instructions, and the computer instructions direct a computing device to perform operations corresponding to the dialogue flow mining method according to any one of claims 1-9.

Images