CN113792819A - Multitask model-based node reusable intelligent outbound method and system - Google Patents
Multitask model-based node reusable intelligent outbound method and system Download PDFInfo
- Publication number
- CN113792819A CN113792819A CN202111276483.4A CN202111276483A CN113792819A CN 113792819 A CN113792819 A CN 113792819A CN 202111276483 A CN202111276483 A CN 202111276483A CN 113792819 A CN113792819 A CN 113792819A
- Authority
- CN
- China
- Prior art keywords
- multitask
- matching model
- intention
- training
- flow tree
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 57
- 238000012549 training Methods 0.000 claims abstract description 141
- 239000013598 vector Substances 0.000 claims description 53
- 230000006870 function Effects 0.000 claims description 15
- 230000015654 memory Effects 0.000 claims description 15
- 239000011159 matrix material Substances 0.000 claims description 9
- 238000004458 analytical method Methods 0.000 claims description 8
- 238000004422 calculation algorithm Methods 0.000 claims description 8
- 238000004590 computer program Methods 0.000 claims description 8
- 238000005457 optimization Methods 0.000 claims description 8
- 239000000126 substance Substances 0.000 claims description 3
- 238000002372 labelling Methods 0.000 claims description 2
- 230000008569 process Effects 0.000 description 16
- 238000012545 processing Methods 0.000 description 12
- 238000010586 diagram Methods 0.000 description 7
- 238000004364 calculation method Methods 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 6
- 239000000463 material Substances 0.000 description 6
- 238000003860 storage Methods 0.000 description 5
- 238000013473 artificial intelligence Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 4
- 238000013145 classification model Methods 0.000 description 3
- 238000004891 communication Methods 0.000 description 2
- 238000012790 confirmation Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000007621 cluster analysis Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 238000005111 flow chemistry technique Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 238000004904 shortening Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 230000008685 targeting Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/24323—Tree-organised classifiers
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Machine Translation (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention provides a node reusable intelligent outbound method and a system based on a multitask model, wherein the method comprises the following steps: obtaining the training corpus and using the training corpus as a training set; obtaining a first flow tree according to a first outbound scene; obtaining a second flow tree according to a second calling scene, and taking each non-leaf node of the first flow tree and the second flow tree as a subtask of multi-task training; initializing a multitask intention matching model, wherein the multitask intention matching model comprises an encoder and a plurality of task layers; obtaining a first multitask intention matching model; training and updating a task layer which is different from the subtasks of the first flow tree in the first multitask intention matching model through the training set to obtain a second multitask intention matching model; and carrying out intelligent outbound according to the second multitask intention matching model. The technical problems that in the prior art, the rule is hard to reuse and cannot be updated in real time are solved.
Description
Technical Field
The invention relates to the field of intelligent outbound, in particular to a node reusable intelligent outbound method and system based on a multitask model.
Background
In recent years, artificial intelligence has been developed rapidly, and in most industrial fields, artificial intelligence has been competent for part of the work. The telephone voice system can automatically dial the telephone and issue the notice, which is a perfect combination of computer technology and voice technology. With the increasing operating costs, telephone customer service centers are also following a major trend of moving from "cost centers to profit centers". When a plurality of service objects at the enterprise end need to carry out service communication or notification, if the service objects are dialed one by manpower, the time and the labor are wasted, and the efficiency is not high. The automatic notification can be performed by using a telephone outgoing call system, and a pre-designed voice or text notification is sent to the other party after the other party is automatically judged to be off-hook. After the other party is judged to hang up, the notification of the next object can be automatically carried out, and the method is very convenient and fast. The intelligent outbound system well solves the problems of high investment cost and low working efficiency at the server side, enterprises can preset the required voice outbound flow according to the business requirements, and the system can realize the flows of identity confirmation, business consultation, business processing and the like of users.
However, in the process of implementing the technical solution of the invention in the embodiments of the present application, the inventors of the present application find that the above-mentioned technology has at least the following technical problems:
the intention recognition based on the rules needs to manually configure the rules for each operation node, but the method is too rigid and has low recognition accuracy, the operation is recognized wrongly if the client is changed slightly, the problems that the same text is under different nodes of the same operation process but has different meanings cannot be solved, the text is difficult to multiplex, and the text cannot be updated in real time.
Disclosure of Invention
Aiming at the defects in the prior art, the embodiment of the application aims to solve the technical problems that the rule-based intention identification in the prior art needs to be manually configured for each dialect node, but the mode is too rigid and the identification accuracy is very low, the identification is wrong if the dialect of a client is slightly changed, the problem that the same text is different in meaning under different nodes of the same dialect process cannot be solved, the multiplexing is difficult, and the real-time updating cannot be realized, so that the language materials and the labels needed for freely configuring different nodes of a process tree are realized through a multi-task intention classification model, the intention labels under different nodes cannot be influenced mutually, and prototype language materials contained under each node can be directly updated to complete the updating and the configuring of the model, the technical effect of the reusability of the model is improved.
In view of the foregoing problems, the present application provides a node reusable intelligent outbound method and system based on a multitasking model.
In one aspect, an embodiment of the present application provides a node reusable intelligent outbound method based on a multitasking model, where the method includes: clustering analysis is carried out on all the training corpuses by using a clustering algorithm, the training corpuses which accord with a preset rule in each category are obtained, and the obtained training corpuses are used as a training set; obtaining a first flow tree according to a first outbound scene; obtaining a second flow tree according to a second call-out scene, wherein each non-leaf node of the first flow tree and each non-leaf node of the second flow tree are respectively used as a subtask of multi-task training; initializing a multitask intention matching model, wherein the multitask intention matching model comprises a plurality of task layers which respectively correspond to each subtask in the flow tree; training the initialized multitask intention matching model by using the corpus of the first flow tree in the training set to obtain a first multitask intention matching model; training and updating a task layer which is different from the subtasks of the first flow tree in the initialized first multitask intention matching model by using the corpus of the second flow tree in the training set to obtain a second multitask intention matching model; and carrying out intelligent outbound according to the second multitask intention matching model.
On the other hand, the embodiment of the present application further provides a node reusable intelligent outbound system based on a multitasking model, wherein the system includes: the first obtaining unit is used for obtaining the training corpora which accord with the preset rules in each category and taking the training corpora as a training set; a second obtaining unit, configured to obtain a first flow tree according to a first outbound scenario; a third obtaining unit, configured to obtain a second flow tree according to a second outbound scenario; a first initialization unit, configured to initialize a multitask intention matching model, where the multitask intention matching model includes a plurality of task layers and corresponds to each subtask in the flow tree respectively; a fourth obtaining unit, configured to train the initialized multitask intention matching model by using the corpus of the first flow tree in the training set, so as to obtain a first multitask intention matching model; a fifth obtaining unit, configured to perform training and updating on a task layer different from the subtasks of the first flow tree in the initialized first multitask intent matching model by using the corpus of the second flow tree in the training set, so as to obtain a second multitask intent matching model; and the first outbound unit is used for carrying out intelligent outbound according to the second multitask intention matching model.
On the other hand, an embodiment of the present application further provides a node reusable intelligent outbound system based on a multitasking model, including a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor implements the steps of the method according to any one of the first aspect when executing the program.
One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages:
clustering analysis is carried out on all the training corpuses by using a clustering algorithm, the training corpuses which accord with a preset rule in each category are obtained, and the obtained training corpuses are used as a training set; obtaining a first flow tree according to a first outbound scene; obtaining a second flow tree according to a second call-out scene, wherein each non-leaf node of the first flow tree and each non-leaf node of the second flow tree are respectively used as a subtask of multi-task training; initializing a multitask intention matching model, wherein the multitask intention matching model comprises a plurality of task layers which respectively correspond to each subtask in the flow tree; training the initialized multitask intention matching model by using the corpus of the first flow tree in the training set to obtain a first multitask intention matching model; training and updating a task layer which is different from the subtasks of the first flow tree in the initialized first multitask intention matching model by using the corpus of the second flow tree in the training set to obtain a second multitask intention matching model; and carrying out intelligent outbound according to the second multitask intention matching model.
The foregoing description is only an overview of the technical solutions of the present application, and the present application can be implemented according to the content of the description in order to make the technical means of the present application more clearly understood, and the following detailed description of the present application is given in order to make the above and other objects, features, and advantages of the present application more clearly understandable.
Drawings
Other features, objects and advantages of the invention will become more apparent upon reading of the detailed description of non-limiting embodiments with reference to the following drawings:
FIG. 1 is a schematic flowchart illustrating a node reusable intelligent outbound method based on a multitasking model according to an embodiment of the present application;
FIG. 2 is a diagram illustrating an embodiment of the present application that optimizes the multitask intent matching model by minimizing a difference between the maximum intra-class distance and the minimum inter-class distance to obtain a first multitask intent matching model;
FIG. 3 is a diagram illustrating an embodiment of the present application in comparison with the node5，node6Corresponding task layer T5，T6Training and updating are carried out, and a second multitask intention matching model is obtained;
FIG. 4 is a diagram illustrating an intelligent outbound call based on the first prediction result according to an embodiment of the present application;
FIG. 5 is a diagram illustrating an example of assigning each subtask with training data and corresponding intent labels to be used;
fig. 6 is a schematic diagram of a data modeling system for improving intelligent marketing efficiency according to an embodiment of the present application.
Fig. 7 is a schematic structural diagram of an exemplary electronic device according to an embodiment of the present application.
Description of reference numerals: a first obtaining unit 11, a second extracting unit 12, a third obtaining unit 13, a first initializing unit 14, a fourth obtaining unit 15, a fifth obtaining unit 16, a first calling-out unit 17, an electronic device 50, a processor 51, a memory 52, an input device 53, and an output device 54.
Detailed Description
The embodiment of the application provides a data modeling method and a data modeling system for improving intelligent marketing efficiency, and solves the technical problems that in the prior art, rule configuration needs to be manually performed on each dialect node based on rule identification, but the mode is too rigid and the identification accuracy is very low, and if the dialect of a client changes slightly, an identification error can be generated, the problem that the same text is different in meaning under different nodes of the same dialect process cannot be solved, multiplexing is difficult, and real-time updating cannot be performed.
Hereinafter, technical solutions in example embodiments of the present application will be described in detail with reference to the accompanying drawings. It should be apparent that the described embodiments are merely some embodiments of the present application and not all embodiments of the present application, and it should be understood that the present application is not limited to the example embodiments described herein. All other embodiments, which can be derived by a person skilled in the art from the embodiments of the present application without making any creative effort, shall fall within the protection scope of the present application.
Summary of the application
In recent years, artificial intelligence has been developed rapidly, and in most industrial fields, artificial intelligence has been competent for part of the work. The telephone voice system can automatically dial the telephone and issue the notice, which is a perfect combination of computer technology and voice technology. With the increasing operating costs, telephone customer service centers are also following a major trend of moving from "cost centers to profit centers". When a plurality of service objects at the enterprise end need to carry out service communication or notification, if the service objects are dialed one by manpower, the time and the labor are wasted, and the efficiency is not high. The automatic notification can be performed by using a telephone outgoing call system, and a pre-designed voice or text notification is sent to the other party after the other party is automatically judged to be off-hook. After the other party is judged to hang up, the notification of the next object can be automatically carried out, and the method is very convenient and fast. The intelligent outbound system well solves the problems of high investment cost and low working efficiency at the server side, enterprises can preset the required voice outbound flow according to the business requirements, and the system can realize the flows of identity confirmation, business consultation, business processing and the like of users. However, in the prior art, the intention identification based on the rules needs to manually configure the rules for each dialect node, but the method is too rigid and has low identification accuracy, and if the client makes a slight change in the dialect, the identification is wrong, so that the problems that the same text is under different nodes of the same dialect process but has different meanings cannot be solved, and the technical problems that the text is difficult to multiplex and cannot be updated in real time are difficult to realize.
For the above technical problems, the general idea of the technical solution provided by the applicant is as follows:
the embodiment of the application provides a node reusable intelligent outbound method and system based on a multitask model, the method is applied to a node reusable intelligent outbound system based on the multitask model, and the method comprises the following steps: clustering analysis is carried out on all the training corpuses by using a clustering algorithm, the training corpuses which accord with a preset rule in each category are obtained, and the obtained training corpuses are used as a training set; obtaining a first flow tree according to a first outbound scene; obtaining a second flow tree according to a second call-out scene, wherein each non-leaf node of the first flow tree and each non-leaf node of the second flow tree are respectively used as a subtask of multi-task training; initializing a multitask intention matching model, wherein the multitask intention matching model comprises a plurality of task layers which respectively correspond to each subtask in the flow tree; training the initialized multitask intention matching model by using the corpus of the first flow tree in the training set to obtain a first multitask intention matching model; training and updating a task layer which is different from the subtasks of the first flow tree in the initialized first multitask intention matching model by using the corpus of the second flow tree in the training set to obtain a second multitask intention matching model; and carrying out intelligent outbound according to the second multitask intention matching model.
Having thus described the general principles of the present application, embodiments thereof will now be described with reference to the accompanying drawings. As can be known to those skilled in the art, with the development of technology and the emergence of new scenarios, the technical solutions provided in the embodiments of the present application are also applicable to similar technical problems.
Example one
As shown in fig. 1, an embodiment of the present application provides a node reusable intelligent outbound method and system based on a multitasking model, where the method is applied to a node reusable intelligent outbound system based on a multitasking model, and the method includes:
step S100: clustering analysis is carried out on all the training corpuses by using a clustering algorithm, the training corpuses which accord with a preset rule in each category are obtained, and the obtained training corpuses are used as a training set;
specifically, the cluster analysis is also called group analysis, which is a statistical analysis method for research, sample, index or classification problems, based on similarity, there is more similarity between patterns in a cluster than between patterns not in the same cluster, that is, before all the training corpuses are obtained, firstly, the domain direction to which the system is applied is determined, after the determination, all the training corpuses are obtained, then, the obtained corpuses are subjected to word segmentation processing, a long text is divided into a plurality of words, secondly, after the text is segmented into words, further conversion into vectors is needed, all words contained in the training corpuses are constructed into a term list, repeated terms are not contained in the list, moreover, for each text, a vector is constructed, the dimension of the vector is the same as that of the term list, the value of the vector is the number of times that each term in the term list appears in the text, and then distributing the clustering center to the most similar clustering according to the similarity of the clustering center, obtaining the training corpus which accords with the preset rule in each category after the distribution is finished, and taking the training corpus as a training set, wherein the obtained training set is used as basic data for model training, performing jargon node setting and intention identification, and the obtaining of the training set is the basis for constructing the multitask model-based node reusable intelligent outbound system.
Step S200: obtaining a first flow tree according to a first outbound scene;
step S300: obtaining a second flow tree according to a second call-out scene, wherein each non-leaf node of the first flow tree and each non-leaf node of the second flow tree are respectively used as a subtask of multi-task training;
specifically, the first outbound scenario refers to selecting any application scenario in the application field of the multitask-model-based node-reusable intelligent outbound system as the first outbound scenario, the second scenario is similar to the first scenario and refers to selecting any application scenario different from the first outbound scenario in the application field of the multitask-model-based node-reusable intelligent outbound system, obtaining a first flow tree according to the first outbound scenario refers to formulating a corresponding flow tree according to a specific service handling flow of the first outbound scenario, obtaining a second flow tree according to the second outbound scenario, the second flow tree refers to formulating a corresponding flow tree according to a specific service handling flow of the second outbound scenario, wherein the flow tree includes leaf nodes and non-leaf nodes, and the leaf nodes included in the flow tree are end points of the flow tree, i.e. service flow ends of the corresponding service scenarios, the non-leaf node is a sub-business process in the business process under the corresponding application scene and is also a sub-task of the multi-task training of the multi-task model to be constructed, the first process tree and the second process tree are obtained to enable the flow of the multi-task model-based node reusable intelligent outbound system to be more standard when in application, and the plurality of sub-tasks corresponding to the non-leaf node are used for processing various possible business modes under the corresponding business scene, so that the business processing flow is more flexible.
Step S400: initializing a multitask intent matching model comprising NnTask layer TnRespectively corresponding to each subtask in the flow tree;
specifically, the multitask intention matching model refers to that for a given plurality of business scenarios, a part of tasks or all tasks have similarity but are less similar, the multitask matching model aims to mutually assist the knowledge contained in the plurality of business scenarios so as to improve the performance of completing each business scenario, the initializing multitask intention matching model refers to firstly making corresponding multitask intention matching model according to the first outbound scenario and the second outbound scenario before training the multitask intention matching model, wherein the multitask intention matching model contains NnNumber of subtasks, i.e. containing NnAnd the multi-task intention matching model is set based on a specific application scene, is more intelligent after being trained by the training set, and can perform corresponding operation processing according to the processed business content.
Step S500: training the initialized multitask intention matching model by using the corpus of the first flow tree in the training set to obtain a first multitask intention matching model;
specifically, the training of the initialized multitask intention matching model by using the corpus of the first flow tree in the training set refers to randomly selecting any multiple objects from the processed training set as initial clustering centers, calculating the remaining unselected objects, allocating the objects to the most similar clusters according to the similarity between the objects and the selected objects, namely, the distance between the objects and the clustering centers, namely, classifying the obtained corpus in the training set according to the similarity, if the objects are found to have more similar clustering centers than the currently classified distance class centers after the allocation is completed, re-classifying according to the similarity principle until the classification result is relatively optimal, extracting the features of the training corpus according to the obtained classification result of the training set, and further obtaining the corresponding feature vectors, different feature vectors correspond to different task layers, intention matching of different texts is achieved, the first multitask intention matching model is obtained, and corresponding expression intents can be obtained by the first multitask intention matching model according to input linguistic data.
Step S600: training and updating a task layer which is different from the subtasks of the first flow tree in the initialized first multitask intention matching model by using the corpus of the second flow tree in the training set to obtain a second multitask intention matching model;
specifically, the corpus of the second flow tree is used in the training set to train and update task layers different from the subtasks of the first flow tree in the initialized first multitask intention matching model, that is, a non-leaf node is randomly selected in the second flow tree as a subtask of the current training, then a data set corresponding to the subtask is used as input data of the current training task, intentions of multiple categories are distinguished from the training set, multiple texts are selected as data required by the training, the input training text is preprocessed to obtain the feature vectors, distances among the feature vectors are calculated, and the maximum value and the minimum value of the distances of the feature vectors in the category where the feature vectors are located, namely the maximum inter-category distance and the minimum inter-category distance, are obtained according to the distances among the feature vectors, optimizing the model by calculating the difference between the maximum inter-class distance and the minimum inter-class distance, reselecting a new subtask node and corresponding training data until an ideal difference is obtained, ending the optimization process to obtain the second multitask intention matching model, training a plurality of related tasks between the first multitask intention model and the second multitask intention model, extracting general features among the tasks in each scene, and transmitting the general features into different subtask scenes for calculation, thereby perfecting the flow processing and text processing results of each application scene.
Step S700: and carrying out intelligent outbound according to the second multitask intention matching model.
Specifically, the second multitask intention matching model is realized on the basis of the first multitask intention matching model, the second multitask intention matching model can configure required corpora for different service scenes, namely different task nodes, the service flow under the corresponding scene is completed, the model is updated once when the service is executed once, and manual periodic updating is not needed, so that the second multitask intention matching model has higher reusability, and intelligent outbound can be performed according to the second multitask intention matching model.
Further, as shown in fig. 2, step S500 further includes:
step S501: obtaining a first subtask task in the first flow treei；
Step S502: from the first subtask taskiRandom selection of NCIntention categories, N being selected from each intention categorySTaking the bar text as a first training set;
step S503: inputting all training texts in the first training set into corresponding task layers T in the multi-task intention matching modeliObtaining a first eigenvector matrix;
step S504: calculating Euclidean distance between each pair of eigenvectors in the first eigenvector matrix;
step S505: for each of said classes, calculating a maximum intra-class distance and a minimum inter-class distance;
step S506: obtaining a difference between the maximum intra-class distance and the minimum inter-class distance for each of the classes;
step S507: and optimizing the multitask intention matching model by minimizing the difference between the maximum intra-class distance and the minimum inter-class distance to obtain a first multitask intention matching model.
Specifically, a first subtask task in the first flow tree is obtainediI.e. randomly selecting a non-leaf node from the first flow tree as a first task, i.e. the first subtask taskiFrom the first subtask taskiRandom selection of NCThe intention category refers to the taskiSelecting N from the data included belowCData as task to the first subtaskiIn each intention category, selecting NSThe item intent labels serve as a first training set, i.e., N for each of the selected itemsCSelecting N from each of the intention categoriesSIntention labels as training data, i.e. selecting NSUsing the text as a first training set, inputting the first training set into an encoder to obtain NC×NSInputting all training texts in the first training set into a corresponding task layer T in the multi-task intention matching modeliMeans that all the training texts in the obtained first training set are input into the first subtask taskiCorresponding task layer TiGet the task layer TiWeighted NC×NSObtaining a first eigenvector matrix according to a plurality of eigenvectors with the length d, then calculating Euclidean distances between the eigenvectors according to the obtained first eigenvector matrix, namely calculating the distance between each eigenvector and other imagination under the same intention category, respectively obtaining the largest intra-class distance and the smallest inter-class distance in each category according to the Euclidean distances between the calculated quantities in each category, then obtaining the difference between the largest intra-class distance and the smallest inter-class distance in each category according to the obtained largest intra-class distance and the smallest inter-class distance, and optimizing the difference between the largest intra-class distance and the smallest inter-class distance as the direction for optimizing the multi-task intention matching model to obtain a first multi-task intention matching model, the first multitask intention matching model is obtained by training according to any service scene and is suitable for service processing under the scene。
Further, step S507 further includes:
step S5071: the optimization function of the first multitask intention matching model is as follows:
wherein the content of the first and second substances,
optimizing a target difference value for the multitask intent matching model;
step S5072: a is the number of the selected text vector in the intention category;
step S5073: p is the number of the positive sample text vector corresponding to the text vector a;
step S5074: n is a negative sample text vector number corresponding to the text vector a;
step S5075: i is the category to which the intent category belongs;
step S5077: d is the distance between vectors;
NCnumber of intention categories
NSNumber of samples contained for each intention category
Specifically, the optimization function of the first multitask intention matching model refers to a function that obtains the maximum intra-class distance and the minimum inter-class distance in the category from the first subtask, and is constructed by targeting shortening the maximum intra-class distance and the minimum inter-class distance, wherein,
optimizing a target difference value for the multitask intention matching model, wherein a, p and N are respectively text numbers selected in the intention categories, namely target parameters needing to be optimized by the optimization function are the difference between the distances between text vectors with the text vector numbers of a and p and the distances between the text vectors of a and N, i is the category to which the intention category belongs, D is the distance between vectors, N is the distance between vectorsCNumber of intention classes, NSNumber of samples contained for each intention category
That is, the optimization function of the first multitask intent matching model may decrease the distance between samples belonging to the same class and increase the distance between samples belonging to different classes.
Further, as shown in fig. 3, step S600 further includes:
step S601: obtaining a non-leaf node of the first flow tree1，node2，node3，node4；
Step S602: obtaining non-leaf node of the second flow tree1，node2，node5，node6；
Step S603: using the encoder of the first multitask intent matching model as the encoder of the second multitask intent matching model. To the node5，node6Corresponding task layer T5，T6And carrying out training updating to obtain a second multitask intention matching model.
Specifically, a non-leaf node of the first flow tree is obtained1，node2，node3，node4Obtaining the non-leaf node of the second flow tree1，node2，node5，node6Wherein non-leaf node nodes of the first flow tree and the second flow tree1And a node2Are identical and the first multi-tasking intent matching model has been trained through the training set by the first flow tree, i.e., the first multi-tasking intent matching model comprises a feature matrix and 4 and a task level T1，T2，T3，T4And the second multitask intention matching model is encoded by using an encoder of the first multitask intention matching model to obtain the second multitask intention matching model and the node5，node6Corresponding task layer T5，T6And for the task layer T5，T6Performing a training update to obtain a second multitask intent matching model, i.e., theThe task layer of the second multitask intention matching model comprises T1，T2，T3，T4，T5，T6The model is more comprehensive in coverage and relates to more detailed business content.
Further, as shown in fig. 4, step S700 further includes:
step S701: for each non-leaf node in the first flow tree and/or the second flow treei；
Step S702: mixing the unidentified corpus and the nodeiInputting a multi-task intention matching model to obtain a first prediction result;
step S703: and carrying out intelligent outbound according to the first prediction result.
Specifically, each non-leaf node in the first flow tree and/or the second flow tree is obtainediI.e. the obtained non-leaf nodeiCan belong to a common subtask in the first flow tree and the second flow tree or a subtask in the first flow tree or the second flow tree, and then according to the obtained non-leaf nodeiAnd inputting the unrecognized corpus into the multitask intention matching model for intention matching, calculating the multitask intention matching model to obtain the intention of the unrecognized corpus, namely obtaining the first prediction result, and calling out by the multitask intention matching model according to the obtained first prediction result to follow up the business process.
Further, as shown in fig. 5, step S701 further includes:
step S7011: the node is put intoiInputting all the texts into the second multitask intention matching model to obtain a first feature vector;
step S7012: inputting the unrecognized corpus into the second multitask intention recognition model to obtain a second feature vector;
step S7013: randomly selecting a plurality of corpora for each category under the non-leaf nodes as a prototype corpora, wherein the prototype corpora comprise texts text1，text2，…，textnAnd a corresponding intention label y1，y2，…，yn；
Step S7014: calculating the Euclidean distance between the first characteristic vector and the second characteristic vector, and labeling the label y corresponding to the characteristic vector with the shortest Euclidean distancejAnd the label is used as the label of the unidentified corpus, and the label is returned.
Specifically, the nodeiInputting all texts into the second multi-task intention matching model to obtain a first feature vector, wherein the first feature vector comprises the feature vectors of all the input texts under the task node, then inputting the unrecognized corpus into the second multi-task intention recognition model to obtain a second feature vector, and simultaneously selecting all the corpora contained by the non-leaf node as prototype corpora of the second multi-task intention recognition model, wherein the prototype corpora comprises text1，text2，…，textnAnd a corresponding intention label y1，y2，…，ynCalculating the Euclidean distance between the first feature vector and the second feature vector according to the obtained text contained in the prototype corpus and the intention label corresponding to the prototype corpus, then using the label with the shortest Euclidean distance as the label of the input unidentified corpus, returning the obtained label of the unidentified corpus to the second multitask intention matching model for continuous calculation, returning the label of the unidentified corpus to the second multitask intention matching model for continuous calculation to obtain the intention label of the unidentified corpus, and performing service processing according to the intention label by the second multitask intention matching model.
Compared with the prior art, the invention has the following beneficial effects:
clustering analysis is carried out on all the training corpuses by using a clustering algorithm, the training corpuses which accord with a preset rule in each category are obtained, and the obtained training corpuses are used as a training set; obtaining a first flow tree according to a first outbound scene; obtaining a second flow tree according to a second call-out scene, wherein each non-leaf node of the first flow tree and each non-leaf node of the second flow tree are respectively used as a subtask of multi-task training; initializing a multitask intention matching model, wherein the multitask intention matching model comprises a plurality of task layers which respectively correspond to each subtask in the flow tree; training the initialized multitask intention matching model by using the corpus of the first flow tree in the training set to obtain a first multitask intention matching model; training and updating a task layer which is different from the subtasks of the first flow tree in the initialized first multitask intention matching model by using the corpus of the second flow tree in the training set to obtain a second multitask intention matching model; and carrying out intelligent outbound according to the second multitask intention matching model, so that the language materials and the labels required by free configuration of different nodes of the flow tree are realized through the multitask intention classification model, the intention labels under different nodes do not influence each other, the prototype language materials contained under each node can be directly updated to complete the updating and configuration of the model, and the technical effect of the reusability of the model is improved.
Example two
Based on the same inventive concept as the data modeling method based on the intelligent marketing scenario in the foregoing embodiment, the present invention further provides a node reusable intelligent outbound system based on a multitask model, as shown in fig. 6, where the system includes:
a first obtaining unit 11, where the first obtaining unit 11 is configured to obtain the corpus conforming to a predetermined rule in each category, and use the corpus as a training set;
a second obtaining unit 12, where the second obtaining unit 12 is configured to obtain a first flow tree according to a first outbound scenario;
a third obtaining unit 13, where the third obtaining unit 13 is configured to obtain a second flow tree according to a second outbound scenario;
a first initialization unit 14, the first initialization unit 14 being configured to initialize a multitask intent matching model, the multitask intentThe graph matching model comprises NnTask layer TnRespectively corresponding to each subtask in the flow tree;
a fourth obtaining unit 15, where the fourth obtaining unit 15 is configured to train the initialized multitask intention matching model by using the corpus of the first flow tree in the training set, so as to obtain a first multitask intention matching model;
a fifth obtaining unit 16, where the fifth obtaining unit 16 is configured to train and update a task layer, which is different from the subtasks of the first flow tree, in the initialized first multitask intention matching model by using the corpus of the second flow tree in the training set, and obtain a second multitask intention matching model;
a first outbound unit 17, wherein the first outbound unit 17 is configured to perform an intelligent outbound according to the second multitask intention matching model.
Further, the system further comprises:
a sixth obtaining unit, configured to obtain a first subtask task in the first flow treei；
A first selection unit for selecting from the first subtask taskiRandom selection of NCIntention categories, N being selected from each intention categorySTaking the bar text as a first training set;
a seventh obtaining unit, configured to input all training texts in the first training set into corresponding task layers T in the multitask intention matching modeliObtaining a first eigenvector matrix;
a first calculation unit, configured to calculate a euclidean distance between each pair of eigenvectors in the first eigenvector matrix;
a second calculation unit for calculating, for each of the classes, a maximum intra-class distance and a minimum inter-class distance;
an eighth obtaining unit, configured to obtain a difference between the maximum intra-class distance and the minimum inter-class distance for each of the classes;
a first optimization unit for optimizing the multitask intention matching model by minimizing a difference between the maximum intra-class distance and the minimum inter-class distance to obtain a first multitask intention matching model.
Further, the system further comprises:
a first function unit, an optimization function of the first multitask intent matching model of the first function unit is:
wherein the content of the first and second substances,
optimizing a target difference value for the multitask intent matching model;
a first numbering unit, wherein the first numbering unit is used for numbering a text vector selected from the intention category;
a second numbering unit, wherein the second numbering unit is used for numbering the text vectors selected in the intention category;
a third numbering unit, wherein the third numbering unit is used for numbering n tasks;
a first category unit for c being a category to which the intention category belongs;
a first intent unit for y being an intent tag;
a first distance unit for D being an inter-vector distance;
a second class unit for classiIs the intention category to which the feature vector belongs.
Further, the system further comprises:
a ninth obtaining unit for obtaining a non-leaf node of the first flow tree1，node2，node3，node4；
A tenth obtaining unit for obtaining a non-leaf node of the second flow tree1，node2，node5，node6；
An eleventh obtaining unit for using an encoder of the first multitask intention matching model as an encoder of the second multitask intention matching model. To the node5，node6Corresponding task layer T5，T6And carrying out training updating to obtain a second multitask intention matching model.
Further, the system further comprises:
a second selection unit for, for each non-leaf node in the first flow tree and/or the second flow treei；
A twelfth obtaining unit configured to obtain the unidentified corpus and the nodeiInputting a multi-task intention matching model to obtain a first prediction result;
the first prediction unit is used for carrying out intelligent outbound according to the first prediction result.
Further, the system further comprises:
a first input unit for inputting the nodeiInputting all the texts into the second multitask intention matching model to obtain a first feature vector;
the second input unit is used for inputting the unidentified corpus into the second multitask intention recognition model to obtain a second feature vector;
and the third selection unit is used for randomly selecting a plurality of corpora for each category under the non-leaf node as a prototype corpora. The prototype corpus contains text and text1，text2，…，textnAnd a corresponding intention label y1，y2，…，yn；
A third calculation unit configured to calculate a euclidean distance between the first feature vector and the second feature vector, and calculate a feature having the shortest euclidean distanceLabel y corresponding to vectorjAnd the label is used as the label of the unidentified corpus, and the label is returned.
Various changes and specific examples of the multitask model-based node reusable intelligent outbound method in the first embodiment of fig. 1 are also applicable to the multitask model-based node reusable intelligent outbound system in the present embodiment, and through the foregoing detailed description of the multitask model-based node reusable intelligent outbound method, those skilled in the art can clearly know that the multitask model-based node reusable intelligent outbound system in the present embodiment is not described in detail herein for the sake of brevity of the description.
EXAMPLE III
The electronic apparatus of the embodiment of the present application is described below with reference to fig. 7.
Fig. 7 illustrates a schematic structural diagram of an electronic device according to an embodiment of the present application.
Based on the inventive concept of the multitasking model-based node reusable intelligent outbound method in the foregoing embodiment, the present invention further provides a multitasking model-based node reusable intelligent outbound system, and an electronic device according to an embodiment of the present application is described below with reference to fig. 7. The electronic device may be a removable device itself or a stand-alone device independent thereof, on which a computer program is stored which, when being executed by a processor, carries out the steps of any of the methods as described hereinbefore.
As shown in fig. 7, the electronic device 50 includes one or more processors 51 and a memory 52.
The processor 51 may be a Central Processing Unit (CPU) or other form of processing unit having data processing capabilities and/or instruction execution capabilities, and may control other components in the electronic device 50 to perform desired functions.
The memory 52 may include one or more computer program products that may include various forms of computer-readable storage media, such as volatile memory and/or non-volatile memory. The volatile memory may include, for example, Random Access Memory (RAM), cache memory (cache), and/or the like. The non-volatile memory may include, for example, Read Only Memory (ROM), hard disk, flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium and executed by the processor 51 to implement the methods of the various embodiments of the application described above and/or other desired functions.
In one example, the electronic device 50 may further include: an input device 53 and an output device 54, which are interconnected by a bus system and/or other form of connection mechanism (not shown).
The embodiment of the invention provides a node reusable intelligent outbound method based on a multitask model, wherein the method comprises the following steps: clustering analysis is carried out on all the training corpuses by using a clustering algorithm, the training corpuses which accord with a preset rule in each category are obtained, and the obtained training corpuses are used as a training set; obtaining a first flow tree according to a first outbound scene; obtaining a second flow tree according to a second call-out scene, wherein each non-leaf node of the first flow tree and each non-leaf node of the second flow tree are respectively used as a subtask of multi-task training; initializing a multitask intention matching model, wherein the multitask intention matching model comprises a plurality of task layers which respectively correspond to each subtask in the flow tree; training the initialized multitask intention matching model by using the corpus of the first flow tree in the training set to obtain a first multitask intention matching model; training and updating a task layer which is different from the subtasks of the first flow tree in the initialized first multitask intention matching model by using the corpus of the second flow tree in the training set to obtain a second multitask intention matching model; and carrying out intelligent outbound according to the second multitask intention matching model, so that the language materials and the labels required by free configuration of different nodes of the flow tree are realized through the multitask intention classification model, the intention labels under different nodes do not influence each other, the prototype language materials contained under each node can be directly updated to complete the updating and configuration of the model, and the technical effect of the reusability of the model is improved.
Through the above description of the embodiments, those skilled in the art will clearly understand that the present application can be implemented by software plus necessary general-purpose hardware, and certainly can also be implemented by special-purpose hardware including special-purpose integrated circuits, special-purpose CPUs, special-purpose memories, special-purpose components and the like. Generally, functions performed by computer programs can be easily implemented by corresponding hardware, and specific hardware structures for implementing the same functions may be various, such as analog circuits, digital circuits, or dedicated circuits. However, for the present application, the implementation of a software program is more preferable. Based on such understanding, the technical solutions of the present application may be substantially embodied in the form of a software product, which is stored in a readable storage medium, such as a floppy disk, a usb disk, a removable hard disk, a ROM, a RAM, a magnetic disk, or an optical disk of a computer, and includes several instructions for causing a computer device to execute the method according to the embodiments of the present application.
In the above embodiments, the implementation may be wholly or partially realized by software, hardware, firmware, or any combination thereof. When implemented in software, may be implemented in whole or in part in the form of a computer program product.
The computer program product includes one or more computer instructions. When loaded and executed on a computer, cause the processes or functions described in accordance with the embodiments of the application to occur, in whole or in part. The computer may be a general purpose computer, a special purpose computer, a network of computers, or other programmable device. The computer instructions may be stored on or transmitted from a computer-readable storage medium to another computer-readable storage medium, which may be magnetic (e.g., floppy disks, hard disks, tapes), optical (e.g., DVDs), or semiconductor (e.g., Solid State Disks (SSDs)), among others.
It should be appreciated that reference throughout this specification to "one embodiment" or "an embodiment" means that a particular feature, structure or characteristic described in connection with the embodiment is included in at least one embodiment of the present application. Thus, the appearances of the phrases "in one embodiment" or "in an embodiment" in various places throughout this specification are not necessarily all referring to the same embodiment. Furthermore, the particular features, structures, or characteristics may be combined in any suitable manner in one or more embodiments. It should be understood that, in the various embodiments of the present application, the sequence numbers of the above-mentioned processes do not mean the execution sequence, and the execution sequence of each process should be determined by its function and inherent logic, and should not constitute any limitation to the implementation process of the embodiments of the present application.
Additionally, the terms "system" and "network" are often used interchangeably herein. The term "and/or" herein is merely an association describing an associated object, meaning that three relationships may exist, e.g., a and/or B, may mean: a exists alone, A and B exist simultaneously, and B exists alone. In addition, the character "/" herein generally indicates that the former and latter related objects are in an "or" relationship.
It should be understood that in the embodiment of the present application, "B corresponding to a" means that B is associated with a, from which B can be determined. It should also be understood that determining B from a does not mean determining B from a alone, but may be determined from a and/or other information.
Those of ordinary skill in the art will appreciate that the elements and algorithm steps of the examples described in connection with the embodiments disclosed herein may be embodied in electronic hardware, computer software, or combinations of both, and that the components and steps of the examples have been described in a functional general in the foregoing description for the purpose of illustrating clearly the interchangeability of hardware and software. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the implementation. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.
In short, the above description is only a preferred embodiment of the present disclosure, and is not intended to limit the scope of the present disclosure. Any modification, equivalent replacement, improvement and the like made within the spirit and principle of the present application shall be included in the protection scope of the present application.
Claims (8)
1. A node reusable intelligent outbound method based on a multitask model, wherein the method comprises the following steps:
clustering analysis is carried out on all the training corpuses by using a clustering algorithm, the training corpuses which accord with a preset rule in each category are obtained, and the obtained training corpuses are used as a training set;
obtaining a first flow tree according to a first outbound scene;
obtaining a second flow tree according to a second call-out scene, wherein each non-leaf node of the first flow tree and each non-leaf node of the second flow tree are respectively used as a subtask of multi-task training;
initializing a multitask intent matching model comprising NnTask layer TnRespectively corresponding to each subtask in the flow tree;
training the initialized multitask intention matching model by using the corpus of the first flow tree in the training set to obtain a first multitask intention matching model;
training and updating a task layer which is different from the subtasks of the first flow tree in the initialized first multitask intention matching model by using the corpus of the second flow tree in the training set to obtain a second multitask intention matching model;
and carrying out intelligent outbound according to the second multitask intention matching model.
2. The method of claim 1, wherein the training the initialized multitask intention matching model by using the corpus of the first flow tree in the training set to obtain a first multitask intention matching model comprises:
obtaining in the first flow treeFirst subtask taski；
From the first subtask taskiRandom selection of NCIntention categories, N being selected from each intention categorySTaking the bar text as a first training set;
inputting all training texts in the first training set into corresponding task layers T in the multi-task intention matching modeliObtaining a first eigenvector matrix;
calculating Euclidean distance between each pair of eigenvectors in the first eigenvector matrix;
for each of said classes, calculating a maximum intra-class distance and a minimum inter-class distance;
obtaining a difference between the maximum intra-class distance and the minimum inter-class distance for each of the classes;
and optimizing the multitask intention matching model by minimizing the difference between the maximum intra-class distance and the minimum inter-class distance to obtain a first multitask intention matching model.
3. The method of claim 2, wherein the optimizing the multitask intent matching model by minimizing the difference between the maximum intra-class distance and the minimum inter-class distance to obtain a first multitask intent matching model comprises:
the optimization function of the first multitask intention matching model is as follows:
wherein the content of the first and second substances,
optimizing a target difference value for the multitask intent matching model;
a is the number of the selected text vector in the intention category;
p is the number of the positive sample text vector corresponding to the text vector a;
n is a negative sample text vector number corresponding to the text vector a;
i is the category to which the intent category belongs;
d is the distance between vectors;
NCthe number of intention categories;
NSthe number of samples contained for each intent category.
4. The method of claim 1, wherein the obtaining a second multitask intent matching model by training and updating a task layer of the initialized first multitask intent matching model different from the subtasks of the first flow tree by using the corpus of the second flow tree in the training set comprises:
obtaining a non-leaf node of the first flow tree1，node2，node3，node4；
Obtaining non-leaf node of the second flow tree1，node2，node5，node6；
Using the encoder of the first multitask intention matching model as the encoder of the second multitask intention matching model to match the node5，node6Corresponding task layer T5，T6And carrying out training updating to obtain a second multitask intention matching model.
5. The method of claim 1, wherein the method further comprises:
for each non-leaf node in the first flow tree and/or the second flow treei，
Mixing the unidentified corpus and the nodeiInputting a multi-task intention matching model to obtain a first prediction result;
and carrying out intelligent outbound according to the first prediction result.
6. The method of claim 5, wherein the method further comprises:
the node is put intoiInputting all the texts into the second multitask intention matching model to obtain a first feature vector;
inputting the unrecognized corpus into the second multitask intention recognition model to obtain a second feature vector;
randomly selecting a plurality of corpora for each category under the non-leaf node as a prototype corpora, wherein the prototype corpora comprise texts and texts1，text2，…，textnAnd a corresponding intention label y1，y2，…，yn；
Calculating the Euclidean distance between the first characteristic vector and the second characteristic vector, and labeling the label y corresponding to the characteristic vector with the shortest Euclidean distancejThe label is used as the label of the unidentified corpus, and the label is returned;
if the number of the non-leaf nodes of the flow tree is NnThen the number of the subtasks is NnEach subtask is assigned training data to be used, and the training data comprises text and text1，text2，…，textnAnd a corresponding intention label y1，y2，…，yn。
7. A multitask model based node multiplexing intelligent outbound system, wherein the system comprises:
the first obtaining unit is used for obtaining the training corpora which accord with the preset rules in each category and taking the training corpora as a training set;
a second obtaining unit, configured to obtain a first flow tree according to a first outbound scenario;
a third obtaining unit, configured to obtain a second flow tree according to a second outbound scenario;
a first initialization unit to initialize a multitask intent matching model comprising NnTask layer TnRespectively corresponding to each subtask in the flow tree;
a fourth obtaining unit, configured to train the initialized multitask intention matching model by using the corpus of the first flow tree in the training set, so as to obtain a first multitask intention matching model;
a fifth obtaining unit, configured to perform training and updating on a task layer different from the subtasks of the first flow tree in the initialized first multitask intent matching model by using the corpus of the second flow tree in the training set, so as to obtain a second multitask intent matching model;
and the first outbound unit is used for carrying out intelligent outbound according to the second multitask intention matching model.
8. A multitasking model based node multiplexing intelligent outbound system comprising a memory, a processor and a computer program stored on the memory and being executable on the processor, wherein said processor when executing said program implements the steps of the method according to any one of claims 1-6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111276483.4A CN113792819B (en) | 2021-10-29 | 2021-10-29 | Multi-task model-based node reusable intelligent outbound method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111276483.4A CN113792819B (en) | 2021-10-29 | 2021-10-29 | Multi-task model-based node reusable intelligent outbound method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113792819A true CN113792819A (en) | 2021-12-14 |
| CN113792819B CN113792819B (en) | 2024-05-14 |
Family
ID=78878376
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111276483.4A Active CN113792819B (en) | 2021-10-29 | 2021-10-29 | Multi-task model-based node reusable intelligent outbound method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113792819B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180124241A1 (en) * | 2016-10-28 | 2018-05-03 | Microsoft Technology Licensing, Llc | Caller assistance system |
| CN110233946A (en) * | 2019-06-17 | 2019-09-13 | 三角兽（北京）科技有限公司 | Execute outbound call service method, electronic equipment and computer readable storage medium |
| US20190286906A1 (en) * | 2018-03-13 | 2019-09-19 | International Business Machines Corporation | Automated intent to action mapping in augmented reality environments |
| CN110276074A (en) * | 2019-06-20 | 2019-09-24 | 出门问问信息科技有限公司 | Distributed training method, device, equipment and the storage medium of natural language processing |
| CN111696576A (en) * | 2020-05-21 | 2020-09-22 | 升智信息科技(南京)有限公司 | Intelligent voice robot talk test system |
| CN112185358A (en) * | 2020-08-24 | 2021-01-05 | 维知科技张家口有限责任公司 | Intention recognition method, model training method, device, equipment and medium |
| CN113094481A (en) * | 2021-03-03 | 2021-07-09 | 北京智齿博创科技有限公司 | Intention recognition method and device, electronic equipment and computer readable storage medium |
-
2021
- 2021-10-29 CN CN202111276483.4A patent/CN113792819B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180124241A1 (en) * | 2016-10-28 | 2018-05-03 | Microsoft Technology Licensing, Llc | Caller assistance system |
| US20190286906A1 (en) * | 2018-03-13 | 2019-09-19 | International Business Machines Corporation | Automated intent to action mapping in augmented reality environments |
| CN110233946A (en) * | 2019-06-17 | 2019-09-13 | 三角兽（北京）科技有限公司 | Execute outbound call service method, electronic equipment and computer readable storage medium |
| CN110276074A (en) * | 2019-06-20 | 2019-09-24 | 出门问问信息科技有限公司 | Distributed training method, device, equipment and the storage medium of natural language processing |
| CN111696576A (en) * | 2020-05-21 | 2020-09-22 | 升智信息科技(南京)有限公司 | Intelligent voice robot talk test system |
| CN112185358A (en) * | 2020-08-24 | 2021-01-05 | 维知科技张家口有限责任公司 | Intention recognition method, model training method, device, equipment and medium |
| CN113094481A (en) * | 2021-03-03 | 2021-07-09 | 北京智齿博创科技有限公司 | Intention recognition method and device, electronic equipment and computer readable storage medium |
Non-Patent Citations (2)
| Title |
|---|
| LISONG CHEN等: "Joint Multiple Intent Detection and Slot Filling Via Self-Distillation", ARXIV:2108.08042V1, 18 August 2021 (2021-08-18) * |
| 张庆恒: "智能机器外呼系统的设计与实现", 中国优秀硕士学位论文全文数据库 信息科技辑, no. 08, 15 August 2019 (2019-08-15) * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113792819B (en) | 2024-05-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11551004B2 (en) | Intent discovery with a prototype classifier | |
| US10747957B2 (en) | Processing communications using a prototype classifier | |
| TWI752455B (en) | Image classification model training method, image processing method, data classification model training method, data processing method, computer device, and storage medium | |
| Paramesh et al. | Automated IT service desk systems using machine learning techniques | |
| CN110781294A (en) | Training corpus refinement and incremental update | |
| US20220100963A1 (en) | Event extraction from documents with co-reference | |
| CN109214407B (en) | Event detection model, method and device, computing equipment and storage medium | |
| CN109299263B (en) | Text classification method and electronic equipment | |
| US20200210776A1 (en) | Question answering method, terminal, and non-transitory computer readable storage medium | |
| Arnold | An active-set evolution strategy for optimization with known constraints | |
| CN110569870A (en) | deep acoustic scene classification method and system based on multi-granularity label fusion | |
| CN112732882A (en) | User intention identification method, device, equipment and computer readable storage medium | |
| US20060179017A1 (en) | Preparing data for machine learning | |
| CN115329075A (en) | Text classification method based on distributed machine learning | |
| CN115600605A (en) | Method, system, equipment and storage medium for jointly extracting Chinese entity relationship | |
| CN110765276A (en) | Entity alignment method and device in knowledge graph | |
| US12001174B2 (en) | Determination of task automation using an artificial intelligence model | |
| CN113792819A (en) | Multitask model-based node reusable intelligent outbound method and system | |
| WO2020095655A1 (en) | Selection device and selection method | |
| US20220292393A1 (en) | Utilizing machine learning models to generate initiative plans | |
| CN115146653A (en) | Dialogue script construction method, device, equipment and storage medium | |
| CN114372148A (en) | Data processing method based on knowledge graph technology and terminal equipment | |
| CN112329440A (en) | Relation extraction method and device based on two-stage screening and classification | |
| CN112445933A (en) | Model training method, device, equipment and storage medium | |
| CN111767735A (en) | Method, apparatus and computer readable storage medium for executing task |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB02 | Change of applicant information | ||
| CB02 | Change of applicant information |
Address after: 100000 floors 1-3, block a, global creative Plaza, No. 10, Furong street, Chaoyang District, Beijing Applicant after: Bairong Zhixin (Beijing) Technology Co.,Ltd. Address before: 100000 floors 1-3, block a, global creative Plaza, No. 10, Furong street, Chaoyang District, Beijing Applicant before: Bairong Zhixin (Beijing) credit investigation Co.,Ltd. |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |