A method, apparatus, device and storage medium for refusing to recognize

Abstract

The application provides a method, device and equipment for refusing to recognize, and a storage medium. The method comprises: obtaining target information to be recognized; inputting the target information into a pre-trained function recognition model to obtain a target business function corresponding to the target information; inputting the target information into a pre-trained intention classification model to obtain a target intention type corresponding to the target information; and determining whether to refuse to recognize the target information according to the target business function and the target intention type. The method for refusing to recognize provided by the application distinguishes whether the target information is a function range utterance or a non-function range utterance of the target business type by judging whether the target intention type corresponding to the target information conforms to the intention type supported by the target business function, and refuses to recognize the non-function range utterance, thereby improving the accuracy of the voice interaction system in refusing to recognize information, optimizing the refusing to recognize task in the voice interaction system, and improving the refusing to recognize effect.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of artificial intelligence, and in particular relates to a rejection method and device, equipment and a storage medium. BACKGROUND

[0002] With the progress of artificial intelligence technology, human-computer voice interaction has also made great progress. In order to improve the user experience, full-duplex voice interaction mode has begun to appear. The characteristic of full-duplex voice interaction mode is that after being woken up once, multiple continuous interactions can be carried out within a certain period of time.

[0003] In the full-duplex voice interaction mode, the intelligent voice system needs to continuously record the user's input voice, and carry out natural language understanding and reply to the recorded voice. In this mode, some of the user's input voice that is not interacting with the system (such as chatting with others, personal self-talk, etc.) needs to be accurately recognized, and the user's voice needs to be rejected when it is determined that the user is not interacting with the system, so as to avoid disturbing the user.

[0004] In order to achieve the above function, the current intelligent voice system generally establishes a classification model such as an out-of-domain (OOD) model to distinguish between functional range utterances and non-functional range utterances (i.e. out-of-function range utterances). Such classification models mainly distinguish based on utterance content, for example, they can distinguish whether a user's sentence is to listen to a song, or to inquire about the weather, or a sentence completely unrelated to the function involved. However, when it comes to similar information content and different intent utterance corpus (for example, "I want to go to Shanghai" and "I have been to Shanghai"), the distinction effect is often not very good, resulting in poor rejection effect and affecting user experience. SUMMARY

[0005] In view of the above problems of the prior art, the purpose of the present application is to provide a rejection method, device, equipment and storage medium, which can improve the accuracy of rejecting information, and thereby improve the interaction effect between the user and the voice interaction system.

[0006] In order to solve the above problems, the present application provides a rejection method, comprising:

[0007] obtaining target information to be recognized;

[0008] inputting the target information into a pre-trained function recognition model to obtain a target business function corresponding to the target information;

[0009] inputting the target information into a pre-trained intent classification model to obtain a target intent type corresponding to the target information;

[0010] determine whether to reject the target information according to the target business function and the target intent type.

[0011] Further, the obtaining the target information to be recognized comprises:

[0012] obtaining voice information to be recognized;

[0013] performing voice recognition processing on the voice information to obtain voice text corresponding to the voice information.

[0014] Further, the determining whether to reject the target information according to the target business function and the target intent type comprises:

[0015] obtaining a set of intent types corresponding to the target business function;

[0016] matching the target intent type with the set of intent types to obtain a matching result;

[0017] determining whether to reject the target information according to the matching result.

[0018] Further, the matching the target intent type with the set of intent types to obtain a matching result comprises:

[0019] matching the target intent type with each intent type in the set of intent types respectively;

[0020] when the target intent type matches any one intent type in the set of intent types, determining that the matching result is a matching success;

[0021] when the target intent type does not match each intent type in the set of intent types, determining that the matching result is a matching failure.

[0022] Further, the determining whether to reject the target information according to the matching result comprises:

[0023] when the matching result is a matching failure, rejecting to recognize the target information;

[0024] when the matching result is a matching success, obtaining semantic information corresponding to the target information and outputting.

[0025] Further, the method further comprises:

[0026] training the function recognition model and the intent classification model by using a pre-obtained sample.

[0027] Another aspect of the present application provides a rejection device, comprising:

[0028] an obtaining module, configured to obtain target information to be recognized;

[0029] a recognition module, configured to input the target information into a pre-trained function recognition model to obtain a target business function corresponding to the target information;

[0030] a classification module, configured to input the target information into a pre-trained intention classification model to obtain a target intention type corresponding to the target information;

[0031] a determination module, configured to determine whether to reject the target information according to the target business function and the target intention type.

[0032] Further, the device further comprises:

[0033] a training module, configured to train the function recognition model and the intention classification model by using pre-acquired samples.

[0034] Another aspect of the present application provides an electronic device, comprising a processor and a memory, wherein the memory stores at least one instruction or at least one program, and the at least one instruction or the at least one program is loaded and executed by the processor to implement the rejection method as described above.

[0035] Another aspect of the present application provides a computer readable storage medium, wherein the computer readable storage medium stores at least one instruction or at least one program, and the at least one instruction or the at least one program is loaded and executed by a processor to implement the rejection method as described above.

[0036] Due to the above technical solutions, the present application has the following beneficial effects:

[0037] According to the rejection method provided by the embodiment of the present application, the target business function corresponding to the target information to be recognized is determined by using the function recognition model, the target intention type corresponding to the target information to be recognized is determined by using the intention classification model, and whether the target intention type conforms to the intention type supported by the target business function is judged to distinguish whether the target information is a function range utterance or a non-function range utterance of the target business type, and finally the non-function range utterance is rejected, which has a good distinguishing effect on utterance corpus with similar information content and different intentions, improves the accuracy of the voice interaction system in rejecting information, optimizes the rejection task in the voice interaction system, improves the rejection effect, and thus can avoid unnecessary disturbance to the user, improves the interaction effect between the user and the voice interaction system, and improves the user experience. BRIEF DESCRIPTION OF DRAWINGS

[0038] In order to make the technical scheme of the present application clearer, the following will briefly introduce the drawings needed in the embodiments or prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained from these drawings without creative labor.

[0039] Figure 1 is a schematic diagram of an implementation environment provided by an embodiment of the present application;

[0040] Figure 2 is a flowchart of a rejection method provided by an embodiment of the present application;

[0041] Figure 3 is a flowchart of a rejection method provided by another embodiment of the present application;

[0042] Figure 4 is a schematic diagram of a rejection device provided by an embodiment of the present application;

[0043] Figure 5 is a schematic diagram of a rejection device provided by another embodiment of the present application;

[0044] Figure 6 is a schematic diagram of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0045] In order to make the technical scheme of the present application clearer, the following will briefly introduce the drawings needed in the embodiments or prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained from these drawings without creative labor.

[0046] It should be noted that the terms "first", "second", and the like in the description, claims, and above drawings of the present application are used to distinguish similar objects, and do not necessarily indicate a specific order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the present application described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, device, product, or apparatus that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but can include other steps or units that are not clearly listed or inherent to these processes, methods, products, or apparatuses.

[0047] For the purposes of the present disclosure, the technical solutions and advantages of the embodiments of the present application are more clearly and explicitly described below in conjunction with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the embodiments of the present application and are not intended to limit the embodiments of the present application. First, the embodiments of the present application explain the following concepts:

[0048] Full-duplex interaction: after one wake-up, the user can perform multiple consecutive interactions without waking up the voice system within a certain period of time.

[0049] Natural Language Understanding (NLU): NLU is an application field of artificial intelligence, which is a technology for communicating with computers using natural language. Through NLU, computers can "understand" natural language and then perform certain language functions expected by humans. In other words, NLU is a bridge for human-machine communication. NLU can include but is not limited to the following: ① can understand the correct order rules and concepts of a sentence, and can also understand sentences without rules; ② knows the exact meaning, form, part of speech, and word formation of a word; ③ understands semantic classification of words, polysemy of words, and ambiguity of words; ④ specifies and indefinite characteristics and all characteristics; ⑤ structural knowledge and practical concepts of the problem domain; ⑥ language tone information and prosodic performance; ⑦ textual knowledge about language expression form; ⑧ background knowledge of the domain.

[0050] Speech acts classification theory: a theory of pragmatics that focuses on the behavior carried by speech in addition to information and classifies the behavior at the intention level.

[0051] Reference is made to the accompanying drawings Figure 1 which shows an implementation environment schematic diagram provided by an embodiment of the present application, as Figure 1 shown, the implementation environment can include at least one terminal device 110 and a server 120. The server 120 and each terminal device 110 can be directly or indirectly connected through wired or wireless communication, and the embodiments of the present application do not limit this.

[0052] The terminal device 110 can include an entity device such as a smart phone, a tablet computer, a notebook computer, a desktop computer, a digital assistant, a smart speaker, a smart wearable device, a vehicle-mounted terminal, a server, etc., and can also include software such as an application program running in an entity device, but is not limited thereto. The operating system running on the terminal device 110 can include but is not limited to an Android system, an IOS system, a Linux system, a Windows system, etc.

[0053] The server 120 can be a stand-alone physical server, a server cluster composed of multiple physical servers, or a distributed system, and can also be a cloud server providing cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDN), and basic cloud computing services such as big data and artificial intelligence platforms.

[0054] In actual applications, the rejection method provided by the embodiments of the present application can be applied to a full-duplex interaction scenario. Specifically, the terminal device 110 can be configured with an intelligent voice system, which can collect voice information issued by a user through a voice collection module and send the collected voice information to the server 120 for processing. The server 120 can determine whether the voice information is voice for interacting with the intelligent voice system through the rejection method provided by the embodiments of the present application. For voice that is not for interacting with the intelligent voice system, the voice is rejected and no feedback is given. For voice for interacting with the intelligent voice system, corresponding semantic information is output, and corresponding interaction data is further generated and sent to the terminal device 110, so that the intelligent voice system can complete the interaction with the user.

[0055] It should be noted that, Figure 1 is only an example. Those skilled in the art can understand that although Figure 1 only two terminal devices 110 are shown, this does not limit the embodiments of the present application, and more or fewer terminal devices 110 can be included than shown.

[0056] Reference is made to the accompanying drawings Figure 2 which show a flow of a rejection method provided by an embodiment of the present application. The method can be applied to the server 120 in Figure 1 , and specifically as shown in Figure 2 , the method can include the following steps:

[0057] S210: Obtain target information to be identified.

[0058] In the embodiments of the present application, the target information is information that needs to be determined whether to be rejected or not. The target information can be various information input by a user in an interaction process between an intelligent voice system and the user.

[0059] Specifically, obtaining the target information to be identified can include:

[0060] Obtaining voice information to be identified;

[0061] The voice information is subjected to voice recognition processing to obtain voice text corresponding to the voice information.

[0062] Specifically, the terminal device can be configured with an intelligent voice system, which can collect voice information input by a user through a voice collection module and send the voice information to a server. The server can use voice recognition technology to recognize the voice information to obtain voice text corresponding to the voice information, and use the voice text as the target information to be recognized. The voice collection module can be a sound sensor, a microphone, etc.

[0063] S220: inputting the target information into a pre-trained function recognition model to obtain a target service function corresponding to the target information.

[0064] In the embodiment of the application, the function recognition model can process the input target information to determine an intent corresponding to the target information, and determine a target service function corresponding to the target information according to the intent corresponding to the target information. The target service function can be one or more of various functions provided by the intelligent voice system to a user. The various functions provided by the intelligent voice system to a user can include various functions commonly used in the prior art, such as a weather query function, a music playing function, and a navigation initiation function, etc. Details are not repeated here.

[0065] In one possible embodiment, the method can further include the step of training the function recognition model using pre-acquired samples.

[0066] Specifically, first training data labeled with corresponding service functions can be pre-acquired, and a first preset neural network model is trained using the first training data to obtain the function recognition model. The first preset neural network model can include but is not limited to a deep neural network model commonly used in the prior art. Details are not repeated here.

[0067] S230: inputting the target information into a pre-trained intent classification model to obtain a target intent type corresponding to the target information.

[0068] In the embodiment of the application, the intent classification model can process the input target information to determine a target intent type corresponding to the target information. The target intent type can be one or more of various intent types in a classification framework in a current application scenario. The classification framework in the current application scenario can be based on a speech intent classification theory to pre-determine user utterance information in the current application scenario.

[0069] Exemplarily, for the automobile scenario, the user utterance information in the in-vehicle intelligent voice system can include utterances of initiating instructions and other non-instruction utterances, according to which the classification framework in the automobile scenario can be determined to include initiating command type, willingness expression type, information query type, and the like, and according to which the classification framework in the automobile scenario can be further determined to include event statement type, emotion expression type, and the like.

[0070] It should be noted that the classification framework in different application scenarios can include the same intent type or different intent types, and the embodiments of the present application do not limit this.

[0071] In one possible embodiment, the method can further include training the intent classification model by using pre-acquired samples.

[0072] Specifically, second training data labeled with corresponding intent types can be pre-acquired, and a second preset neural network model is trained by using the second training data to obtain the intent classification model. The second preset neural network model can include, but is not limited to, a deep neural network model commonly used in the prior art, and the embodiments of the present application will not be repeated here.

[0073] In the embodiments of the present application, the intent classification model can have a good distinguishing effect on utterance information with similar information content and different intent types. For example, the same event "listening to music" can have different expression modes of intent types, as shown in Table 1, and distinguishing these intent types is the task of the intent classification model.

[0074] Table 1

[0075] Script information Intention type Whether corresponding to business function Listen to music Initiate command Corresponding to music playing function I want to listen to music Express desire Corresponding to music playing function I don't want to listen to music Express desire Corresponding to music playing function I love listening to music Express preference Not corresponding to business function I often listen to music Inform info Not corresponding to business function Do you listen to music Request to others Not corresponding to business function

[0076] Through the intent classification model, when a user expresses utterance of the same event but different intent types, the user can more accurately distinguish whether the user intent is to initiate an instruction or not (such as chatting with others), and this accurate judgment of whether the user initiates an instruction to the system can better realize the full-duplex and wake-up-free function of the intelligent voice system.

[0077] It should be noted that the steps S220 and S230 can be executed in sequence according to the above order, can be executed in sequence according to other orders, or can be executed in parallel, and the embodiments of the present application do not limit this.

[0078] S240: determining whether to reject the target information according to the target business function and the target intent type.

[0079] Specifically, as Figure 3As shown, the determining whether to reject the target information according to the target service function and the target intent type can include:

[0080] S241: Obtain an intent type set corresponding to the target service function.

[0081] In the embodiment of the application, the corresponding logic of various intent types in the classification framework in the current application scenario and various service functions supported by the intelligent voice system can be determined in advance. Specifically, for each service function supported by the intelligent voice system, there is a function range speech for the function, so the function range speech corresponding to each service function can be analyzed, and specific several intent types can be summarized to obtain the intent type set corresponding to the service function. For example, for the function of querying the weather, the information inquiry type "How is the weather today?" or the command type "Tell me the weather today" and the like are included, so it can be determined that the intent type set corresponding to the weather query function can include the information inquiry type and the command type.

[0082] S242: Match the target intent type with the intent type set to obtain a matching result.

[0083] In the embodiment of the application, the matching the target intent type with the intent type set to obtain a matching result can include:

[0084] Matching the target intent type with each intent type in the intent type set respectively;

[0085] When the target intent type matches any one of the intent types in the intent type set, it is determined that the matching result is a matching success;

[0086] When the target intent type does not match each intent type in the intent type set, it is determined that the matching result is a matching failure.

[0087] In the embodiment of the application, the target intent type can be matched with each intent type in the intent type set one by one, and it is determined whether the target intent type is the same as any one of the intent types in the intent type set, that is, whether the target intent type is included in the intent type set. If the target intent type is included in the intent type set, the matching is successful, otherwise the matching is failed.

[0088] S243: Determine whether to reject the target information according to the matching result.

[0089] In the embodiment of the application, the determining whether to reject the target information according to the matching result can include:

[0090] when the matching result is a matching failure, rejecting to recognize the target information;

[0091] when the matching result is a matching success, obtaining semantic information corresponding to the target information and outputting.

[0092] In the embodiment of the application, when the target intent type is contained in the intent type set, it can be determined that the target information is a functional range utterance for the target business type, that is, the target information needs to be recognized, at this time, the semantic information corresponding to the target information can be determined based on natural language understanding technology and output.

[0093] In the embodiment of the application, when the target intent type is not contained in the intent type set, it can be determined that the target information is a non-functional range utterance, that is, the target information does not need to be recognized, at this time, the target information can be rejected.

[0094] Exemplarily, assuming that the target information input by the user is "I want to go to the People's Square", it is determined by the function recognition model that the target business function corresponding to the target information is "initiating navigation function", it is determined by the intent classification model that the target intent type corresponding to the target information is "initiating command class", and assuming that the intent type set corresponding to the "initiating navigation function" includes "initiating command class", "will expression class" and "information question class", then the result of matching the target intent type with the intent type set is a matching success, that is, the target information needs to be recognized and an interactive behavior is initiated. Assuming that the target information input by the user is "I have been to the People's Square" or "I will go to the People's Square tomorrow", it can be determined by the intent classification model that the target intent type corresponding to the target information is "event statement class", then the result of matching the target intent type with the intent type set is a matching failure, that is, the target information does not need to be recognized and an interactive behavior is not initiated.

[0095] In a possible embodiment, the method can further include generating interaction information for the target information according to the semantic information, and sending the interaction information to a terminal device, so that a voice interaction system in the terminal device performs a corresponding interaction operation according to the interaction information. The interaction information can include reply information for the target information, or can include an interaction operation for a target execution object. When the interaction information is the reply information, the voice interaction system can display the reply information, for example, displaying the reply information in a specified interface, or playing the reply information through voice. When the interaction information is the interaction operation for the target execution object, the voice interaction system can generate an operation instruction for performing the interaction operation and send the operation instruction to the target execution object. The target execution object can perform the interaction operation in response to the operation instruction, and obtain an execution result.

[0096] In summary, according to the rejection method provided by the embodiment of the application, the target service function corresponding to the target information to be recognized is determined by using the function recognition model, the target intent type corresponding to the target information to be recognized is determined by using the intent classification model, and whether the target intent type conforms to the intent type supported by the target service function is determined, so as to distinguish whether the target information is a function range utterance or a non-function range utterance of the target service type, and finally reject the non-function range utterance. The utterance corpus with similar information content and different intents can be well distinguished, the accuracy of the voice interaction system in rejecting information is improved, the rejection task in the voice interaction system is optimized, the rejection effect is improved, and unnecessary disturbance to the user can be avoided, the interaction effect between the user and the voice interaction system is improved, and the user experience is improved.

[0097] Reference is made to the accompanying drawings Figure 4 which show the structure of a rejection device 400 provided by an embodiment of the application. As shown in the drawings, Figure 4 the device 400 can include:

[0098] The acquisition module 410 is configured to acquire target information to be recognized.

[0099] The recognition module 420 is configured to input the target information into a pre-trained function recognition model to obtain a target service function corresponding to the target information.

[0100] The classification module 430 is configured to input the target information into a pre-trained intent classification model to obtain a target intent type corresponding to the target information.

[0101] The determination module 440 is configured to determine whether to reject the target information according to the target service function and the target intent type.

[0102] In one possible embodiment, as shown in Figure 5 The apparatus 400 can further include:

[0103] a training module 450, configured to train the function recognition model and the intent classification model by using pre-acquired samples.

[0104] It should be noted that the apparatus provided in the above embodiments, when realizing its functions, is only exemplified by the above-mentioned division of functional modules, and in actual application, the above-mentioned functions can be completed by different functional modules according to needs, that is, the internal structure of the device is divided into different functional modules to complete all or part of the above-described functions. In addition, the apparatus provided in the above embodiments and the corresponding method embodiments belong to the same concept, and the specific implementation process is detailed in the corresponding method embodiments, which will not be described here.

[0105] One embodiment of the present application also provides an electronic device, which includes a processor and a memory, the memory storing at least one instruction or at least one program, the at least one instruction or the at least one program being loaded and executed by the processor to implement the rejection method provided in the above method embodiments.

[0106] The memory can be used to store software programs and modules, and the processor can execute various functional applications and data processing by running the software programs and modules stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, application programs required by functions, etc.; the data storage area can store data created according to the use of the device, etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as at least one magnetic disk storage device, a flash memory device, or other volatile solid-state memory device. Accordingly, the memory can also include a memory controller to provide access of the processor to the memory.

[0107] The method embodiments provided in the embodiments of the present application can be executed in a terminal, a server or a similar computing device, that is, the above-mentioned electronic device can include a terminal, a server or a similar computing device. Taking the case of running on a server as an example, as shown in Figure 6As shown, it shows a structural schematic diagram of a server for implementing the method of refusing provided by the embodiments of the present application. The server 600 can be quite different due to different configurations or performances, and can include one or more central processing units (CPU) 610 (for example, one or more processors) and a memory 630, one or more storage media 620 (for example, one or more mass storage devices) for storing application programs 623 or data 622. Among them, the memory 630 and the storage medium 620 can be temporary storage or persistent storage. The programs stored in the storage medium 620 can include one or more modules, each of which can include a series of instruction operations in the server. Further, the central processing unit 610 can be configured to communicate with the storage medium 620 to execute a series of instruction operations in the storage medium 620 on the server 600. The server 600 can also include one or more power supplies 660, one or more wired or wireless network interfaces 650, one or more input / output interfaces 640, and / or one or more operating systems 621, such as Windows ServerTM, Mac OS XTM, UnixTM, LinuxTM, FreeBSDTM, etc.

[0108] The input / output interface 640 can be used to receive or send data via a network. The above-mentioned specific examples of the network can include a wireless network provided by a communication provider of the server 600. In one example, the input / output interface 640 includes a network adapter (NIC) which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the input / output interface 640 can be a radio frequency (RF) module for communicating with the Internet in a wireless manner, and the wireless communication can use any communication standard or protocol, including but not limited to global system for mobile communication (GSM), general packet radio service (GPRS), code division multiple access (CDMA), wideband code division multiple access (WCDMA), long term evolution (LTE), email, short message service (SMS), etc.

[0109] Those skilled in the art can understand that Figure 6 The structure shown is only schematic, and the server 600 can also include more or fewer components than shown, or have a different configuration of components than shown. Figure 6 The server 600 can also have a different configuration of components than shown. Figure 6 The server 600 can also have a different configuration of components than shown.

[0110] One embodiment of the present application also provides a computer readable storage medium, which can be arranged in an electronic device to store at least one instruction or at least one program related to the method of rejecting. The at least one instruction or the at least one program is loaded and executed by the processor to implement the method of rejecting provided by the above-mentioned method embodiment.

[0111] Optionally, in the embodiment of the present application, the storage medium can include but is not limited to: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and various storage program codes.

[0112] One embodiment of the present application also provides a computer program product or computer program, which includes computer instructions stored in a computer readable storage medium. The processor of the computer device reads the computer instructions from the computer readable storage medium, and the processor executes the computer instructions to make the computer device execute the method of rejecting provided in the above-mentioned various optional implementation examples.

[0113] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description, not representing the advantages and disadvantages of the embodiments. The above-mentioned specific embodiments of the present application are described. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be executed in different order from the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are possible or can be advantageous.

[0114] Each embodiment in the present specification is described in a progressive manner, and the same or similar parts between each embodiment can be referred to each other. Each embodiment focuses on the difference from other embodiments. Especially, for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the related parts can be referred to the part of the method embodiment.

[0115] Those skilled in the art can understand that all or part of the steps of the above-mentioned embodiments can be completed by hardware, or can be instructed to relevant hardware by program. The program can be stored in a computer readable storage medium, and the storage medium mentioned above can be a read-only memory, a magnetic disk or an optical disk, etc.

[0116] The above description is merely preferred embodiments of the present application, and is not intended to limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method of refusing, characterized by, The method comprises the following steps: obtaining target information to be recognized; inputting the target information into a pre-trained function recognition model to obtain a target business function corresponding to the target information; wherein the target business function is one or more of various functions provided by an intelligent voice system to a user; inputting the target information into a pre-trained intent classification model to obtain a target intent type corresponding to the target information; wherein the target intent type is one or more of various intent types in a classification framework in a current application scenario, and the classification framework in the current application scenario is determined based on a speech intent classification theory; the various intent types include an initiation command type, a willingness expression type, an information query type, an event statement type, and an emotion expression type; determining whether to reject the target information according to the target business function and the target intent type, obtaining an intent type set corresponding to the target business function, matching the target intent type with the intent type set to obtain a matching result, and determining whether to reject the target information according to the matching result; wherein the intent type set corresponding to each business function is obtained by analyzing the function range of each business function.

2. The method of claim 1, wherein, The method further comprises the following steps: obtaining voice information to be recognized; performing voice recognition processing on the voice information to obtain voice text corresponding to the voice information.

3. The method of claim 1, wherein, The matching of the target intent type with the intent type set to obtain a matching result comprises the following steps: matching the target intent type with each intent type in the intent type set respectively; when the target intent type matches any one of the intent types in the intent type set, determining that the matching result is a matching success; when the target intent type does not match any of the intent types in the intent type set, determining that the matching result is a matching failure.

4. The method of claim 1, wherein, The determination of whether to reject the target information according to the matching result comprises the following steps: when the matching result is a matching failure, rejecting the recognition of the target information; when the matching result is a matching success, obtaining semantic information corresponding to the target information and outputting the semantic information.

5. The method of claim 1, wherein, The method further comprises the following steps: training the function recognition model and the intent classification model by using pre-obtained samples.

6. A device for refusing recognition, characterized in that The method comprises the following steps: an obtaining module, configured to obtain target information to be recognized; an identifying module, configured to input the target information into a pre-trained function recognition model to obtain a target business function corresponding to the target information; wherein the target business function is one or more of various functions provided by an intelligent voice system to a user; a classifying module, configured to input the target information into a pre-trained intent classification model to obtain a target intent type corresponding to the target information; wherein the target intent type is one or more of various intent types in a classification framework in a current application scenario, and the classification framework in the current application scenario is determined based on a speech intent classification theory; the various intent types include an initiation command type, a willingness expression type, an information query type, an event statement type, and an emotion expression type. The determining module is configured to determine whether to reject the target information according to the target service function and the target intent type, and obtain an intent type set corresponding to the target service function, wherein a function range language of each service function is analyzed to obtain the intent type set corresponding to each service function; the target intent type is matched with the intent type set to obtain a matching result; and whether to reject the target information is determined according to the matching result.

7. The apparatus of claim 6, wherein, The apparatus further includes: The training module is configured to train the function recognition model and the intent classification model by using a pre-acquired sample.

8. An electronic device, comprising: The electronic device includes a processor and a memory, and the memory stores at least one instruction or at least one program, which is loaded and executed by the processor to implement the rejection method according to any one of claims 1-5.

9. A computer-readable storage medium, characterized in that, The computer readable storage medium stores at least one instruction or at least one program, which is loaded and executed by the processor to implement the rejection method according to any one of claims 1-5.

Images