Information identification method, system, computing device and storage medium

Abstract

The embodiment of the present application provides an information recognition method, system, computing device and storage medium. In the embodiment of the present application, a communication voice request is received. When the source of the communication voice request does not belong to the preset source, the voice text corresponding to the communication voice is obtained; based on the voice text, the voice dialogue text corresponding to the voice text is obtained; based on the voice dialogue text, the category to which the voice dialogue text belongs is identified. Thus, the type of unknown communication voice can be directly and automatically identified through the unknown communication voice. Further, according to the type, the user can be helped to classify the unknown voice information, especially for bad information, to help the user avoid the harassment of bad information. At the same time, the user is not aware of the recognition process, which improves the user experience, and can also help the user provide the required category of unknown voice information.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to an information identification method, system, computing device and storage medium. Background Art

[0002] With the development of information technology, communication networks can realize the interaction between information, thus facilitating people's lives. For example, users can use communication networks for voice communication, online shopping, online entertainment, online financial management, and online payment, etc. Correspondingly, the carriers of information have also become an indispensable part of users' lives, such as mobile phones and computers. Based on this, the carriers of information will receive a large amount of unknown information, especially voice information, which will bring many bad experiences to users, and even worse, economic losses to users. Summary of the invention

[0003] The present application provides an information identification method, system, computing device and storage medium for relatively automatically identifying the category of unknown information and improving user experience.

[0004] An embodiment of the present application provides a method for identifying information, including: receiving a communication voice request, and when the source of the communication voice request does not belong to a preset source, obtaining a voice text corresponding to the communication voice; based on the voice text, obtaining a voice dialogue text corresponding to the voice text; and based on the voice dialogue text, identifying a category to which the voice dialogue text belongs.

[0005] The embodiment of the present application provides a method for establishing a voiceprint library, comprising: establishing an abnormal voiceprint library, wherein the abnormal voiceprint library is obtained based on the abnormal voiceprints.

[0006] The embodiment of the present application provides a method for identifying information, including: accepting a forwarded communication voice request and obtaining a voice text corresponding to the voice information; obtaining a voice dialogue text corresponding to the voice text based on the voice text; and identifying a category to which the voice dialogue text belongs based on the voice dialogue text.

[0007] An embodiment of the present application provides a method for identifying information, including: when a communication voice request is received, and the source of the communication voice request does not belong to a preset source, receiving the communication voice through a smart speaker, and shielding the reception of user voice and shielding the provision of communication voice to the user; obtaining a voice text corresponding to the communication voice through the smart speaker; obtaining a voice dialogue text corresponding to the voice text based on the voice text through the smart speaker; and identifying the category to which the voice dialogue text belongs based on the voice dialogue text through the smart speaker.

[0008] An embodiment of the present application also provides a method for setting information, including: obtaining a user communication voice identifier; binding the user communication voice identifier with a transfer communication voice identifier so that under preset conditions, the communication voice request is forwarded to the device to which the transfer communication voice identifier belongs according to the binding relationship; and setting the preset conditions.

[0009] An embodiment of the present application also provides a method for setting information, including: providing a binding interface for a communication voice identifier, wherein the binding interface displays a user communication voice identifier input area, a transfer communication voice identifier input area, and a preset condition input area, and the user communication voice identifier and the transfer communication voice identifier have a binding relationship; in response to a user's input operation, determining the user communication voice identifier and the transfer communication voice identifier in the corresponding input area; in response to the user's input operation, determining the preset condition in the corresponding input area, and under the preset condition, forwarding a communication voice request to the device to which the transfer communication voice identifier belongs; in response to the binding operation, completing the operation of the binding relationship.

[0010] An embodiment of the present application also provides an information identification system, including: a first device and a second device; the first device receives a communication voice request, and forwards the communication voice request to the second device when the source of the communication voice request does not belong to a preset source; the second device receives the communication voice request and obtains a voice text corresponding to the communication voice; based on the voice text, obtains a voice dialogue text corresponding to the voice text; based on the voice dialogue text, identifies the category to which the voice dialogue text belongs.

[0011] An embodiment of the present application also provides a computing device, including a memory and a processor; the memory is used to store a computer program; the processor is used to execute the computer program to: establish an abnormal voiceprint library, wherein the abnormal voiceprint library is obtained based on the abnormal voiceprint.

[0012] An embodiment of the present application also provides a computing device, including a memory, a processor, and a communication component; the memory is used to store a computer program; the communication component is used to receive a communication voice request; the processor is used to execute the computer program to: obtain a voice text corresponding to the communication voice when the source of the communication voice request does not belong to a preset source; obtain a voice dialogue text corresponding to the voice text based on the voice text; and identify the category to which the voice dialogue text belongs based on the voice dialogue text.

[0013] An embodiment of the present application also provides a computing device, including a memory and a processor; the memory is used to store a computer program; the processor is used to execute the computer program to: obtain a user communication voice identifier; bind the user communication voice identifier with a transferred communication voice identifier, so that under preset conditions, according to the binding relationship, the communication voice request is forwarded to the device to which the transferred communication voice identifier belongs; and set the preset conditions.

[0014] An embodiment of the present application also provides a computing device, including a memory and a processor; the memory is used to store a computer program; the processor is used to execute the computer program, so as to: provide a binding interface for a communication voice identifier, wherein the binding interface displays a user communication voice identifier input area, a transfer communication voice identifier input area and a preset condition input area, and the user communication voice identifier and the transfer communication voice identifier have a binding relationship; in response to a user's input operation, determine the user communication voice identifier and the transfer communication voice identifier in the corresponding input area; in response to a user's input operation, determine the preset condition in the corresponding input area, and under the preset condition, forward a communication voice request to the device to which the transfer communication voice identifier belongs; in response to the binding operation, complete the operation of the binding relationship.

[0015] An embodiment of the present application also provides a computing device, including a memory and a processor; the memory is used to store a computer program; the processor is used to execute the computer program, so as to: accept a forwarded communication voice request and obtain a voice text corresponding to the voice information; obtain a voice conversation text corresponding to the voice text based on the voice text; and identify the category to which the voice conversation text belongs based on the voice conversation text.

[0016] An embodiment of the present application also provides a computing device, including a memory and a processor; the memory is used to store a computer program; the processor is used to execute the computer program, so as to: when a communication voice request is received, and the source of the communication voice request does not belong to a preset source, receive the communication voice through a smart speaker, and shield the reception of user voice and shield the provision of communication voice to the user; obtain the voice text corresponding to the communication voice through the smart speaker; obtain the voice dialogue text corresponding to the voice text through the smart speaker based on the voice text; identify the category to which the voice dialogue text belongs based on the voice dialogue text through the smart speaker.

[0017] An embodiment of the present application also provides a computer-readable storage medium storing a computer program. When the computer program is executed by one or more processors, the one or more processors implement the steps in the above method.

[0018] In the embodiment of the present application, a communication voice request is received, and when the source of the communication voice request does not belong to the preset source, the voice text corresponding to the communication voice is obtained; based on the voice text, the voice dialogue text corresponding to the voice text is obtained; based on the voice dialogue text, the category to which the voice dialogue text belongs is identified. Thus, the type of unknown communication voice can be directly and automatically identified through the unknown communication voice, and further, according to the type, the user can be helped to classify the unknown voice information, especially for bad information, to help the user avoid the harassment of bad information, and at the same time, the user is not aware of the identification process, which improves the user experience, and can also help the user provide the required category of unknown voice information.

[0019] In addition, in the embodiment of the present application, an abnormal voiceprint library can also be established, and the abnormal voiceprint library is obtained based on the abnormal voiceprint. So that the voiceprint of unknown voice information can be automatically identified through the abnormal voiceprint in the abnormal voiceprint library, and it is determined whether the unknown voice information belongs to the abnormal voiceprint, so that the unknown voice information can be further processed, and the user can be automatically helped to identify and process the abnormal voice information, reduce the bad harassment brought to the user by the unknown voice information, and improve the user's service experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0021] Figure 1 A flowchart of a method for identifying information according to an exemplary embodiment of the present application;

[0022] Figure 2 A flowchart of a caller identification method according to an exemplary embodiment of the present application;

[0023] Figure 3 A schematic diagram of an interface of a notification message provided by an exemplary embodiment of the present application;

[0024] Figure 4 A schematic diagram of a flow chart of a method for establishing a voiceprint database according to an exemplary embodiment of the present application;

[0025] Figure 5 A flowchart of a method for identifying information according to another exemplary embodiment of the present application;

[0026] Figure 6 A flowchart of a method for setting information of another exemplary embodiment of the present application is provided;

[0027] Figure 7 A flowchart of a method for setting information of another exemplary embodiment of the present application is provided;

[0028] Figure 8 A schematic diagram of an interface for setting conditions provided for an exemplary embodiment of the present application;

[0029] Fig. 9 A schematic diagram of the structure of an information identification system of an exemplary embodiment of the present application;

[0030] Fig.10 A flowchart of a method for identifying information according to an exemplary embodiment of the present application;

[0031] Fig.11 A schematic diagram of the structure of an information identification device provided by an exemplary embodiment of the present application;

[0032] Fig.12 A schematic diagram of the structure of an information identification device provided by an exemplary embodiment of the present application;

[0033] Fig.13 A schematic diagram of the structure of a device for establishing a voiceprint database provided by an exemplary embodiment of the present application;

[0034] Fig.14 A schematic diagram of the structure of an information identification device provided by an exemplary embodiment of the present application;

[0035] Fig.15 A schematic diagram of the structure of a device for setting information provided by another exemplary embodiment of the present application;

[0036] Fig.16 A schematic diagram of the structure of a device for setting information provided by another exemplary embodiment of the present application;

[0037] Fig.17 A schematic diagram of the structure of a computing device provided for an exemplary embodiment of the present application;

[0038] Fig.18 A schematic diagram of the structure of a computing device provided for another exemplary embodiment of the present application;

[0039] Fig.19 A schematic diagram of the structure of a computing device provided for another exemplary embodiment of the present application;

[0040] Fig. 20 A schematic diagram of the structure of a computing device provided for another exemplary embodiment of the present application;

[0041] Fig.21 A schematic diagram of the structure of a computing device provided for another exemplary embodiment of the present application;

[0042] Fig. 22 A schematic diagram of the structure of a computing device provided for yet another exemplary embodiment of the present application. DETAILED DESCRIPTION

[0043] In order to make the purpose, technical solution and advantages of the present application clearer, the technical solution of the present application will be clearly and completely described below in combination with the specific embodiments of the present application and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present application.

[0044] With the development of communication networks, people receive a large number of unknown calls every day, most of which are abnormal calls, such as harassment calls, fraud calls, and sales calls. Since it is not known whether the current call is an abnormal call, there are often problems such as misanswering calls and missing calls, which also brings a lot of inconvenience and troubles to users.

[0045] In the prior art, it is usually necessary to rely on users to manually mark abnormal calls among unfamiliar calls, especially to mark bad unfamiliar calls. However, when users manually mark unfamiliar calls, since the marking of unfamiliar calls depends on the manual triggering of the user, it will take up a large amount of the user's time cost, so the proportion of users reporting to the server is relatively low. Users may only report 10% of the harassment and fraud calls they receive.

[0046] In addition, when users receive bad stranger calls, they may not report them in a timely manner. Perhaps less than 3% of harassed users will immediately report data to the server through the application after being harassed by calls.

[0047] In addition, it is very common for black and gray industry practitioners to frequently change their phone numbers or use number changers. When a number is reported by many people, black and gray industry practitioners can change another number at a very low cost and start harassing again. Therefore, reporting the phone number has very limited defensive capabilities to block unknown calls.

[0048] In the embodiment of the present application, a communication voice request is received, and when the source of the communication voice request does not belong to the preset source, the voice text corresponding to the communication voice is obtained; based on the voice text, the voice dialogue text corresponding to the voice text is obtained; based on the voice dialogue text, the category to which the voice dialogue text belongs is identified. Thus, the type of unknown communication voice can be directly and automatically identified through the unknown communication voice, and further, according to the type, the user can be helped to classify the unknown voice information, especially for bad information, to help the user avoid the harassment of bad information, and at the same time, the user is not aware of the identification process, which improves the user experience, and can also help the user provide the required category of unknown voice information.

[0049] The technical solutions provided by various embodiments of the present application are described in detail below in conjunction with the accompanying drawings.

[0050] Figure 1 This is a flow chart of a method for identifying information of an exemplary embodiment of the present application. The method 100 provided in the embodiment of the present application is executed by an electronic terminal device, such as a mobile phone, a tablet computer, a desktop computer, etc. The method 100 includes the following steps:

[0051] 101: receiving a communication voice request, and obtaining a voice text corresponding to the communication voice when the source of the communication voice request does not belong to a preset source.

[0052] 102: According to the voice text, obtain the voice dialogue text corresponding to the voice text.

[0053] 103: According to the voice conversation text, identify the category to which the voice conversation text belongs.

[0054] The following is a detailed description of steps 101-103:

[0055] 101: receiving a communication voice request, and obtaining a voice text corresponding to the communication voice when the source of the communication voice request does not belong to a preset source.

[0056] Among them, communication voice refers to voice transmitted through the communication network, that is, voice information, such as calls between communication devices (such as mobile phones, landlines, computers, etc.) and Internet calls between applications.

[0057] The source of the communication voice request refers to the source information of the voice, such as mobile phone number, landline number, computer IP, etc.

[0058] The preset source refers to the source of the communication voice request set by the user, or the source of the communication voice request automatically set by the device, such as the mobile phone directory, the mobile phone local phone blacklist, etc.

[0059] Voice text refers to the text content that records the transmission of voice information.

[0060] For example, according to the above, if Figure 2As shown, the user's mobile phone receives a call request, the mobile phone obtains the phone number of the call request, and matches it according to the locally stored phone book. When it does not belong to the phone book, it is determined that the call request is an unfamiliar call request, and the AI ​​application or smart speaker installed on the mobile phone can accept the request. After the call connection is established, the other party of the unfamiliar call will start speaking, and the mobile phone receives the communication voice sent by the other party, and converts the communication voice into voice text through a preset ASR (speech recognition, also known as automatic speech recognition) model, such as "Hello, this is xx insurance" or "Hello, is this Ms. xx?"

[0061] It should be noted that in order to improve the user experience and make the user unaware of the execution of this step, the mobile phone can directly block the communication voice received to the user, that is, the mobile phone temporarily turns off the voice speaker earpiece, and blocks the reception of the user's voice, that is, turns off the microphone for receiving the user's voice.

[0062] In addition, when requesting the source of the communication voice request belonging to the user setting of the user setting, the user can be prompted to answer the call.

[0063] Specifically, the method 100 may further include: obtaining the source of the communication voice request, matching the source with a normal source set preset by the user; and providing the communication voice to the user in case of a successful match.

[0064] The user presets a normal source set, that is, the user sets a trusted phone number set, that is, a phone book.

[0065] For example, as described above, the mobile phone obtains the phone number of the incoming call request and matches it according to the locally stored phone book. If it belongs to the phone book, the name of the number in the phone book is obtained, such as "Mom", and a display interface is provided to display the prompt message to the user, "Mom is calling, answer the call."

[0066] In addition, in order to more accurately determine the source of the unknown call, after matching the number of the unknown caller with the phone book, it can also be compared with the local phone blacklist to expand the comparison range and improve the comparison accuracy.

[0067] Specifically, the method 100 may further include: obtaining the source of the communication voice request, matching the source with a preset abnormal source set; and in case of a successful match, providing a prompt message to the user on whether to accept the communication voice.

[0068] Among them, a telephone number blacklist automatically set by users or applications from pre-set abnormal sources is included, wherein the list may include illegal lists, such as fraudulent telephone numbers, and may also include gray list telephone numbers, such as sales promotion telephone numbers.

[0069] For example, as described above, the mobile phone obtains the phone number of the incoming call request and matches it with the locally stored phone book. If it does not belong to the phone book, it is matched with the blacklist. If it belongs to the blacklist, the user can be prompted according to the type in the blacklist. The mobile phone provides an interface and displays the type, such as "sales call", and the prompt information of "answer or reject".

[0070] More specifically, a match is performed with a preset abnormal source set; in the event of a match with an illegal source, the communication voice request is rejected.

[0071] For example, according to the above, when it belongs to the blacklist and it is an illegal call in the blacklist, such as a fraud call, the mobile phone directly rejects the incoming call request. 102: According to the voice text, obtain the voice conversation text corresponding to the voice text.

[0072] The voice conversation text refers to the information of the voice conversation between at least two parties.

[0073] Specifically, the method of obtaining the voice conversation text may include:

[0074] 1) The acquired voice text is directly used as the voice dialogue text.

[0075] For example, as mentioned above, the AI ​​application or smart speaker on the mobile phone can directly use the first voice text of the unfamiliar call as the voice conversation text, such as "Hello, this is xx insurance."

[0076] 2) Determine the interactive text corresponding to the voice text; take the acquired voice text and the corresponding interactive text as a group of dialogue texts, and take at least one group of dialogue texts as voice dialogue texts; when there are multiple groups of voice dialogue texts, for any two groups of dialogue texts, obtain the latter group of dialogue texts based on the former group of dialogue texts.

[0077] The interactive text refers to the communication text corresponding to the voice text based on human language logic or grammatical logic. For example, the voice text is "How are you?" and the interactive text can be "I am fine", thus forming a set of dialogue texts.

[0078] For example, according to the above, if Figure 2As shown, the AI ​​application or smart speaker on the mobile phone converts the received voice information into voice text through the preset ASR model, such as "Hello, this is xx insurance" or "Hello, are you Ms. xx?". The server determines that the interactive text can be "Hello, who are you?". At this time, the AI ​​application or smart speaker on the mobile phone can form the voice text and the interactive text into a first group of dialogue texts, that is, "Hello, Ms. xx" + "Hello, who are you?"; the voice dialogue text can generate multiple groups of dialogue texts with the interaction of voices. For example, the other party of the strange call receives the interactive text "Hello, who are you?" and can continue to send communication voices, such as "This is xx insurance, do you have any insurance needs?" After receiving the voice text, the AI ​​application or smart speaker on the mobile phone can determine the interactive text as "What types of insurance are there?", and so on to determine multiple groups of dialogue texts. It can be seen that any two groups of dialogue texts have a language logical relationship, that is, the latter group of dialogue texts is obtained based on the former group of dialogue texts.

[0079] Specifically, obtaining a second group of dialogue texts based on a first group of dialogue texts includes: obtaining voice texts in a second group of dialogue texts based on interactive texts in a first group of dialogue texts; after obtaining the voice texts, executing the step of determining interactive texts corresponding to the voice texts.

[0080] Since this example has been described in detail above, it will not be repeated here. This execution method can further improve the voice dialogue text and increase the integrity of the voice dialogue text.

[0081] The acquisition of the voice text in the latter group of conversation texts includes: sending the interactive text through the communication voice, and receiving the corresponding communication voice to acquire the voice text. Acquiring the voice text by receiving the voice information can better understand the intention of the caller, so as to accurately classify the voice conversation text, while consuming the other party's time cost and even manpower cost.

[0082] For example, according to the above, after determining the interactive text, the AI ​​application or smart speaker on the mobile phone can synthesize the interactive text into sound through the preset TTS (Text To Speech) model, and send it to the unknown caller through the established voice call. Then, the communication voice sent by the unknown caller (also known as the other party of the unknown call) is received, and the corresponding voice text is obtained.

[0083] It should be noted that the TTS model uses a neural network to intelligently convert text into a natural speech stream. The TTS model can convert text in real time, and the conversion time can be calculated in seconds. Under its unique intelligent voice control, the voice rhythm of the text output is smooth, making the listener feel natural when listening to the information, without the indifference and awkwardness of machine voice output. The TTS model can cover the first and second level Chinese characters of the national standard, has an English interface, automatically recognizes Chinese and English, and supports mixed reading of Chinese and English. All sounds use real mandarin as the standard pronunciation, achieving fast speech synthesis of 120-150 Chinese characters / minute, and a reading speed of 3-4 Chinese characters / second, so that users can hear clear and pleasant sound quality and coherent and fluent intonation.

[0084] The TTS model can be generated based on artificial learning. The specific generation process belongs to the existing technology and will not be described here.

[0085] Specifically, determining the interactive text corresponding to the voice text includes: determining the semantics of the voice text; and determining the interactive text corresponding to the voice text from the dialogue set according to the semantics. Determining the interactive text by semantics can better help users understand more of the content they need and increase the user's good experience. At the same time, the voice dialogue text can also be classified.

[0086] Among them, semantics refers to the linguistic meaning of spoken text.

[0087] A dialog collection is a collection of multiple sentences.

[0088] For example, as mentioned above, after the AI ​​application or smart speaker on the mobile phone obtains the voice text "Hello, is this Ms. xx?", it can determine the semantics based on the NLU (Natural Language Understanding) model, and then search for the interaction text corresponding to the semantics from the dialogue collection.

[0089] It should be noted that the goal of the NLU model is to convert text into semantic representation. The exact meaning of the words in the text is not important. What is important is the semantic information conveyed by the text. NLU is also called semantic decoding. The NLU model can be generated based on artificial learning. Its specific generation process belongs to the existing technology and will not be repeated here.

[0090] The dialogue set may be the idioms that people usually use, and each idiom may be annotated with its own semantics and the semantics of the speech text that may become part of a set of dialogue texts.

[0091] It should be noted that in order to make the user unaware of this processing method and improve the user experience, when sending voice text via communication voice, the external sound can be shielded from the user, that is, the user cannot hear the sent communication voice.

[0092] 103: According to the voice conversation text, identify the category to which the voice conversation text belongs.

[0093] The category refers to the type of voice conversation text, which may refer to the purpose of language communication, such as sales promotion category, harassment category, fraud category, etc.

[0094] Specifically, identifying the category to which the voice conversation text belongs includes: determining the semantics of the voice conversation text, and determining the category corresponding to the voice conversation text according to the semantics. Determining the voice conversation text by semantics can more accurately classify the voice conversation text.

[0095] For example, according to the above, if Figure 2 As shown, when the AI ​​application or smart speaker on the mobile phone cannot determine the category of the voice dialogue text, it can continuously update the voice dialogue text. When the AI ​​application or smart speaker on the mobile phone can determine the purpose of the subsequent voice dialogue text based on the first voice text, such as the voice text "Hello, this is xx insurance", the semantics can be determined according to the NLU model, and the purpose of the voice text can be further determined according to the NLU model, that is, the category is sales promotion, more specifically, insurance sales promotion. At this time, the first voice text can be directly used as the voice dialogue text.

[0096] When the NLU model cannot temporarily determine the semantics of the current voice dialogue text, the voice dialogue text can be continuously updated until the NLU model can determine the purpose, that is, the category, of the current voice dialogue text.

[0097] In the embodiment of the present application, the NLU model can be placed on the server side, and the other models mentioned above can also be set on the server side. That is to say, the specific implementation method of obtaining voice text and identifying the category to which it belongs can be implemented through the server side. When the server side is implemented, a terminal is required, such as a mobile phone, to forward the communication voice request to the server side.

[0098] When implemented on a local terminal, a higher level of user authorization is required, but it can effectively prevent user information from being leaked.

[0099] When the phone list cannot accurately determine the incoming call request, in order to improve the recognition progress and prevent the user from receiving calls that the user dislikes and threatens the user's economic security, the recognition can continue, such as through the voiceprint library, to avoid the resource occupation caused by recognizing the voice conversation text.

[0100] Specifically, the method 100 may further include: when the source of the communication voice request does not belong to a preset source, obtaining the voiceprint of the communication voice; when the voiceprint does not belong to an abnormal voiceprint library, executing the step of obtaining the voice text corresponding to the communication voice.

[0101] The abnormal voiceprint library is used to store multiple abnormal voiceprints. Voiceprint refers to the sound wave spectrum that carries speech information displayed by electroacoustic instruments. It is an attribute that can uniquely identify a sound, such as the sound that can uniquely identify a communication voice. It should be understood that the voiceprint refers to the voiceprint of the initiator of the communication voice (also called the caller).

[0102] Abnormal voiceprints refer to voiceprints of black and gray industries, which can include voiceprints belonging to illegal industries (also called illegal voiceprints), fraud voiceprints, and voiceprints belonging to industries that are on the edge of the law (also called gray voiceprints), such as harassment voiceprints and sales promotion voiceprints.

[0103] For example, as described above, after receiving the communication voice, the AI ​​application or smart speaker on the mobile phone extracts the voiceprint from the communication voice through the voiceprint recognition model, and matches the voiceprint with the abnormal voiceprint library. Only when it does not belong to the abnormal voiceprint library, step 101 is executed.

[0104] In addition, the method 100 may further include: establishing an abnormal voiceprint library, where the abnormal voiceprint library is obtained based on the abnormal voiceprint.

[0105] For example, a mobile phone can receive multiple abnormal voiceprints reported to the local mobile phone by the user through a reporting application installed on the mobile phone, collect these abnormal voiceprints together, and generate an abnormal voiceprint library for storing these abnormal voiceprints. The abnormal voiceprint library can have an ID and a label "abnormal category". Among them, each abnormal voiceprint can have an ID and can also be labeled according to its specific category. For example, the voiceprint label of the fraud category with ID "aa" is set to "illegal category-fraud category", etc., and the voiceprint label of the sales promotion category with ID "bb" is set to "abnormal category-sales promotion category", etc.

[0106] In addition, each voiceprint in the abnormal voiceprint library can be divided into illegal categories and other categories (ie, gray categories), and two sub-libraries are set in the abnormal voiceprint library to store voiceprints of illegal categories and voiceprints of other categories respectively.

[0107] It should be noted that in addition to the reporting method, the method of obtaining abnormal voiceprints can also be through the application server, the relevant public security management platform or other platforms to obtain multiple abnormal voiceprints stored locally to generate an abnormal voiceprint library.

[0108] By building a voiceprint database for the black and gray industries, even if the black and gray industries change their phone numbers, their voiceprints cannot be changed, and thus their voiceprints can be quickly identified. If they want to continue marketing or fraud, they will need to keep changing people while changing their numbers, which greatly increases the manpower costs of the black and gray industries.

[0109] In order to help users identify more voiceprints, the established abnormal voiceprint library should be continuously updated to ensure that more voiceprints can be effectively identified. The update method may include:

[0110] 1) When the category is an abnormal category, the voiceprint is updated as an abnormal voiceprint to the abnormal voiceprint library.

[0111] For example, as mentioned above, the AI ​​application or smart speaker on the mobile phone can determine the semantics of the speech text based on the NLU (Natural Language Understanding) model, and after determination, the corresponding semantics can be searched from the semantic sets of different categories. When it is determined that the speech belongs to the sales promotion category in the abnormal category, the communication speech can be determined to be an abnormal speech. The voiceprint is then extracted from the communication speech through the voiceprint recognition model, and the extracted voiceprint is updated as an abnormal voiceprint to the abnormal voiceprint library.

[0112] It should be noted that the ASR model is a preset speech recognition model, and its goal is to convert the vocabulary content in human speech into computer-readable input text, including characters, etc. The ASR model is generated based on artificial learning, and its specific generation process belongs to the existing technology and will not be repeated here.

[0113] The goal of the NLU model is to convert text into semantic representations. The exact meaning of the words in the text is not important. What is important is the semantic information conveyed by the text. NLU is also called semantic decoding. The NLU model can be generated based on artificial learning. Its specific generation process belongs to the existing technology and will not be repeated here.

[0114] In addition, the voiceprint recognition model is a service that identifies the speaker based on the speaker's voice characteristics. Identity recognition has nothing to do with accent or language, and can be used for speaker recognition and confirmation, and is widely used in financial security, smart homes, smart buildings and other fields.

[0115] The voiceprint recognition model can only extract the voiceprint, that is, extract the characteristics of the sound, which can be the sound spectrum and sound wavelength, etc. Alternatively, the voiceprint recognition model can have a built-in voiceprint library to directly identify the voiceprint, such as whether the voiceprint is a bad voiceprint or not.

[0116] Regardless of the function of the voiceprint recognition model, the model establishment process belongs to the prior art and will not be repeated here. It should be understood that the voiceprint library is a carrier for recording bad voiceprints. The voiceprint library is not limited to one implementation form, and all implementation forms that can realize the function of the carrier belong to the protection scope of the embodiments of this application.

[0117] The updating method may also include: 2) receiving the reported voiceprint and the reporting category, and when the reporting category is an abnormal category, updating the reported voiceprint as an abnormal voiceprint to the abnormal voiceprint library.

[0118] For example, according to the above, a mobile phone can receive multiple voiceprints reported by a user to the local mobile phone through a reporting application installed on the mobile phone, and the voiceprint category, such as sales category, is also reported together with the voiceprint. After receiving the reported voiceprint, the mobile phone determines that the voiceprint category belongs to an abnormal category, and then stores the reported voiceprint as an abnormal voiceprint in the abnormal voiceprint library to update the voiceprint.

[0119] When the category of the reported voiceprint belongs to other categories, such as entertainment category, friends and relatives category, etc., the server will store the reported voiceprint in the voiceprint library of the corresponding category, so that when the user misses a call or the user is not next to the phone, the server can automatically record the voice information of the corresponding category of the unfamiliar call for the user and inform the user so that the user can deal with it as soon as possible.

[0120] It should be noted that when users report voiceprints on their own, after answering a call voice message through a mobile phone, the user can extract the voiceprint of the voice message through the local voiceprint recognition module of the mobile phone, report the voiceprint locally, and store it locally on the mobile phone. In addition, the voiceprint can be reported to the server. When reporting the stored voiceprint to the server, the voiceprint category is specified. After reporting, the mobile phone stores the reported voiceprint in the local voiceprint library of the mobile phone through the installed reporting application for record.

[0121] In order to prevent the existence of multiple redundant voiceprints in the abnormal voiceprint library after the update, and save the resources occupied by the abnormal voiceprint library, the method 100 may also include: when it is determined that the voiceprint does not exist in the abnormal voiceprint library, the voiceprint is updated as an abnormal voiceprint to the abnormal voiceprint library.

[0122] For example, according to the above, before the AI ​​application or smart speaker on the mobile phone updates the voiceprint as an abnormal voiceprint to the abnormal voiceprint library, it can first match the voiceprint with multiple voiceprints in the abnormal voiceprint library. If no match is found, it is deemed that it does not exist in the abnormal voiceprint library, and the voiceprint can be updated to the abnormal voiceprint library. Otherwise, the voiceprint can be discarded. The same is true when the server is updated.

[0123] It should be noted that this method is also used for other types of voiceprint libraries, so it will not be repeated here.

[0124] In addition, the identification method of the abnormal voiceprint library may include: matching the voiceprint with multiple voiceprints in the abnormal voiceprint library, and when the match is successful, determining that the voiceprint is an abnormal voiceprint.

[0125] For example, according to the above, if Figure 2 As shown, after the call connection is established by the AI ​​application or smart speaker on the mobile phone, the other party of the unfamiliar call will start speaking, and the voiceprint is extracted from the communication voice through the voiceprint recognition model, and the extracted voiceprint is matched with each voiceprint in the abnormal voiceprint library. When there is an abnormal voiceprint that is the same as the extracted voiceprint, the extracted voiceprint is determined to be an abnormal voiceprint.

[0126] When the received communication voice belongs to a communication voice request, and the extracted voiceprint is determined to be an abnormal voiceprint, the voice information can be processed, thereby further automatically helping the user to process the voice information, improving the user's service experience, and reducing the probability of the user being harassed. Among them, the processing method may include: when the matched voiceprint belongs to an illegal voiceprint among the abnormal voiceprints, the corresponding communication voice is determined to be an illegal communication voice, and the reception of the voice information is terminated.

[0127] For example, according to the foregoing, after the AI ​​application or smart speaker on the mobile phone determines that the extracted voiceprint is an abnormal voiceprint and that the corresponding communication voice is an abnormal communication voice, the AI ​​application or smart speaker on the mobile phone can end the communication voice call by itself.

[0128] When the AI ​​application or smart speaker on the mobile phone determines that the voiceprint does not exist in the abnormal voiceprint library, it can also make further judgments on the communication voice information to which the voiceprint belongs in other ways, determine the category of the communication voice information, and then process it.

[0129] By determining that the category is an abnormal category through steps 101-103, the type of voice information can be directly determined, thereby further determining the voiceprint category or processing the communication voice.

[0130] Among them, abnormal voice information refers to voice corresponding to abnormal voiceprints, such as voice of black and gray industries, which can include voice belonging to illegal industries (also called illegal category voice), fraud category telephone voice, and voice belonging to industries on the edge of the law (also called gray category voice), such as harassment category telephone voice and sales category telephone voice.

[0131] The method of determining the communication voice type includes: when the category is an abnormal category, determining the voice information as abnormal voice information.

[0132] For example, according to the above, Figure 2 As shown, when the AI ​​application or smart speaker on the mobile phone determines that the category of the voice conversation text is a sales promotion category, it determines that the voice conversation text is an abnormal category and determines that the corresponding communication voice is a sales promotion voice.

[0133] In order to more accurately determine the way to process voice information, the method 100 also includes: when the category is an illegal category in the abnormal category, determining that the communication voice is an illegal communication voice in the abnormal communication voice and the voiceprint is an illegal voiceprint in the abnormal voiceprint, and terminating the communication receiving voice.

[0134] For example, according to the above, if Figure 2 As shown, when the AI ​​application or smart speaker on the mobile phone determines that the category of the voice conversation text is a fraud category, it determines that the voice conversation text is an illegal category, and determines that the corresponding communication voice is a fraud voice, that is, an illegal voice, and can directly end the voice call. At the same time, since the illegal category is determined to be an abnormal category, the voiceprint can be extracted, and the voiceprint can be stored as an illegal voiceprint in the abnormal voiceprint library, and the abnormal voiceprint library is further updated.

[0135] In addition, when the user has a need to obtain other categories in the abnormal category, for example, the user wants to buy insurance or buy a house, then the user has a need for insurance promotion voice information and house purchase promotion voice information. In order to meet the needs of the user and improve the user's service experience at the same time, and help the user eliminate unnecessary and illegal communication voices, the method 100 also includes: when the category is the category required by the user, then according to the content of the category required by the user, obtain the voice dialogue text that matches the content, and the category required by the user may belong to the abnormal category.

[0136] The categories required by users can be pre-set. For example, users can directly enter information on the setting condition page provided by the AI ​​application on the mobile phone or the smart speaker to obtain more information. More specifically, they can also inquire about the price and compensation of major disease insurance. This helps to understand the content required by users and further complete the subsequent acquisition of the content required by users, especially for the selection of dialogue sets.

[0137] For example, according to the above, when the AI ​​application or smart speaker on the mobile phone determines that the voice dialogue text is for insurance promotion according to the NLU model, the purchase of critical illness insurance in the insurance promotion required by the user is determined according to the information of the user's setting conditions, and the AI ​​application or smart speaker on the mobile phone replaces the user to talk to the insurance promotion phone. At this time, the server can select a dialogue set related to insurance consultation. For example, according to the information that the user needs to know, such as the price of critical illness insurance, the interactive text with the semantics of consulting insurance, such as "the price of critical illness insurance" or "how to pay for critical illness", etc., can be selected from the corresponding dialogue set for purchasing insurance, and the interactive text is generated into voice in turn, sent to the stranger caller, and wait for the other party's voice reply. And determine the voice text of the other party's reply, continue to update the voice dialogue text, until the information required by the user is obtained, and then hang up the phone. The AI ​​application or smart speaker on the mobile phone obtains the voice dialogue text required by the user.

[0138] It should be noted that the categories required by users may overlap with abnormal categories, such as sales calls. When users have needs, except for illegal fraud calls from strangers, which will cause losses to users, other categories of strange calls can be based on user needs.

[0139] When the category of the voice conversation text is other categories, such as entertainment category, friends and relatives category, etc., the corresponding voiceprint can be determined as other category voiceprint. And the other category voiceprint is set to the corresponding category voiceprint library, such as setting the voiceprint of the friends and relatives category to the friends and relatives category voiceprint library. So that when the user misses a call later or the user is not next to the phone, the voice information of the corresponding category of the unfamiliar call can be automatically recorded for the user and informed so that the user can handle it as soon as possible. Other categories can also be regarded as categories required by the user.

[0140] In addition to identifying abnormal voice information based on the abnormal voiceprint library and the category to which the voice conversation text belongs, other methods can also be used to identify abnormal voice information. In particular, when the abnormal voiceprint library cannot be identified, other methods can be used to improve the recognition rate. Or before the abnormal voiceprint library is identified, other methods can be used to identify first to simplify the recognition process.

[0141] Specifically, the method 100 further includes: determining the category of the source of the communication voice request according to the category.

[0142] The source of the communication voice request may also be referred to as a communication identifier, which is an identifier that can uniquely identify the source of the communication voice, such as a telephone number.

[0143] For example, as mentioned above, after the AI ​​application or smart speaker on the mobile phone determines the category of the voice conversation text, the category of the phone number of the unknown caller is determined based on the category. If the category is fraud, then the phone number is a fraud phone number.

[0144] It should be noted that telephone numbers can be marked by tagging.

[0145] In addition, the method 100 further includes: setting the source of the communication voice request to a corresponding source category set, thereby enriching each source category set and identifying more categories.

[0146] The source category set, also called the communication identification category set, is a set consisting of multiple sources or communication identifications. The set can be sources or communication identifications of the same category, or multiple categories. When there are multiple categories, each category will be marked. The set can be a blacklist or an abnormal number library.

[0147] For example, according to the above, if Figure 2 As shown, the AI ​​application or smart speaker on the mobile phone can also store categorized phone numbers, such as fraudulent phone numbers, in a blacklist of fraudulent phone numbers.

[0148] In order to better identify voice information, widely identify a large amount of voice information, and further improve the recognition rate, the communication identification can be updated. The updating method may include: receiving the source and reporting category of the reported communication voice request, and storing the source of the reported communication voice request in the corresponding communication identification category set according to the reporting type.

[0149] For example, as described above, the AI ​​application or smart speaker on the mobile phone can receive multiple strange caller numbers reported to the local mobile phone by the user through the reporting application installed on the mobile phone. Along with the strange caller numbers, the category of the strange caller numbers, such as sales category, is also reported. After the AI ​​application or smart speaker on the mobile phone receives the reported strange caller numbers, it determines that the category of the strange caller numbers belongs to the abnormal category, and then stores the reported strange caller numbers as abnormal numbers in the abnormal number library and updates the numbers.

[0150] In order to update the number database more accurately and eliminate factors of intentional human disturbance, the method 100 also includes: for any reporting category, according to the number of reports of the source of the communication voice request, the source of the reported communication voice request is stored in the corresponding source category set.

[0151] For example, as mentioned above, when an AI application or smart speaker on a mobile phone receives a reported fraud number and the number of reports of the fraud number exceeds a threshold, the fraud number is identified as a real fraud number and is updated as a fraud number in the abnormal number database.

[0152] In addition to accurately updating the number database by reporting the number of times, the accuracy of the update can be further improved by reporting the number of users.

[0153] Specifically, the method may include: for any reporting category, according to the number of reports of the source of the communication voice request and the number of reporting users, storing the reported source of the communication voice request in a corresponding source category set.

[0154] For example, as mentioned above, the AI ​​application or smart speaker on the mobile phone receives a fraud number reported to the local mobile phone. When the number of times the fraud number is reported is greater than a threshold and the number of users who report the number is greater than the threshold, the fraud number is identified as a real fraud number and is updated as a fraud number in the abnormal number database.

[0155] The process of identifying abnormal voice information through an abnormal number library may include: when a forwarded communication voice request is received, obtaining the source carried by the communication voice request; matching the source with multiple source category sets, and when the source matches a source of an illegal category, rejecting the communication voice request.

[0156] For example, as mentioned above, the AI ​​application or smart speaker on the mobile phone first obtains the number of the strange caller, such as the mobile phone number "186xxxxxx", and compares the mobile phone number with the phone number blacklist that can be stored locally to determine whether the mobile phone number exists in the blacklist. If so, the AI ​​application or smart speaker on the mobile phone can directly hang up the strange call and make a record of the situation, such as "The call from the mobile phone number "186xxxxxx" is a fraudulent call" and make the situation available to the user for review.

[0157] It should be noted that the blacklist can record fraudulent phone numbers, sales promotion phone numbers, harassment phone numbers, etc., and can also record multiple categories of phone numbers and distinguish them by various phone labels. It should be understood that the blacklist is a carrier for recording bad numbers, and the blacklist is not limited to one implementation form. All implementation forms that can realize the function of the carrier belong to the protection scope of the embodiments of this application.

[0158] In addition, the server can also update the number database according to the user's reporting method.

[0159] In addition: if the source of the communication voice request of the illegal category is not matched, the communication voice request and the voice information are received, and the voiceprint corresponding to the voice information is determined. The voice information is identified through the voiceprint.

[0160] In order to further improve services and enhance user experience, the automatically identified communication voice, number and voiceprint can be informed to the user.

[0161] Specifically, the method 100 further includes: providing the category and the source of the communication voice to the user.

[0162] For example, according to the above, if Figure 2 As shown, after the AI ​​application or smart speaker on the mobile phone determines the category of the voice information, it can directly provide the user with an interface message, which can record the phone number of the incoming call and the category of the phone number as fraud. After receiving the message, the user can click on the message to view it.

[0163] In addition, since the server can also perform category recognition, after determining the category of the voice information, the server can also send a notification message to the user's mobile phone in the form of text messages or application notifications, and the user can click on the message to view it.

[0164] It should be noted that when the incoming call does not belong to the category required by the user, only the above content can be sent.

[0165] In order to improve the user experience, when the incoming call belongs to the category required by the user, the method 100 further includes: notifying the user of the voice conversation text.

[0166] For example, according to the above, if Figure 2 As shown, the AI ​​application or smart speaker on the mobile phone can also set the voice conversation text in the message and display it to the user together with the incoming call phone number and its category. When the user views the notification message, the voice conversation text can be seen. The server can also send these contents to the user's mobile phone so that the user can view the message.

[0167] in, Figure 3 The display interface 300 of the notification message is shown. What is recorded in the interface 300 is the call record, that is, the phone record, which includes the incoming number, such as the phone number of an unfamiliar caller, the category, that is, the mark: marketing category. Loan (also known as sales promotion category. Loan), the number status and voiceprint status: automatically reported and stored, and the voice conversation text, that is, the content of the chat record.

[0168] In addition, you can also record the time of unknown calls and call duration, etc.

[0169] It should be noted that, regardless of whether the category is what the user wants or is an illegal category, the notification message can carry the chat history, which can help the user better understand the purpose of the unfamiliar call.

[0170] After discovering an abnormal call, especially an illegal call, you can use the simplest strategy to hang up, or you can pretend to be a real person to increase the other party's cost. If the user really needs marketing information, you can also ask for information on their behalf. The user can configure the strategy and let the server execute it on their behalf. In addition, the embodiment of the present application can also realize the function of automatically recording the content of incoming calls. When the user is inconvenient to answer the phone for a period of time or under certain conditions, but needs to know the content of the calls during the period, the method of the embodiment of the present application can be used to realize the text recording of the calls that are inconvenient to answer. When it is convenient for the user to check the recorded text, he will know what happened during the period and whether he needs to call back.

[0171] In order to improve the accuracy of the category recognition of the voice dialogue text in the artificial intelligence, the user's correction information on the voice information category can also be received. Specifically, the method 200 also includes: receiving the category adjustment information; adjusting the category according to the adjustment information; using the adjusted category and the corresponding voice dialogue text as samples for optimizing the model; wherein the model is used to determine the semantics of the voice dialogue text, and the category corresponding to the voice dialogue text is determined according to the semantics.

[0172] The adjustment information refers to the modification information used to modify the category.

[0173] model may refer to a NUL model.

[0174] For example, according to the above, if Figure 2As shown, after receiving the notification message, the user can determine whether the category determined by the AI ​​application or smart speaker on the current mobile phone is correct based on the content recorded in the notification message, such as the phone number and / or chat history of the unknown caller. When the user determines that the category is incorrect, the user can send adjustment information to the AI ​​application or smart speaker through the mobile phone, such as changing the category "fraud" to "harassment". The AI ​​application or smart speaker on the mobile phone receives the adjustment information sent by the user's mobile phone. The AI ​​application or smart speaker on the mobile phone will send an audit notification to the front end of the manual reviewer, such as a computer, and the manual reviewer will review the adjustment information. When the review is passed, the AI ​​application or smart speaker on the mobile phone will receive a message sent by the manual reviewer through the computer that the review is passed. After receiving the message, the AI ​​application or smart speaker on the mobile phone can modify the category according to the adjustment information. For example, change the fraud category to the harassment category. In addition, the voiceprint, phone number, and chat history associated with this category must be modified to the harassment category. In addition, the AI ​​application or smart speaker on the mobile phone will also store the chat history as a negative sample for training the NLU model to optimize the NLU model, so that the NLU model can be more optimized and have higher processing accuracy.

[0175] In addition, the server can also execute the above content implemented by AI applications on mobile phones or smart speakers.

[0176] in, Figure 3 The interface 300 also shows an adjustment box “Inaccurate, manually re-mark” for adjusting information. The user can click on the adjustment box and re-mark the category, such as inputting the type by the user himself, or selecting from multiple categories provided in the interface 300. After marking the new category, the user performs a confirmation operation, so that the user's mobile phone can respond to the operation and send the adjustment information to the AI ​​application or smart speaker on the mobile phone.

[0177] In addition, the method 100 further includes: after acquiring the voice dialogue text corresponding to the voice text, the method 100 further includes: providing the user with a viewing address of the voice dialogue text.

[0178] For example, as mentioned above, when an AI application or smart speaker on a mobile phone is talking to an unknown caller, it can obtain the voice conversation text while recording or saving the voice conversation text, and display the recorded local address or local storage address (i.e., local viewing address) to the user in the form of a message. The user can click on the recorded local address to view the voice conversation text.

[0179] It should be noted that since the voice conversation text can be continuously updated, when the user sees the voice conversation text, he can continue to view the updated content by refreshing. At this time, when the voice conversation text is updated, the AI ​​application or smart speaker on the mobile phone can send an update message to the voice conversation text page, prompting the user to perform the update operation and view the updated content.

[0180] In order to enhance the user's sense of participation, better understand artificial intelligence, and know whether their needs are met, the method 100 also includes: recording the conversation voice to which the voice conversation text belongs; and providing the user with a listening address for the conversation voice to which the voice conversation text belongs.

[0181] For example, as mentioned above, when an AI application or smart speaker on a mobile phone is talking to an unknown caller, it can obtain the voice conversation text while recording or saving the communication voice of the unknown caller and the communication voice of the interactive text, that is, the communication voice of the semantic conversation text, and display the recorded local address or saved local address (that is, the local listening address) to the user in the form of a message. The user can click on the recorded address to listen to the communication voice of the voice conversation text.

[0182] Since the voice conversation text is continuously updated, the communication voice is also continuously updated. The communication voice can be continuously updated in the form of live broadcast without the need for users to update it. The AI ​​application or smart speaker on the mobile phone can continuously record or save the media stream of the communication voice, so that users can continue to listen to the communication voice until the call ends.

[0183] It should be noted that, for the embodiments of the present application, the categories of communication voice, the categories of sources, the categories of voice dialogue text, and even the categories of voiceprints are related to each other. For example, the category of a mobile phone number is the same as the category of the voiceprint corresponding to the number, the category of the corresponding communication voice, and the category of voice dialogue text. Knowing one of the categories, the other categories can also be known.

[0184] The embodiment of this application adopts artificial learning and all-round active discovery to build a black and gray industry phone number database, such as mining marketing fraud calls through the NLU model to expand the blacklist of phone numbers and greatly improve the efficiency of discovering black and gray industry calls. Its mining efficiency and issuance efficiency can be 10 times or even higher than that of manual mining.

[0185] The embodiment of the present application can be embedded in an APP in the form of software and installed in a terminal, or embedded in a small program, and the user can directly pull the small program application through the network without downloading, thereby reducing the resource occupation of the terminal.

[0186] The server can also execute the content implemented by the terminal device, such as executing method 100. The execution steps are similar to the server execution steps in the previous text, and will not be repeated here. When the server executes method 100, it is necessary to forward the communication voice request, receive the request, obtain the voiceprint, determine the voiceprint category, determine the category of the voice conversation text, and determine the category of the number.

[0187] Since the cost of frequently changing phone numbers is very low for people working in the black and gray industries, it is very difficult to accurately identify many different numbers. In order to curb the frequent number changes that harass users, the identification of people working in the black and gray industries is strengthened. The present invention also provides a method for establishing a voiceprint library.

[0188] Figure 4 The method 400 provided in the embodiment of the present application is executed by an electronic device, such as a conventional server, a cloud server, a mid-terminal device, such as a mobile phone, a tablet computer, and a desktop computer, etc. The method 400 includes the following steps:

[0189] 401: Establish an abnormal voiceprint database, which is obtained based on abnormal voiceprints.

[0190] Since the specific implementation of step 401 has been described in detail above, it will not be repeated here.

[0191] By building a voiceprint database for the black and gray industries, even if the black and gray industries change their phone numbers, their voiceprints cannot be changed, and thus their voiceprints can be quickly identified. If they want to continue marketing or fraud, they will need to keep changing people while changing their numbers, which greatly increases the manpower costs of the black and gray industries.

[0192] In order to help users identify more voiceprints, the established abnormal voiceprint library should be continuously updated to ensure effective identification of more voiceprints. The method 400 also includes: receiving communication voice and determining the voiceprint corresponding to the communication voice; when the communication voice is determined to be an abnormal communication voice, updating the voiceprint as an abnormal voiceprint to the abnormal voiceprint library.

[0193] The method 400 further includes: receiving a reported voiceprint and a reported category, and when the reported category is an abnormal category, updating the reported voiceprint as an abnormal voiceprint to an abnormal voiceprint library.

[0194] In addition, the method 400 also includes: receiving communication voice and determining a voiceprint corresponding to the communication voice; matching the voiceprint with multiple voiceprints in an abnormal voiceprint library, and when the match is successful, determining that the voiceprint is an abnormal voiceprint.

[0195] After the voiceprint is identified, the communication voice corresponding to the voiceprint can be directly processed, which is convenient and fast. The method 400 also includes: when the matched voiceprint belongs to an illegal voiceprint among the abnormal voiceprints, determining that the corresponding communication voice is an illegal communication voice, and terminating the reception of the communication voice.

[0196] In order to prevent the existence of multiple redundant voiceprints in the abnormal voiceprint library after the update, and to save the resources occupied by the abnormal voiceprint library, before updating the abnormal voiceprint library, the method 400 also includes: when it is determined that the voiceprint does not exist in the abnormal voiceprint library, the voiceprint is updated as an abnormal voiceprint to the abnormal voiceprint library.

[0197] When the voiceprint cannot be identified in the voiceprint library, the source of the incoming call can be accurately identified by identifying the category to which the voice conversation text belongs, thereby improving the recognition progress. The method 400 also includes: when the voiceprint does not exist in the abnormal voiceprint library, obtaining the voice text corresponding to the communication voice; based on the voice text, obtaining the voice conversation text corresponding to the voice text; based on the voice conversation text, identifying the category to which the voice conversation text belongs.

[0198] It should also be noted that since the specific implementation methods of the embodiments of the present application have been explained in detail above, they will not be repeated here.

[0199] In addition, for other parts not described in detail in this embodiment, reference may be made to the relevant description of the embodiment shown in the above method 100, and will not be repeated here.

[0200] In order to efficiently and massively mine unfamiliar calls from potential black and gray industries, while increasing the human cost of black and gray industries, thereby reducing the generation of black and gray industries, the embodiment of the present application also provides an information identification method.

[0201] Figure 5 This is a flow chart of an information identification method according to an exemplary embodiment of the present application. The method 500 provided in the embodiment of the present application is executed by an electronic device, such as a conventional server or a cloud server, and the method 500 includes the following steps:

[0202] 501: Accept the forwarded communication voice request and obtain the voice text corresponding to the voice information.

[0203] 502: According to the voice text, obtain the voice dialogue text corresponding to the voice text.

[0204] 503: Identify the category to which the voice conversation text belongs according to the voice conversation text.

[0205] Since the specific implementation of steps 501 to 503 has been described in detail above, they will not be repeated here. It is only explained that the server receives a communication voice request for call forwarding from a user's mobile phone through the virtual number operator deployed thereon. The request comes from an unknown caller. The server accepts the request through the virtual number operator, thereby establishing a call connection between the unknown caller and the server through the virtual number operator. After the call connection is established, the other party of the unknown caller will start speaking, and the server receives the communication voice sent by the other party and converts the communication voice into speech text through the preset ASR model.

[0206] Specifically, obtaining the voice dialogue text corresponding to the voice text includes: directly using the obtained voice text as the voice dialogue text.

[0207] Specifically, obtaining the voice dialogue text corresponding to the voice text includes: determining the interactive text corresponding to the voice text; taking the obtained voice text and the corresponding interactive text as a group of dialogue texts, and taking at least one group of dialogue texts as voice dialogue text; when there are multiple groups of voice dialogue texts, for any two groups of dialogue texts, obtaining the latter group of dialogue texts according to the former group of dialogue texts.

[0208] Specifically, obtaining a second group of dialogue texts based on a first group of dialogue texts includes: obtaining voice texts in a second group of dialogue texts based on interactive texts in a first group of dialogue texts; after obtaining the voice texts, executing the step of determining interactive texts corresponding to the voice texts.

[0209] Specifically, obtaining the voice text in the latter group of dialogue texts includes: sending the interactive text through communication voice, and receiving the corresponding communication voice to obtain the voice text.

[0210] Specifically, determining the interactive text corresponding to the voice text includes: determining the semantics of the voice text; and determining the interactive text corresponding to the voice text from the dialogue set according to the semantics.

[0211] Specifically, identifying the category to which the voice conversation text belongs includes: determining the semantics of the voice conversation text, and determining the category corresponding to the voice conversation text according to the semantics.

[0212] In addition, the method 500 further includes: receiving the source of the reported communication voice request and the reporting category, and storing the source of the communication voice request in a corresponding source category set according to the reporting type.

[0213] In addition, the method 500 further includes: for any reporting category, according to the number of reports of the source, storing the source of the reported communication voice request into a corresponding source category set.

[0214] In addition, the method 500 further includes: for any reporting category, according to the number of reporting times of the source and the number of reporting users, storing the source of the reported communication voice request into a corresponding source category set.

[0215] In addition, the method 500 further includes: sending the updated source category set to the corresponding device for updating.

[0216] It should be noted that since the specific implementation has been described in detail in the embodiment of method 100 above, and the specific implementation of method 500 is similar to the specific implementation of method 100 in the above text, it will not be described in detail here.

[0217] It should also be noted that for other parts not described in detail in this embodiment, reference may be made to the relevant description of the embodiment shown in the above method 100, and will not be repeated here.

[0218] In order to ensure that users are unaware of automatic voice information recognition and at the same time help users screen unfamiliar calls that may cause harassment and economic losses to users, an embodiment of the present application also provides an information recognition method.

[0219] Figure 6 This is a flow chart of a method for setting information in an exemplary embodiment of the present application. The method 600 provided in the embodiment of the present application is executed by an electronic terminal device, such as a mobile phone terminal, and the method 600 includes the following steps:

[0220] 601: Obtain the user communication voice identifier.

[0221] 602: Bind the user communication voice identifier with the transfer communication voice identifier, so that under preset conditions and according to the binding relationship, the communication voice request is forwarded to the device to which the transfer communication voice identifier belongs.

[0222] 603: Set the preconditions.

[0223] The following is a detailed description of steps 601 to 603:

[0224] 601: Obtain the user communication voice identifier.

[0225] The communication voice identifier is also called a communication identifier, such as a telephone number.

[0226] For example, the mobile phone obtains the execution code of the application applet from the installed application, runs the application applet on the mobile phone according to the execution code, and displays the interface of the application applet. The application applet can provide a number binding interface. The binding interface requires the user to enter the user's mobile phone number and the number to which the incoming call is forwarded.

[0227] The user enters the user's mobile phone number on the interface, and the mobile phone responds to the input operation, obtains the mobile phone number, and can display it on the interface.

[0228] 602: Bind the user communication voice identifier with the transfer communication voice identifier, so that under preset conditions and according to the binding relationship, the communication voice request is forwarded to the device to which the transfer communication voice identifier belongs.

[0229] The forwarding communication voice identifier is also a communication identifier, which is a communication identifier of a device for receiving a forwarding voice request, such as a forwarding number.

[0230] For example, according to the above, the user enters the incoming call forwarding number on the interface, and the mobile phone responds to the input operation, obtains the incoming call forwarding number, and can display it on the interface. In response to the user's binding operation, the two numbers are bound to achieve call forwarding.

[0231] 603: Set the preconditions.

[0232] The preset conditions refer to the conditions for voice request transfer that are preset by the user, such as an unfamiliar call, that is, the unfamiliar call will be transferred.

[0233] For example, according to the above, the user enters the preset condition: unfamiliar call on the interface, and the mobile phone responds to the input operation and obtains the preset condition: unfamiliar call, which can be displayed on the interface. After completing the binding operation, when an unfamiliar call comes in, the unfamiliar call request can be forwarded to the device of the forwarding number. It should be understood that the device is deployed with a virtual number operator.

[0234] In addition, the method 600 further includes: setting a processing method for the corresponding speech under preset conditions.

[0235] The processing methods may include: ending the voice message, that is, hanging up the call, recording the text of the incoming voice message, etc.

[0236] For example, according to the above, the user inputs the processing method on the interface: hang up the call, and the mobile phone responds to the input operation, obtains the processing method, and can display it on the interface. After completing the binding operation, when an unfamiliar call comes in, the unfamiliar call request can be forwarded to the device of the forwarding number. When the unfamiliar call belongs to an illegal category, the call is hung up directly.

[0237] In addition, the method 600 also includes: receiving a communication voice request and obtaining a communication voice identifier corresponding to the communication voice request; when the obtained communication voice identifier meets a preset condition, forwarding the communication voice request to a device to which the communication voice identifier belongs according to a binding relationship.

[0238] In addition, the method 600 further includes: obtaining the voiceprint of the communication voice, and reporting the voiceprint and the category of the voiceprint.

[0239] In addition, the method 600 further includes: matching the voiceprint with the reported voiceprint in the local voiceprint database, and when no corresponding voiceprint is matched, performing the step of reporting the voiceprint and the category of the voiceprint.

[0240] In addition, the method 600 further includes: acquiring a communication voice identifier, and reporting the communication voice identifier and a category of the communication voice identifier.

[0241] In addition, the method 600 also includes: receiving and displaying a communication voice category corresponding to the communication voice request and a voice dialogue text corresponding to the communication voice.

[0242] In addition, the method 600 further includes: receiving a communication voice request corresponding to a voice dialogue text of the communication voice; determining the category of the communication voice in response to an input operation of the user; and reporting the category and the corresponding voice dialogue text.

[0243] It should be noted that, since the above specific implementation has been described in detail in the embodiment of method 100 above, it will not be described in detail here.

[0244] It should also be noted that for other parts not described in detail in this embodiment, reference may be made to the relevant description of the embodiment shown in the above method 100, and will not be repeated here.

[0245] In order to ensure that users are unaware of automatic voice information recognition and at the same time help users screen unfamiliar calls that may cause harassment and economic losses to users, an embodiment of the present application also provides an information recognition method.

[0246] Figure 7 This is a flowchart of a method for setting information in an exemplary embodiment of the present application. The method 700 provided in the embodiment of the present application is executed by an electronic terminal device, such as a mobile phone terminal, and the method 700 includes the following steps:

[0247] 701: providing a binding interface for a communication voice identifier, wherein the binding interface displays a user voice communication identifier input area, a transfer communication voice identifier input area, and a preset condition input area, and the user communication voice identifier and the transfer communication voice identifier have a binding relationship.

[0248] 702: In response to the user's input operation, determine the user communication voice identifier and the transfer communication voice identifier in the corresponding input area.

[0249] 703: In response to the user's input operation, determine the preset conditions in the corresponding input area, and forward the communication voice request to the device to which the communication voice transfer identifier belongs under the preset conditions.

[0250] 704: In response to the binding operation, the binding relationship operation is completed.

[0251] It should be noted that since the specific implementation of the above steps 701 to 704 has been described in detail in the embodiment of the method 600 above, they will not be described in detail here.

[0252] In addition, the method 700 further includes: sending the binding relationship to the communication voice switching device.

[0253] For example, a user activates call forwarding in his mobile phone, and the operator of the mobile phone number supports the call forwarding operation. When the user activates the call forwarding operation in the mobile phone, he enters the forwarding number and forwarding conditions of the call forwarding. At the same time, he needs to access the server through the web page or the applet to obtain the setting condition page provided by the server, and enter the call forwarding conditions on the setting condition page, such as entering the local number, the forwarding number, the forwarding conditions and the dialogue strategy, etc. Among them, Figure 8 The interface 800 for setting conditions is shown. In the interface 800, an input box for the local number is set, in which the user enters "138xxxxxxx", an input box for the virtual forwarding number is set, in which the user enters "17xxxxxxx", an input box for the forwarding condition is set, in which the user selects "not in the phone book", and an input box for setting the dialogue strategy is set, in which the user selects "more information", and can also select "sales category: insurance, more information". In addition, after setting the information, the server will receive the setting information, and the user's mobile phone will also start to perform the call forwarding operation.

[0254] After the settings are completed, when the user's mobile phone receives an unfamiliar call, the user's mobile phone will first intercept the unfamiliar call, that is, it will not perform the ringing operation, first obtain the number of the unfamiliar caller, and query the number recorded in the phone book stored locally on the user's mobile phone or the phone book. When the number of the unfamiliar caller does not exist in the phone book, the user's mobile phone will not perform the ringing operation, and directly dial the virtual forwarding number according to the call forwarding setting information. The virtual forwarding number is set on the server side of the operator with the virtual number, thereby completing the call forwarding.

[0255] When the server receives the communication voice request for call transfer, it can accept the request, thereby establishing a call connection between the unknown caller and the server. After the call connection is established, the other party of the unknown caller will start speaking, and the server receives the communication voice sent by the other party and converts the communication voice into voice text through a preset ASR (speech recognition, also known as automatic speech recognition) model, such as "Hello, this is xx insurance" or "Hello, is this Ms. xx?"

[0256] It should be noted that in addition to the user's mobile phone automatically forwarding calls, it can also automatically forward calls when it is inconvenient for the user to answer the phone. For example, if the user is driving and has not answered the phone for a long time, a prompt message can be displayed to the user, allowing the user to choose whether to forward the call independently, and it can also avoid the occurrence of misdirection. Or, when the user has not answered the phone for a long time, the user may not be next to the phone, and the user's mobile phone can directly forward the call, and at the same time make text records and text classification for the user.

[0257] It should also be noted that for other parts not described in detail in this embodiment, reference may be made to the relevant descriptions of the embodiments shown in the above-mentioned method 100 and method 600, and will not be repeated here.

[0258] In order to ensure that users are not aware of automatic voice information recognition, and at the same time be able to quickly and automatically recognize voice information, help users avoid being harassed by too many unfamiliar calls, and reduce resource consumption and occupation, the embodiment of the present application also provides an information recognition system.

[0259] Fig. 9 This is a schematic diagram of a structure of an information recognition system provided by an exemplary embodiment of the present application. Fig. 9 As shown, the identification system 900 may include: a first device 901 and a second device 902 , and the identification system 900 may also include a third device 903 .

[0260] Among them, the first device 901 can be a device with certain computing capabilities, which is mainly responsible for transferring the incoming call request to the second device 902. The basic structure of the first device 901 may include: at least one processor. The number of processors may depend on the configuration and type of the device with certain computing capabilities. The device with certain computing capabilities may also include a memory, which may be volatile, such as RAM, or non-volatile, such as read-only memory (ROM), flash memory, etc., or may include both types. The memory usually stores an operating system (OS), one or more application programs, and may also store program data, etc. In addition to the processing unit and the memory, the device with certain computing capabilities also includes some basic configurations, such as a network card chip, an IO bus, a display component, and some peripheral devices, etc. Optionally, some peripheral devices may include, for example, a keyboard, an input pen, etc. Other peripheral devices are well known in the art and will not be described in detail here. Optionally, the first device 901 may be an intelligent terminal, such as a mobile phone.

[0261] The second device 902 refers to a device that can provide computing and processing services in a network virtual environment, and generally refers to a server that uses the network to identify information. In physical implementation, the second device 902 can be any device that can provide computing services, respond to service requests, and perform processing, such as a conventional server, a cloud server, a cloud host, a virtual center, etc. The second device 902 mainly includes a processor, a hard disk, a memory, a system bus, etc., which is similar to a general computer architecture.

[0262] The third device 903 may be a device with certain computing capabilities. The basic structure of the third device 903 may include: at least one processor. The number of processors may depend on the configuration and type of the device with certain computing capabilities. The device with certain computing capabilities may also include a memory, which may be volatile, such as RAM, or non-volatile, such as read-only memory (ROM), flash memory, etc., or may include both types at the same time. The memory usually stores an operating system (OS), one or more application programs, and may also store program data, etc. In addition to the processing unit and the memory, the device with certain computing capabilities also includes some basic configurations, such as a network card chip, an IO bus, a display component, and some peripheral devices, etc. Optionally, some peripheral devices may include, for example, a keyboard, an input pen, etc. Other peripheral devices are well known in the art and will not be described in detail here. Optionally, the third device 903 may be an intelligent terminal, such as a mobile phone.

[0263] In this embodiment, the first device 901 receives a communication voice request, and forwards the communication voice request to the second device 902 when the source of the communication voice request does not belong to a preset source.

[0264] The second device 902 receives a communication voice request and obtains a voice text corresponding to the voice information; obtains a voice dialogue text corresponding to the voice text according to the voice text; and identifies a category to which the voice dialogue text belongs according to the voice dialogue text.

[0265] The third device 903 sends a communication voice request to the first device 901.

[0266] Specifically, the second device 902 determines the interactive text corresponding to the voice text; takes the acquired voice text and the corresponding interactive text as a group of dialogue texts, and takes at least one group of dialogue texts as voice dialogue text; when there are multiple groups of voice dialogue texts, for any two groups of dialogue texts, the latter group of dialogue texts is obtained according to the former group of dialogue texts.

[0267] Specifically, the second device 902 obtains the voice text in the latter group of dialogue texts according to the interactive text in the former group of dialogue texts; after obtaining the voice text, the step of determining the interactive text corresponding to the voice text is performed.

[0268] Specifically, the second device 902 sends the interactive text via the communication voice, and receives the corresponding communication voice via the communication voice to obtain the voice text.

[0269] Specifically, the second device 902 determines the semantics of the voice text; and determines the interactive text corresponding to the voice text from the dialogue set according to the semantics.

[0270] Specifically, the second device 902 determines the semantics of the voice dialogue text, and determines the category corresponding to the voice dialogue text according to the semantics.

[0271] Specifically, the second device 902 sends the category, the voice conversation text, and the communication identifier corresponding to the communication voice request.

[0272] It should also be noted that for other parts not described in detail in this embodiment, reference may be made to the relevant descriptions of the embodiments shown in the above-mentioned method 100 and method 400 to method 700, and will not be repeated here.

[0273] In addition, in this embodiment, the first device 901 forwards the communication voice request to the second device 902. The second device 902 receives the communication voice request and obtains the voiceprint corresponding to the voice information; when the voiceprint is an abnormal voiceprint, an abnormal voiceprint library is established according to the abnormal voiceprint.

[0274] The first device 901 may also directly report abnormal voiceprints to the second device 902 .

[0275] In the application scenario of the embodiment of the present application, for example, in the application scenario of making a phone call, the mobile phone of the unknown caller initiates a call request to the user's mobile phone. After the user's mobile phone receives the call request, it first intercepts the request, and according to the preset forwarding rules, it searches whether the mobile phone number of the unknown caller belongs to the phone book of the user's mobile phone. If not, it executes the call forwarding function, dials the preset forwarding number, and transfers the call request to the preset forwarding number, such as the virtual operator platform of the server. After receiving the call request, the virtual operator platform of the server can first search whether the mobile phone number of the unknown caller belongs to the blacklist of the server. If it does, it can directly reject the call request, and the blacklist can record the fraudulent phone number. If it does not belong, the server can accept the call request and establish a voice call. The server receives the voice sent by the mobile phone of the unknown caller. The server can also obtain the voiceprint of the voice and compare the voiceprint with the voiceprint library. If the comparison is successful, the server can hang up the phone, and the voiceprint library can record the voiceprint of the fraudster. If the comparison is unsuccessful, the server can convert the speech into speech text, determine the semantics of the speech text, obtain the interactive text corresponding to the semantics from the semantic library, and send the speech text to the unknown caller's mobile phone through the communication voice. The server thus generates a speech dialogue text, which should be understood to be continuously updated. The server determines the dialogue scene in which the dialogue occurs based on the understanding of the semantics of the speech dialogue text, thereby determining the category of the speech dialogue text, and thus determining the intention of the unknown caller's mobile phone.

[0276] When the category belongs to the fraud category, the server can directly hang up the call and send the category situation, the mobile phone number of the unfamiliar caller and the voice conversation text to the user's mobile phone so that the user can know.

[0277] When the category belongs to the type required by the user, such as insurance sales, the server can generate the interactive text required by the user according to the user's needs, send the interactive text through the communication voice, and get the content required by the user, and update the voice dialogue text. The final voice dialogue text, the category situation and the mobile phone number of the unknown caller are sent to the user's mobile phone so that the user can know it.

[0278] In addition, in this embodiment, the second device 902 receives the category and the corresponding voice conversation text, and the second device 902 sends the voice conversation text.

[0279] The first device 901 obtains the voice conversation text; identifies the category to which the voice conversation text belongs based on the voice conversation text, displays the category and the corresponding voice conversation text; and sends the category and the corresponding voice conversation text to the second device 902.

[0280] In this application scenario, it is similar to the application scenario described above, except that the server returns the voice conversation text to the user's mobile phone, and the voice conversation text is continuously updated and sent to the user's mobile phone. After the user's mobile phone receives it, the user's mobile phone identifies the voice conversation text, determines the category, and directly displays the final voice conversation text and the category to the user. It can also obtain the mobile phone number of the unfamiliar caller from the server for display. The voice conversation text and the category are returned to the server so that the server can store them. At the same time, the user's mobile phone can directly display the final voice conversation text, the category, and the mobile phone number of the unfamiliar caller to the user.

[0281] In the above embodiment, the first device 901 can establish a network connection with the second device 902 and the third device 903, and the network connection can be a wireless connection. If the first device 901 is connected to the second device 902 and the third device 903 for communication, the network standard of the mobile network can be any one of 2G (GSM), 2.5G (GPRS), 3G (WCDMA, TD-SCDMA, CDMA2000, UTMS), 4G (LTE), 4G+ (LTE+), WiMax, 5G, etc.

[0282] In order to better improve user services and reduce the possibility of information being attacked and leaked, an embodiment of the present application also provides an information identification method.

[0283] Fig.10 This is a flow chart of an information identification method of an exemplary embodiment of the present application. The method 1000 provided in the embodiment of the present application is executed by an electronic terminal device, such as a mobile phone terminal, and the method 700 includes the following steps:

[0284] 1001: When a communication voice request is received, and the source of the communication voice request is not a preset source, the communication voice is received through the smart speaker, and the reception of the user voice and the provision of the communication voice to the user are blocked.

[0285] 1002: Obtain the voice text corresponding to the communication voice through the smart speaker.

[0286] 1003: Obtain a voice dialogue text corresponding to the voice text through the smart speaker according to the voice text.

[0287] 1004: Identify the category to which the voice conversation text belongs based on the voice conversation text through the smart speaker.

[0288] It should be noted that since the specific implementation methods of the above steps 1001 to 1004 have been explained in detail in the embodiment of the method 1000 above, they will not be described in detail here.

[0289] It is only explained that a smart speaker can be set in the execution device terminal, and the smart speaker can execute steps 1001-1004, and the specific execution method is similar to method 100, which will not be repeated here. After receiving the request, the smart speaker will temporarily turn off the user terminal, such as the function of playing the sound of the mobile phone and the function of collecting the user's voice, such as pausing the handset, the speaker and the microphone.

[0290] In addition, before using the smart speaker, the smart speaker needs to be set up. The method 100 also includes: in response to the user's power-on and authorization operations, turning on the smart speaker, and enabling the smart speaker to receive communication voice when the source of the communication voice request does not belong to the preset source.

[0291] For example, a setting interface is provided to the user, and the user turns on the button to use the smart speaker according to the prompt information on the interface, that is, the smart speaker is started, and the smart speaker is authorized so that the smart speaker has the right to obtain communication voice and implement this solution. In addition, the service provided by the smart speaker may incur communication fees. When the user turns on the smart speaker, it is deemed that the communication fee is agreed to be charged. Otherwise, the smart speaker will not be turned on.

[0292] In addition, the method 1000 also includes: providing prompt information for the user to receive communication voice; in response to the user's receiving operation, canceling the shielding action, receiving the user's voice and shielding the communication voice provided to the user.

[0293] When the smart speaker is working, it can continuously provide prompt information to the user, asking the user whether to answer the call by voice. When the user agrees to answer the call, the smart speaker turns on the receiver and microphone functions, allowing the user to talk to the caller, thereby improving the user experience.

[0294] In addition, when the smart speaker cannot accurately answer the caller's call, it can display the question to the user in text form, prompting the user whether to answer the caller in person. When the user agrees, the smart speaker turns on the receiver and microphone functions, allowing the user to talk to the caller. The smart speaker stops the call. After the user receives the call, if the user hangs up the phone directly, the smart speaker will no longer execute. When the user simply stops talking to the caller, the smart speaker continues the call.

[0295] Fig.11 A schematic diagram of the structural framework of an information identification device provided in an exemplary embodiment of the present application. The device 1100 can be applied to an electronic terminal device, such as a mobile phone, a tablet computer, and a desktop computer. The device 1100 includes an acquisition module 1101 and an identification module 1102. The functions of each module are described in detail below:

[0296] The acquisition module 1101 is used to receive a communication voice request, and when the source of the communication voice request does not belong to a preset source, acquire the voice text corresponding to the communication voice.

[0297] The acquisition module 1101 is used to acquire the voice dialogue text corresponding to the voice text according to the voice text.

[0298] The recognition module 1102 is used to recognize the category to which the speech dialogue text belongs based on the speech dialogue text.

[0299] In addition, the acquisition module 1101 is also used to obtain the source of the communication voice request and match the source with the user's preset normal source set; the device 1100 also includes: a providing module, which is used to provide the communication voice to the user when the match is successful.

[0300] The acquisition module 1101 is also used to obtain the source of the communication voice request and match the source with a preset abnormal source set; the providing module is also used to provide the user with prompt information on whether to accept the communication voice when the match is successful.

[0301] The acquisition module 1101 is also used to obtain the source of the communication voice request and match the source with a preset abnormal source set; the device 1100 also includes: a rejection module, which is used to reject the communication voice request when an illegal source is matched.

[0302] The acquisition module 1101 is also used to acquire the voiceprint of the communication voice when the source of the communication voice request does not belong to the preset source; and to execute the step of acquiring the voice text corresponding to the communication voice when the voiceprint does not belong to the abnormal voiceprint library.

[0303] In addition, the device 1100 further includes: a building module, which is used to build an abnormal voiceprint library, where the abnormal voiceprint library is obtained based on the abnormal voiceprint.

[0304] In addition, the device 1100 further includes: an updating module, which is used to update the voiceprint as an abnormal voiceprint to an abnormal voiceprint library when the category is an abnormal category.

[0305] In addition, the device 1100 further includes: a determination module, which is used to match the voiceprint with multiple voiceprints in the abnormal voiceprint library, and when the match is successful, determine that the voiceprint is an abnormal voiceprint.

[0306] The determination module is also used to determine that the corresponding communication voice is an illegal communication voice when the matched voiceprint belongs to an illegal voiceprint among the abnormal voiceprints, and terminate the reception of the communication voice.

[0307] The updating module is further used to receive the reported voiceprint and the reported category. When the reported category is an abnormal category, the reported voiceprint is updated as an abnormal voiceprint to the abnormal voiceprint library.

[0308] Before updating the abnormal voiceprint library, the updating module is used to update the voiceprint as an abnormal voiceprint to the abnormal voiceprint library when it is determined that the voiceprint does not exist in the abnormal voiceprint library.

[0309] The determination module is further used to determine that the communication voice is an abnormal communication voice when the category is an abnormal category.

[0310] The determination module is also used to determine that the communication voice is an illegal communication voice in the abnormal communication voice and the voiceprint is an illegal voiceprint in the abnormal voiceprint when the category is an illegal category in the abnormal category, and terminate the reception of the communication voice.

[0311] Specifically, the acquisition module 1101 is used to directly use the acquired voice text as the voice dialogue text.

[0312] Specifically, the acquisition module 1101 includes: a determination unit, used to determine the interactive text corresponding to the voice text; taking the acquired voice text and the corresponding interactive text as a group of dialogue texts, and taking at least one group of dialogue texts as voice dialogue text; an acquisition unit, used to obtain the latter group of dialogue texts according to the former group of dialogue texts for any two groups of dialogue texts when there are multiple groups of voice dialogue texts.

[0313] Specifically, the acquisition unit is used to acquire the voice text in the latter group of dialogue texts according to the interactive text in the former group of dialogue texts; after acquiring the voice text, the step of determining the interactive text corresponding to the voice text is executed.

[0314] Specifically, the acquisition unit is used to send the interactive text via communication voice and receive the corresponding communication voice to acquire the voice text.

[0315] Specifically, the determination unit is used to determine the semantics of the voice text; based on the semantics, determine the interactive text corresponding to the voice text from the dialogue set.

[0316] Specifically, the recognition module 1102 is used to determine the semantics of the voice dialogue text, and determine the category corresponding to the voice dialogue text according to the semantics.

[0317] The acquisition module 1101 is further used to acquire the voice dialogue text matching the content according to the content of the category required by the user when the category is the category required by the user. The category required by the user may belong to an abnormal category.

[0318] In addition, the determination module is also used to determine that the voiceprint is a voiceprint of other categories when the category is other categories.

[0319] In addition, the device 1100 further includes: a setting module, which is used to set other types of voiceprints to the corresponding type of voiceprint library.

[0320] In addition, the determination module is further used to determine the category of the source of the communication voice request according to the category.

[0321] In addition, the setting module is also used to set the source of the communication voice request to the corresponding source category set.

[0322] In addition, a module is provided for providing the category and the source of the communication voice request to the user.

[0323] In addition, a module is provided for providing the user with a text of the voice conversation.

[0324] In addition, the device 1100 also includes: a receiving module for receiving adjustment information of the category; an adjustment module for adjusting the category according to the adjustment information; using the adjusted category and the corresponding voice dialogue text as samples for optimizing the model; wherein the model is used to determine the semantics of the voice dialogue text, and determine the category corresponding to the voice dialogue text according to the semantics.

[0325] After obtaining the voice dialogue text corresponding to the voice text, a module is provided, which is also used to provide the user with a viewing address of the voice dialogue text.

[0326] In addition, the device 1100 further includes: a recording module for recording the conversation voice to which the voice conversation text belongs; and a providing module for providing the user with a listening address for the conversation voice to which the voice conversation text belongs.

[0327] In addition, the device 1100 further includes: a shielding module, which is used to accept a communication voice request and shield the reception of user voice and shield the provision of communication voice to the user when the source of the communication voice request does not belong to a preset source.

[0328] In addition, a module is provided, which is also used to provide the source of the communication voice request to the user when the source of the communication voice request belongs to a preset source.

[0329] Fig.12 A schematic diagram of a structural framework of an information identification device provided as another exemplary embodiment of the present application. The device 1200 can be applied to an electronic device, such as a conventional server; the device 1200 includes: an acquisition module 1201 and an identification module 1202. The functions of each module are described in detail below:

[0330] The acquisition module 1201 is used to receive the forwarded communication voice request and acquire the voice text corresponding to the voice information.

[0331] The recognition module 1202 is used to obtain the speech dialogue text corresponding to the speech text according to the speech text; and to recognize the category to which the speech dialogue text belongs according to the speech dialogue text.

[0332] Specifically, the acquisition module 1201 is used to directly use the acquired voice text as the voice dialogue text.

[0333] Specifically, the acquisition module 1201 includes: a determination unit, used to determine the interactive text corresponding to the voice text; taking the acquired voice text and the corresponding interactive text as a group of dialogue texts, and taking at least one group of dialogue texts as voice dialogue text; an acquisition unit, used to obtain the latter group of dialogue texts according to the former group of dialogue texts for any two groups of dialogue texts when there are multiple groups of voice dialogue texts.

[0334] Specifically, the acquisition unit is used to acquire the voice text in the latter group of dialogue texts according to the interactive text in the former group of dialogue texts; after acquiring the voice text, the step of determining the interactive text corresponding to the voice text is executed.

[0335] Specifically, the acquisition unit is used to send the interactive text via communication voice and receive the corresponding communication voice to acquire the voice text.

[0336] Specifically, the determination unit is used to determine the semantics of the voice text; based on the semantics, determine the interactive text corresponding to the voice text from the dialogue set.

[0337] Specifically, the recognition module 1202 is used to determine the semantics of the speech dialogue text, and determine the category corresponding to the speech dialogue text according to the semantics.

[0338] In addition, the device 1200 also includes: a storage module, which is used to receive the source of the reported communication voice request and the reporting category, and store the source of the communication voice request into a corresponding source category set according to the reporting type.

[0339] The storage module is also used to store the source of the reported communication voice request into the corresponding source category set according to the number of reports of the source for any reporting category.

[0340] The storage module is used to store the source of the reported communication voice request into the corresponding source category set according to the number of reports of the source and the number of reporting users for any reporting category.

[0341] In addition, the apparatus 1200 further includes: a sending module, which is used to send the updated source category set to the corresponding device for updating.

[0342] It should be noted that, for the parts not described in detail in this embodiment, reference may be made to the relevant description of the embodiment shown in the above identification device, and will not be repeated here.

[0343] Fig.13 A schematic diagram of the structural framework of a device for establishing a voiceprint library provided by another exemplary embodiment of the present application. The device 1300 can be applied to electronic devices, such as conventional servers or mobile phones; the device 1300 includes: an establishment module 1301, and the functions of each module are described in detail below:

[0344] The establishing module 1301 is used to establish an abnormal voiceprint library, where the abnormal voiceprint library is obtained based on the abnormal voiceprints.

[0345] In addition, the device 1300 includes: a determination module for receiving communication voice and determining a voiceprint corresponding to the communication voice; an update module for updating the voiceprint as an abnormal voiceprint to an abnormal voiceprint library when the communication voice is determined to be an abnormal communication voice.

[0346] In addition, the determination module is also used to receive the communication voice and determine the voiceprint corresponding to the communication voice; match the voiceprint with multiple voiceprints in the abnormal voiceprint library, and when the match is successful, determine that the voiceprint is an abnormal voiceprint.

[0347] In addition, the determination module is also used to determine that the corresponding communication voice is an illegal communication voice when the matched voiceprint belongs to an illegal voiceprint among the abnormal voiceprints, and terminate the reception of the communication voice.

[0348] In addition, the device 1300 includes: a receiving module, which is used to receive the reported voiceprint and the reporting category, and when the reporting category is an abnormal category, the reported voiceprint is updated as an abnormal voiceprint to the abnormal voiceprint library.

[0349] Before updating to the abnormal voiceprint library, the updating module is further used to update the voiceprint as an abnormal voiceprint to the abnormal voiceprint library when it is determined that the voiceprint does not exist in the abnormal voiceprint library.

[0350] In addition, the device 1300 includes: an acquisition module, which is used to acquire the voice text corresponding to the communication voice when the voiceprint does not exist in the abnormal voiceprint library; acquire the voice conversation text corresponding to the voice text based on the voice text; and an identification module, which is used to identify the category to which the voice conversation text belongs based on the voice conversation text.

[0351] It should be noted that, for the parts not described in detail in this embodiment, reference may be made to the relevant description of the embodiment shown in the above identification device, and will not be repeated here.

[0352] Fig.14A schematic diagram of a structural framework of an information identification device provided as another exemplary embodiment of the present application. The device 1400 can be applied to an electronic device, such as a terminal; the device 1400 includes: a shielding module 1401, an acquisition module 1402, and an identification module 1403. The functions of each module are described in detail below:

[0353] The shielding module 1401 is used to receive the communication voice through the smart speaker when a communication voice request is received and the source of the communication voice request does not belong to the preset source, and shield the reception of the user voice and shield the provision of communication voice to the user.

[0354] The acquisition module 1402 is used to acquire the voice text corresponding to the communication voice through the smart speaker.

[0355] The acquisition module 1402 is used to acquire the voice dialogue text corresponding to the voice text through the smart speaker according to the voice text.

[0356] The recognition module 1403 is used to recognize the category to which the voice conversation text belongs based on the voice conversation text through the smart speaker.

[0357] In addition, the device 1400 also includes: an activation module, which is used to activate the smart speaker in response to the user's activation and authorization operations, and enable the smart speaker to receive communication voice when the source of the communication voice request does not belong to the preset source.

[0358] In addition, the device 1400 also includes: a providing module, which is used to provide prompt information for the user to receive communication voice; and a revocation module, which is used to respond to the user's receiving operation, cancel the shielding action, receive the user's voice, and shield the communication voice provided to the user.

[0359] It should be noted that, for the parts not described in detail in this embodiment, reference may be made to the relevant description of the embodiment shown in the above identification device, and will not be repeated here.

[0360] Fig.15 A schematic diagram of a structural framework of an information setting device provided as another exemplary embodiment of the present application. The device 1500 can be applied to an electronic device, such as a terminal; the device 1500 includes: an acquisition module 1501, a binding module 1502, and a setting module 1503. The functions of each module are described in detail below:

[0361] The acquisition module 1501 is used to acquire the user communication voice identifier.

[0362] The binding module 1502 is used to bind the user communication voice identifier with the transfer communication voice identifier, so that under preset conditions and according to the binding relationship, the communication voice request is forwarded to the device to which the transfer communication voice identifier belongs.

[0363] The setting module 1503 is used to set the preset conditions.

[0364] The setting module 1503 is further used to set a processing method for the corresponding communication voice under preset conditions.

[0365] In addition, the device 1500 also includes: a receiving module, which is used to receive a communication voice request and obtain a communication voice identifier corresponding to the communication voice request; a forwarding module, which is used to forward the communication voice request to the device to which the communication voice identifier belongs according to the binding relationship when the acquired voice communication identifier meets the preset conditions.

[0366] The acquisition module 1501 is further used to acquire the voiceprint of the communication voice and report the voiceprint and the type of the voiceprint.

[0367] In addition, the device 1500 further includes: a matching module, which is used to match the voiceprint with the reported voiceprint in the local voiceprint library, and when the corresponding voiceprint is not matched, the step of reporting the voiceprint and the category of the voiceprint is executed.

[0368] The acquisition module 1501 is further used to acquire the communication voice identifier and report the communication voice identifier and the category of the communication voice identifier.

[0369] The receiving module is also used to receive and display the communication voice category corresponding to the communication voice request and the voice dialogue text corresponding to the communication voice.

[0370] The receiving module is also used to receive the voice dialogue text corresponding to the communication voice request; the device 1500 also includes: a determination module, which is used to determine the category of the communication voice in response to the user's input operation; and a reporting module, which is used to report the category and the corresponding voice dialogue text.

[0371] It should be noted that, for the parts not described in detail in this embodiment, reference may be made to the relevant description of the embodiment shown in the above identification device, and will not be repeated here.

[0372] Fig.16 A schematic diagram of a structural framework of an information setting device provided for another exemplary embodiment of the present application. The device 1600 can be applied to an electronic device, such as a terminal; the device 1600 includes: a providing module 1601, a determining module 1602, and a binding module 1603. The functions of each module are described in detail below:

[0373] A module 1601 is provided for providing a binding interface for a communication voice identifier. The binding interface displays a user communication voice identifier input area, a transfer communication voice identifier input area, and a preset condition input area. The user communication voice identifier and the transfer communication voice identifier have a binding relationship.

[0374] The determination module 1602 is used to determine the user communication voice identifier and the voice transfer communication identifier in the corresponding input area in response to the user's input operation.

[0375] The determination module 1602 is used to determine the preset conditions in the corresponding input area in response to the user's input operation, and forward the communication voice request to the device to which the communication voice transfer identifier belongs under the preset conditions.

[0376] The binding module 1603 is used to complete the operation of the binding relationship in response to the binding operation.

[0377] In addition, the apparatus 1600 further includes: a sending module, configured to send the binding relationship to the communication voice switching device.

[0378] It should be noted that, for the parts not described in detail in this embodiment, reference may be made to the relevant description of the embodiment shown in the above identification device, and will not be repeated here.

[0379] The above describes Fig.11 The internal functions and structure of the identification device 1100 shown in FIG. 1100 are shown in FIG. 1100 . In one possible design, Fig.11 The structure of the identification device 1100 shown can be implemented as a terminal device, such as Fig.17 As shown, the device 1700 may include: a memory 1701, a processor 1702, and a communication component 1703;

[0380] Memory 1701, used for storing computer programs;

[0381] The communication component 1703 is used to receive a communication voice request.

[0382] Processor 1702 is used to execute a computer program to: when the source of the communication voice request does not belong to a preset source, obtain the voice text corresponding to the communication voice; obtain the voice dialogue text corresponding to the voice text based on the voice text; and identify the category to which the voice dialogue text belongs based on the voice dialogue text.

[0383] In addition, the processor 1702 is also used to obtain the source of the communication voice request and match the source with the normal source set preset by the user; if the match is successful, the communication voice is provided to the user.

[0384] In addition, the processor 1702 is also used to obtain the source of the communication voice request and match the source with a preset abnormal source set; if the match is successful, a prompt message of whether to accept the communication voice is provided to the user.

[0385] In addition, the processor 1702 is also used to obtain the source of the communication voice request and match the source with a preset abnormal source set; if an illegal source is matched, the communication voice request is rejected.

[0386] In addition, the processor 1702 is also used to obtain the voiceprint of the communication voice when the source of the communication voice request does not belong to the preset source; and when the voiceprint does not belong to the abnormal voiceprint library, execute the step of obtaining the voice text corresponding to the communication voice.

[0387] In addition, the processor 1702 is also used to establish an abnormal voiceprint library, where the abnormal voiceprint library is obtained based on the abnormal voiceprint.

[0388] In addition, the processor 1702 is further configured to, when the category is an abnormal category, update the voiceprint as an abnormal voiceprint to an abnormal voiceprint library.

[0389] In addition, the processor 1702 is further configured to match the voiceprint with multiple voiceprints in the abnormal voiceprint library, and when the match is successful, determine that the voiceprint is an abnormal voiceprint.

[0390] In addition, the processor 1702 is further configured to, when the matched voiceprint is an illegal voiceprint among the abnormal voiceprints, determine that the corresponding communication voice is an illegal communication voice, and terminate receiving the communication voice.

[0391] In addition, the processor 1702 is further configured to receive the reported voiceprint and the reported category, and when the reported category is an abnormal category, update the reported voiceprint as an abnormal voiceprint to the abnormal voiceprint library.

[0392] Before updating the abnormal voiceprint library, in addition, the processor 1702 is further configured to update the voiceprint as an abnormal voiceprint to the abnormal voiceprint library when it is determined that the voiceprint does not exist in the abnormal voiceprint library.

[0393] In addition, the processor 1702 is further configured to, when the category is an abnormal category, determine that the communication voice is an abnormal communication voice.

[0394] In addition, the processor 1702 is also used to, when the category is an illegal category in the abnormal category, determine that the communication voice is an illegal communication voice in the abnormal communication voice and the voiceprint is an illegal voiceprint in the abnormal voiceprint, and terminate receiving the communication voice.

[0395] Specifically, the processor 1702 is specifically configured to directly use the acquired voice text as the voice dialogue text.

[0396] Specifically, processor 1702 is specifically used to determine the interactive text corresponding to the voice text; use the acquired voice text and the corresponding interactive text as a group of dialogue texts, and use at least one group of dialogue texts as voice dialogue text; when there are multiple groups of voice dialogue texts, for any two groups of dialogue texts, obtain the latter group of dialogue texts based on the former group of dialogue texts.

[0397] Specifically, processor 1702 is specifically used to obtain voice text in a later group of dialogue texts based on interactive text in a previous group of dialogue texts; after obtaining the voice text, execute the step of determining the interactive text corresponding to the voice text.

[0398] Specifically, the processor 1702 is specifically configured to send the interactive text via the communication voice and receive the corresponding communication voice to obtain the voice text.

[0399] Specifically, the processor 1702 is specifically configured to determine the semantics of the voice text; and determine the interactive text corresponding to the voice text from the dialogue set according to the semantics.

[0400] Specifically, the processor 1702 is specifically configured to determine the semantics of the voice dialogue text, and determine the category corresponding to the voice dialogue text according to the semantics.

[0401] In addition, the processor 1702 is also used to, when the category is the category required by the user, obtain the voice dialogue text matching the content according to the content of the category required by the user, and the category required by the user may belong to an abnormal category.

[0402] In addition, the processor 1702 is further configured to, when the category is other categories, determine that the voiceprint is a voiceprint of other categories.

[0403] In addition, the processor 1702 is also used to set other types of voiceprints to the corresponding types of voiceprint libraries.

[0404] In addition, the processor 1702 is further configured to determine the category of the source of the communication voice request according to the category.

[0405] In addition, the processor 1702 is further configured to set the source of the communication voice request to a corresponding source category set.

[0406] In addition, the processor 1702 is further configured to provide the user with a category and a source of the communication voice request.

[0407] In addition, the processor 1702 is further configured to provide the user with a voice conversation text.

[0408] In addition, processor 1702 is also used to receive category adjustment information; adjust the category according to the adjustment information; and use the adjusted category and the corresponding voice dialogue text as samples for optimizing the model; wherein the model is used to determine the semantics of the voice dialogue text, and determine the category corresponding to the voice dialogue text according to the semantics.

[0409] After acquiring the voice dialogue text corresponding to the voice text, the processor 1702 is further configured to provide the user with a viewing address for the voice dialogue text.

[0410] In addition, the processor 1702 is further configured to record the conversation voice to which the voice conversation text belongs, and provide the user with a listening address for the conversation voice to which the voice conversation text belongs.

[0411] In addition, the processor 1702 is also used to accept the communication voice request and block receiving the user voice and blocking providing the communication voice to the user when the source of the communication voice request does not belong to the preset source.

[0412] In addition, the processor 1702 is further configured to provide the source of the communication voice request to the user when the source of the communication voice request belongs to a preset source.

[0413] In addition, an embodiment of the present invention provides a computer storage medium, when the computer program is executed by one or more processors, causing the one or more processors to implement Figure 1-Figure 3 The steps of the information identification method in the method embodiment.

[0414] The above describes Fig.12 The internal functions and structure of the identification device 1200 shown in FIG. 1 are shown in FIG. 1 . In one possible design, Fig.12 The structure of the identification device 1200 shown can be implemented as a server, such as Fig.18 As shown, the device 1800 may include: a memory 1801, a processor 1802;

[0415] Memory 1801, used for storing computer programs;

[0416] Processor 1802 is used to execute a computer program to: accept a forwarded communication voice request and obtain a voice text corresponding to the voice information; obtain a voice dialogue text corresponding to the voice text according to the voice text; and identify a category to which the voice dialogue text belongs according to the voice dialogue text.

[0417] Specifically, the processor 1802 is specifically configured to: directly use the acquired voice text as the voice dialogue text.

[0418] Specifically, processor 1802 is specifically used to: determine the interactive text corresponding to the voice text; use the acquired voice text and the corresponding interactive text as a group of dialogue texts, and use at least one group of dialogue texts as voice dialogue texts; when there are multiple groups of voice dialogue texts, for any two groups of dialogue texts, obtain the latter group of dialogue texts based on the former group of dialogue texts.

[0419] Specifically, processor 1802 is specifically used to: obtain voice text in a later group of dialogue texts according to the interactive text in a previous group of dialogue texts; after obtaining the voice text, execute the step of determining the interactive text corresponding to the voice text.

[0420] Specifically, the processor 1802 is specifically configured to: send the interactive text via communication voice, and receive the corresponding communication voice to obtain the voice text.

[0421] Specifically, processor 1802 is specifically configured to: determine the semantics of the voice text; and determine, based on the semantics, an interactive text corresponding to the voice text from a dialog set.

[0422] Specifically, the processor 1802 is specifically configured to: determine the semantics of the voice dialogue text, and determine the category corresponding to the voice dialogue text according to the semantics.

[0423] In addition, the processor 1802 is further configured to: receive the source of the reported communication voice request and the reporting category, and store the source of the communication voice request in a corresponding source category set according to the reporting type.

[0424] In addition, the processor 1802 is further configured to: for any reporting category, store the source of the reported communication voice request into a corresponding source category set according to the number of reports of the source.

[0425] In addition, the processor 1802 is further configured to: for any reporting category, store the source of the reported communication voice request into a corresponding source category set according to the number of reports of the source and the number of reporting users.

[0426] In addition, the processor 1802 is further configured to send the updated source category set to the corresponding device for updating.

[0427] It should be noted that, for the parts not described in detail in this embodiment, reference may be made to the relevant description of the embodiment shown in the above-mentioned electronic device 1700, and will not be repeated here.

[0428] In addition, an embodiment of the present invention provides a computer storage medium, when the computer program is executed by one or more processors, causing the one or more processors to implement Figure 5 The steps of the information identification method in the method embodiment.

[0429] The above describes Fig.13 The internal functions and structure of the building device 1300 shown, in one possible design, Fig.13 The structure of the establishment device 1300 shown can be implemented as a server, such as Fig.19 As shown, the device 1900 may include: a memory 1901, a processor 1902;

[0430] Memory 1901, used for storing computer programs;

[0431] Processor 1902 is used to execute a computer program to establish an abnormal voiceprint library, where the abnormal voiceprint library is obtained based on the abnormal voiceprint.

[0432] In addition, the device 1900 may include: a communication component for receiving communication voice. In addition, the processor 1902 is also used to determine the voiceprint corresponding to the communication voice; an update module is used to update the voiceprint as an abnormal voiceprint to an abnormal voiceprint library when the communication voice is determined to be an abnormal communication voice.

[0433] In addition, the communication component is also used to receive communication voice. In addition, the processor 1902 is also used to determine the voiceprint corresponding to the communication voice; match the voiceprint with multiple voiceprints in the abnormal voiceprint library, and when the match is successful, determine that the voiceprint is an abnormal voiceprint.

[0434] In addition, the processor 1902 is further configured to determine that the corresponding communication voice is an illegal communication voice when the matched voiceprint is an illegal voiceprint among the abnormal voiceprints, and terminate the reception of the communication voice.

[0435] In addition, the processor 1902 is further configured to receive the reported voiceprint and the reported category, and when the reported category is an abnormal category, update the reported voiceprint as an abnormal voiceprint to the abnormal voiceprint library.

[0436] Before updating to the abnormal voiceprint library, in addition, the processor 1902 is further configured to update the voiceprint as an abnormal voiceprint to the abnormal voiceprint library when it is determined that the voiceprint does not exist in the abnormal voiceprint library.

[0437] In addition, the processor 1902 is also used to obtain the voice text corresponding to the communication voice when the voiceprint does not exist in the abnormal voiceprint library; obtain the voice conversation text corresponding to the voice text according to the voice text; and the recognition module is used to identify the category to which the voice conversation text belongs according to the voice conversation text.

[0438] It should be noted that, for the parts not described in detail in this embodiment, reference may be made to the relevant description of the embodiment shown in the above-mentioned electronic device 1700, and will not be repeated here.

[0439] In addition, an embodiment of the present invention provides a computer storage medium, when the computer program is executed by one or more processors, causing the one or more processors to implement Figure 4 The steps of the method for establishing a voiceprint library in the method embodiment.

[0440] The above describes Fig.14 The internal functions and structure of the identification device 1400 shown in FIG. 1400 are shown in FIG. 1400 . In one possible design, Fig.14 The structure of the identification device 1400 shown can be implemented as a server, such as Fig. 20 As shown, the device 2000 may include: a memory 2001, a processor 2002;

[0441] Memory 2001, used for storing computer programs;

[0442] Processor 2002 is used to execute a computer program for: when a communication voice request is received and the source of the communication voice request is not a preset source, receiving the communication voice through a smart speaker, and shielding the reception of user voice and shielding the provision of communication voice to the user; obtaining a voice text corresponding to the communication voice through the smart speaker; obtaining a voice dialogue text corresponding to the voice text based on the voice text through the smart speaker; and identifying a category to which the voice dialogue text belongs based on the voice dialogue text through the smart speaker.

[0443] In addition, processor 2002 is also used to turn on the smart speaker in response to the user's power-on and authorization operations, and enable the smart speaker to receive communication voice when the source of the communication voice request does not belong to the preset source.

[0444] In addition, the processor 2002 is also used to provide prompt information for the user to receive communication voice; the revocation module is used to respond to the user's receiving operation, cancel the shielding action, receive the user's voice and shield the communication voice provided to the user.

[0445] It should be noted that, for the parts not described in detail in this embodiment, reference may be made to the relevant description of the embodiment shown in the above-mentioned electronic device 1700, and will not be repeated here.

[0446] In addition, an embodiment of the present invention provides a computer storage medium, when the computer program is executed by one or more processors, causing the one or more processors to implement Fig.10 The steps of the information identification method in the method embodiment.

[0447] The above describes Fig.15 The internal functions and structure of the setting device 1500 shown in FIG. 1 are shown in FIG. 1 , and in one possible design, Fig.15 The structure of the setting device 1500 shown can be implemented as an intelligent terminal device, such as a mobile phone, such as Fig.21As shown, the device 2100 may include: a memory 2101 and a processor 2102;

[0448] Memory 2101, used for storing computer programs;

[0449] Processor 2102 is used to execute a computer program to: obtain a user communication voice identifier; bind the user communication voice identifier with the transferred communication voice identifier so that under preset conditions, according to the binding relationship, the communication voice request is forwarded to the device to which the transferred communication voice identifier belongs; and set preset conditions.

[0450] In addition, the processor 2102 is also used to set a processing method for the corresponding communication voice under preset conditions.

[0451] In addition, the processor 2102 is also used to receive a communication voice request and obtain a communication voice identifier corresponding to the communication voice request; when the obtained voice communication identifier meets the preset conditions, the communication voice request is forwarded to the device to which the communication voice identifier belongs according to the binding relationship.

[0452] In addition, the processor 2102 is also used to obtain the voiceprint of the communication voice and report the voiceprint and the type of the voiceprint.

[0453] In addition, the processor 2102 is further configured to match the voiceprint with the reported voiceprints in the local voiceprint database, and when no corresponding voiceprint is matched, perform the step of reporting the voiceprint and the category of the voiceprint.

[0454] In addition, the processor 2102 is further configured to obtain a communication voice identifier and report the communication voice identifier and a category of the communication voice identifier.

[0455] In addition, the processor 2102 is further configured to receive and display a communication voice category corresponding to the communication voice request and a voice dialogue text corresponding to the communication voice.

[0456] In addition, the processor 2102 is further configured to receive a communication voice request and a voice dialogue text corresponding to the communication voice; determine the category of the communication voice in response to an input operation of the user; and report the category and the corresponding voice dialogue text.

[0457] It should be noted that, for the parts not described in detail in this embodiment, reference may be made to the relevant description of the embodiment shown in the above-mentioned electronic device 1700, and will not be repeated here.

[0458] In addition, an embodiment of the present invention provides a computer storage medium, when the computer program is executed by one or more processors, causing the one or more processors to implement Figure 6 Steps of the method for setting information in the method embodiment.

[0459] The above describes Fig.16 The internal functions and structure of the setting device 1600 shown in FIG. 1 are shown in FIG. 1 , in one possible design. Fig.16 The structure of the setting device 1600 shown can be implemented as an intelligent terminal device, such as a mobile phone, such as Fig. 22 As shown, the device 2200 may include: a memory 2201 and a processor 2202;

[0460] Memory 2201, used for storing computer programs;

[0461] Processor 2202 is used to execute a computer program to: provide a binding interface for a communication voice identifier, in which a user communication voice identifier input area, a transfer communication voice identifier input area and a preset condition input area are displayed, and the user communication voice identifier and the transfer communication voice identifier have a binding relationship; in response to a user's input operation, determine the user communication voice identifier and the voice transfer communication identifier in the corresponding input area; in response to a user's input operation, determine the preset conditions in the corresponding input area, and under the preset conditions, forward the communication voice request to the device to which the transfer communication voice identifier belongs; in response to the binding operation, complete the operation of the binding relationship.

[0462] In addition, the processor 2202 is also used to send the binding relationship to the communication voice switching device.

[0463] It should be noted that, for the parts not described in detail in this embodiment, reference may be made to the relevant description of the embodiment shown in the above-mentioned electronic device 1700, and will not be repeated here.

[0464] In addition, an embodiment of the present invention provides a computer storage medium, when the computer program is executed by one or more processors, causing the one or more processors to implement Figure 7 Steps of the method for setting information in the method embodiment.

[0465] In addition, in some of the processes described in the above embodiments and the accompanying drawings, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this article or executed in parallel, and the sequence numbers of the operations, such as 201, 202, 203, etc., are only used to distinguish between different operations, and the sequence numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this article are used to distinguish different messages, devices, modules, etc., do not represent the order of precedence, and do not limit the "first" and "second" to be different types.

[0466] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0467] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by adding a necessary general hardware platform, and of course can also be implemented by combining hardware and software. Based on such an understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a computer product, and the present invention can be in the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0468] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable multimedia data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable multimedia data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0469] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable multimedia data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0470] These computer program instructions may also be loaded onto a computer or other programmable multimedia data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for executing on the computer or other programmable device to implement the process. Figure 1A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0471] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0472] The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0473] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0474] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for identifying information, characterized in that: include: receiving a communication voice request, and obtaining a voiceprint of the communication voice of the caller when the source of the communication voice request does not belong to a preset source; If the voiceprint does not belong to the abnormal voiceprint library, obtaining the voice text corresponding to the communication voice of the caller; Determine the semantics of the speech text, search for interactive text corresponding to the semantics of the speech text from the dialogue set, and the speech text and its corresponding interactive text are used as speech dialogue text; Using a speech synthesis model, synthesizing the interactive text into speech, and sending the synthesized speech to the calling party; In the conversation, the speech conversation text is processed using a natural language understanding model to obtain the category to which the speech conversation text belongs; When the category to which the voice conversation text belongs is an abnormal category, updating the voiceprint as an abnormal voiceprint to the abnormal voiceprint library; The method further comprises: providing the voice conversation text and the category to which the voice conversation text belongs to the user; Receive adjustment information of the category, and use the adjusted category corresponding to the adjustment information and the corresponding voice dialogue text as samples for optimizing a natural language understanding model for obtaining the category to which the voice dialogue text belongs.

2. The method according to claim 1, characterized in that: The method further comprises: Obtaining the source of the communication voice request, and matching the source with a user preset normal source set; If the match is successful, the communication voice is provided to the user.

3. The method according to claim 1, characterized in that: The method further comprises: Obtaining a source of the communication voice request, and matching the source with a preset abnormal source set; If the match is successful, a prompt message is provided to the user indicating whether to accept the communication voice.

4. The method according to claim 1, characterized in that: The method further comprises: Obtaining a source of the communication voice request, and matching the source with a preset abnormal source set; In the event that an illegal source is matched, the communication voice request is rejected.

5. The method according to claim 1, characterized in that: The method further comprises: An abnormal voiceprint database is established, wherein the abnormal voiceprint database is obtained based on the abnormal voiceprint.

6. The method according to claim 1, characterized in that: The method further comprises: The voiceprint is matched with multiple voiceprints in the abnormal voiceprint library, and when the match is successful, the voiceprint is determined to be an abnormal voiceprint.

7. The method according to claim 6, characterized in that: The method further comprises: When the matched voiceprint belongs to an illegal voiceprint among the abnormal voiceprints, the corresponding communication voice is determined to be an illegal communication voice, and the reception of the communication voice is terminated.

8. The method according to claim 1, characterized in that: The method further comprises: The reported voiceprint and the reported category are received, and when the reported category is an abnormal category, the reported voiceprint is updated into the abnormal voiceprint library as the abnormal voiceprint.

9. The method according to claim 1, characterized in that: Before updating the abnormal voiceprint library, the method further includes: When it is determined that the voiceprint does not exist in the abnormal voiceprint library, the voiceprint is updated to the abnormal voiceprint library as an abnormal voiceprint.

10. The method according to claim 1, characterized in that: The method further comprises: When the category is an abnormal category, the communication voice is determined to be an abnormal communication voice.

11. The method according to claim 6, characterized in that: The method further comprises: When the category is an illegal category in the abnormal category, the communication voice is determined to be an illegal communication voice in the abnormal communication voice and the voiceprint is an illegal voiceprint in the abnormal voiceprint, and the reception of the communication voice is terminated.

12. The method according to claim 1, characterized in that The method further comprises: When the category is the category required by the user, a voice dialogue text matching the content is obtained according to the content of the category required by the user, and the category required by the user may belong to an abnormal category.

13. The method according to claim 1, characterized in that The method further comprises: When the category is other categories, the voiceprint is determined to be a voiceprint of other categories.

14. The method according to claim 13, characterized in that The method further comprises: The other types of voiceprints are set to the corresponding types of voiceprint libraries.

15. The method according to claim 1, characterized in that The method further comprises: According to the category, the category of the source of the communication voice request is determined.

16. The method according to claim 15, characterized in that The method further comprises: The source of the communication voice request is set to a corresponding source category set.

17. The method according to claim 1, characterized in that The user is provided with a viewing address for the voice conversation text.

18. The method according to claim 1, characterized in that The method further comprises: Recording the conversation voice to which the voice conversation text belongs; The user is provided with a listening address of the conversation voice to which the voice conversation text belongs.

19. The method according to claim 1, characterized in that The method further comprises: In the case that the source of the communication voice request does not belong to the preset source, the communication voice request is accepted, and the reception of the user voice and the provision of the communication voice to the user are blocked.

20. The method according to claim 1, characterized in that The method further comprises: In a case where the source of the communication voice request belongs to a preset source, the source of the communication voice request is provided to the user.

21. A method for identifying information, characterized in that: include: Accept the forwarded communication voice request and obtain the voiceprint of the caller's communication voice; If the voiceprint does not belong to the abnormal voiceprint library, obtaining a voice text corresponding to the voice information of the caller; Determine the semantics of the speech text, search for interactive text corresponding to the semantics of the speech text from the dialogue set, and the speech text and its corresponding interactive text are used as speech dialogue text; Using the speech synthesis model, the interactive text is synthesized into speech, and the synthesized speech is sent to the caller; In the conversation, the speech conversation text is processed using a natural language understanding model to obtain the category to which the speech conversation text belongs; When the category to which the voice conversation text belongs is an abnormal category, updating the voiceprint as an abnormal voiceprint to the abnormal voiceprint library; The method further comprises: providing the voice conversation text and the category to which the voice conversation text belongs to the user; Receive adjustment information of the category, and use the adjusted category corresponding to the adjustment information and the corresponding voice dialogue text as samples for optimizing a natural language understanding model for obtaining the category to which the voice dialogue text belongs.

22. The method according to claim 21, characterized in that The method further comprises: The source and reporting category of the reported communication voice request are received, and according to the reporting type, the source of the communication voice request is stored in a corresponding source category set.

23. The method according to claim 22, characterized in that The method further comprises: For any reporting category, the source of the reported communication voice request is stored in the corresponding source category set according to the reporting times of the source.

24. The method according to claim 23, characterized in that The method further comprises: For any reporting category, the source of the reported communication voice request is stored in a corresponding source category set according to the number of reports of the source and the number of reporting users.

25. The method according to any one of claims 22 to 24, characterized in that: The method further comprises: The updated source category set is sent to the corresponding device for updating.

26. A method for identifying information, characterized in that: include: When a communication voice request is received, and the source of the communication voice request is not a preset source, the voiceprint of the caller's communication voice is obtained through the smart speaker, and the reception of the user's voice and the provision of communication voice to the user are blocked; In the case where the voiceprint does not belong to the abnormal voiceprint library, obtaining the voice text corresponding to the communication voice of the caller through the smart speaker; Determine the semantics of the voice text through the smart speaker, search for interactive text corresponding to the semantics of the voice text from the dialogue set, and the voice text and its corresponding interactive text are used as voice dialogue text; Using a speech synthesis model through a smart speaker, the interactive text is synthesized into speech, and the synthesized speech is sent to the caller; Using a natural language understanding model to process the voice conversation text during the conversation through the smart speaker, and obtaining the category to which the voice conversation text belongs; When the category to which the voice conversation text belongs is an abnormal category, updating the voiceprint as an abnormal voiceprint to the abnormal voiceprint library through the smart speaker; The method further includes: providing the voice conversation text and the category to which the voice conversation text belongs to the user through the smart speaker; The adjustment information of the category is received through the smart speaker, and the adjusted category corresponding to the adjustment information and the corresponding voice dialogue text are used as samples for optimizing the natural language understanding model for obtaining the category to which the voice dialogue text belongs.

27. The method according to claim 26, characterized in that: The method further comprises: In response to the user's power-on and authorization operations, the smart speaker is turned on, and the smart speaker receives the communication voice when the source of the communication voice request does not belong to the preset source.

28. The method according to claim 26, characterized in that: The method further comprises: Providing prompt information for users to receive communication voice; In response to the user's receiving operation, the shielding action is cancelled, the user's voice is received, and the communication voice provided to the user is shielded.

29. A method for setting information, characterized in that: include: Obtaining the user's communication voice identifier; Binding the user communication voice identifier with the transfer communication voice identifier, so that under preset conditions, according to the binding relationship, the communication voice request is forwarded to the device to which the transfer communication voice identifier belongs, so that the device obtains the voiceprint of the caller's communication voice, and when the voiceprint does not belong to the abnormal voiceprint library, obtains the voice text corresponding to the voice information of the caller, determines the semantics of the voice text, searches for the interactive text corresponding to the semantics of the voice text from the dialogue set, the voice text and its corresponding interactive text are used as voice dialogue text, synthesizes the interactive text into speech using a speech synthesis model, and sends the synthesized speech to the caller, in the dialogue, uses a natural language understanding model to process the voice dialogue text to obtain the category to which the voice dialogue text belongs, and when the category to which the voice dialogue text belongs is an abnormal category, updates the voiceprint as an abnormal voiceprint to the abnormal voiceprint library, and provides the voice dialogue text and the category to which the voice dialogue text belongs to the user; Setting the preset condition; The method further includes: receiving and displaying a voice conversation text and a category to which the voice conversation text belongs; In response to an input operation by a user, determining adjustment information of a category; The adjustment information of the category is reported, so that the adjusted category corresponding to the adjustment information and the corresponding voice dialogue text are used as samples for optimizing the natural language understanding model for obtaining the category to which the voice dialogue text belongs.

30. The method according to claim 29, characterized in that The method further comprises: Set the processing method for the corresponding communication voice under preset conditions.

31. The method according to claim 29, characterized in that The method further comprises: Receiving a communication voice request and obtaining a communication voice identifier corresponding to the communication voice request; When the acquired voice communication identifier meets the preset condition, the communication voice request is forwarded to the device to which the transferred communication voice identifier belongs according to the binding relationship.

32. The method according to claim 29, characterized in that The method further comprises: The voiceprint of the communication voice is obtained, and the voiceprint and the type of the voiceprint are reported.

33. The method according to claim 32, characterized in that The method further comprises: The voiceprint is matched with the reported voiceprint in the local voiceprint database. When no corresponding voiceprint is matched, the step of reporting the voiceprint and the type of the voiceprint is performed.

34. A method for setting information, characterized in that: include: Providing a binding interface for a communication voice identifier, wherein the binding interface displays a user communication voice identifier input area, a transfer communication voice identifier input area, and a preset condition input area, and the user communication voice identifier and the transfer communication voice identifier have a binding relationship; In response to the user's input operation, determining the user communication voice identifier and the transfer communication voice identifier in the corresponding input area; In response to the user's input operation, the preset condition in the corresponding input area is determined. Under the preset condition, the communication voice request is forwarded to the device to which the forwarded communication voice identifier belongs, so that the device obtains the voiceprint of the caller's communication voice. When the voiceprint does not belong to an abnormal voiceprint library, the voice text corresponding to the voice information of the caller is obtained, the semantics of the voice text is determined, and the interactive text corresponding to the semantics of the voice text is searched from the dialogue set. The voice text and its corresponding interactive text are used as voice dialogue text. The interactive text is synthesized into speech using a speech synthesis model, and the synthesized speech is sent to the caller. In the dialogue, the voice dialogue text is processed using a natural language understanding model to obtain the category to which the voice dialogue text belongs. When the category to which the voice dialogue text belongs is an abnormal category, the voiceprint is updated as an abnormal voiceprint to the abnormal voiceprint library, and the voice dialogue text and the category to which the voice dialogue text belongs are provided to the user; In response to the binding operation, completing the operation of the binding relationship; The method further includes: receiving and displaying a voice conversation text and a category to which the voice conversation text belongs; In response to an input operation by a user, determining adjustment information of a category; The adjustment information of the reported category is used to send an audit notice to the auditor to review the adjustment information, so that the adjusted category corresponding to the adjustment information and the corresponding voice dialogue text can be used as samples for optimizing the natural language understanding model used to obtain the category to which the voice dialogue text belongs.

35. The method according to claim 34, characterized in that The method further comprises: The binding relationship is sent to the communication voice switching device.

36. An information recognition system, characterized in that: include: a first device and a second device; The first device receives a communication voice request, and forwards the communication voice request to the second device if a source of the communication voice request does not belong to a preset source; The second device receives a communication voice request and obtains a voiceprint of the communication voice of the caller; In the case where the voiceprint does not belong to the abnormal voiceprint library, obtaining the voice text corresponding to the voice information of the caller; determining the semantics of the voice text, searching for the interactive text corresponding to the semantics of the voice text from the dialogue set, and the voice text and the interactive text corresponding to the voice text are used as the voice dialogue text; Using a speech synthesis model, synthesizing the interactive text into speech, and sending the synthesized speech to the calling party; In the conversation, the voice conversation text is processed using a natural language understanding model to obtain the category to which the voice conversation text belongs; when the category to which the voice conversation text belongs is an abnormal category, the voiceprint is updated as an abnormal voiceprint to the abnormal voiceprint library, and the voice conversation text and the category to which the voice conversation text belongs are provided to the user; The first device receives and displays the voice dialogue text and the category to which the voice dialogue text belongs; determines the adjustment information of the category in response to the user's input operation; and reports the adjustment information of the category; The second device uses the adjusted category corresponding to the adjustment information and the corresponding voice dialogue text as samples for optimizing a natural language understanding model for obtaining the category to which the voice dialogue text belongs.

37. A computing device comprising a memory, a processor, and a communication component; The memory is used to store computer programs; The communication component is used to receive a communication voice request; The processor is configured to execute the computer program to: When the source of the communication voice request does not belong to the preset source, obtaining the voiceprint of the communication voice of the caller; If the voiceprint does not belong to the abnormal voiceprint library, obtaining the voice text corresponding to the communication voice of the caller; Determine the semantics of the speech text, search for interactive text corresponding to the semantics of the speech text from the dialogue set, and the speech text and its corresponding interactive text are used as speech dialogue text; Using a speech synthesis model, synthesizing the interactive text into speech, and sending the synthesized speech to the calling party; In the conversation, the speech conversation text is processed using a natural language understanding model to obtain the category to which the speech conversation text belongs; When the category to which the voice conversation text belongs is an abnormal category, updating the voiceprint as an abnormal voiceprint to the abnormal voiceprint library; Providing the user with the text of the voice conversation and the category to which the text of the voice conversation belongs; Receive adjustment information of the category, and use the adjusted category corresponding to the adjustment information and the corresponding voice dialogue text as samples for optimizing a natural language understanding model for obtaining the category to which the voice dialogue text belongs.

38. A computing device comprising a memory and a processor; The memory is used to store computer programs; The processor is configured to execute the computer program to: Obtaining the user's communication voice identifier; Binding the user communication voice identifier with the transfer communication voice identifier, so that under preset conditions and according to the binding relationship, the communication voice request is forwarded to the device to which the transfer communication voice identifier belongs so that the device obtains the voiceprint of the caller's communication voice, and when the voiceprint does not belong to the abnormal voiceprint library, obtaining the voice text corresponding to the voice information of the caller, determining the semantics of the voice text, searching for the interactive text corresponding to the semantics of the voice text from the dialogue set, the voice text and its corresponding interactive text are used as voice dialogue text, using the speech synthesis model to synthesize the interactive text into speech, and sending the synthesized speech to the caller, in the dialogue, using the natural language understanding model to process the voice dialogue text to obtain the category to which the voice dialogue text belongs, and when the category to which the voice dialogue text belongs is an abnormal category, updating the voiceprint as an abnormal voiceprint to the abnormal voiceprint library, and providing the voice dialogue text and the category to which the voice dialogue text belongs to the user; Setting the preset condition; Receive and display the voice conversation text and the category to which the voice conversation text belongs; In response to an input operation by a user, determining adjustment information of a category; The adjustment information of the category is reported, so that the adjusted category corresponding to the adjustment information and the corresponding voice dialogue text are used as samples for optimizing the natural language understanding model for obtaining the category to which the voice dialogue text belongs.

39. A computing device comprising a memory and a processor; The memory is used to store computer programs; The processor is used to execute the computer program to: provide a binding interface for a communication voice identifier, wherein the binding interface displays a user communication voice identifier input area, a transfer communication voice identifier input area, and a preset condition input area, and the user communication voice identifier has a binding relationship with the transfer communication voice identifier; In response to the user's input operation, determining the user communication voice identifier and the transfer communication voice identifier in the corresponding input area; In response to the user's input operation, the preset condition in the corresponding input area is determined. Under the preset condition, the communication voice request is forwarded to the device to which the forwarded communication voice identifier belongs, so that the device obtains the voiceprint of the caller's communication voice. When the voiceprint does not belong to an abnormal voiceprint library, the voice text corresponding to the voice information of the caller is obtained, the semantics of the voice text is determined, and the interactive text corresponding to the semantics of the voice text is searched from the dialogue set. The voice text and its corresponding interactive text are used as voice dialogue text. The interactive text is synthesized into speech using a speech synthesis model, and the synthesized speech is sent to the caller. In the dialogue, the voice dialogue text is processed using a natural language understanding model to obtain the category to which the voice dialogue text belongs. When the category to which the voice dialogue text belongs is an abnormal category, the voiceprint is updated as an abnormal voiceprint to the abnormal voiceprint library, and the voice dialogue text and the category to which the voice dialogue text belongs are provided to the user; In response to the binding operation, completing the operation of the binding relationship; Receive and display the voice conversation text and the category to which the voice conversation text belongs; In response to an input operation by a user, determining adjustment information of a category; The adjustment information of the category is reported, so that the adjusted category corresponding to the adjustment information and the corresponding voice dialogue text are used as samples for optimizing the natural language understanding model for obtaining the category to which the voice dialogue text belongs.

40. A computing device comprising a memory and a processor; The memory is used to store computer programs; The processor is used to execute the computer program to: accept the forwarded communication voice request and obtain the voiceprint of the communication voice of the caller; If the voiceprint does not belong to the abnormal voiceprint library, obtaining a voice text corresponding to the voice information of the caller; Determine the semantics of the speech text, search for interactive text corresponding to the semantics of the speech text from the dialogue set, and the speech text and its corresponding interactive text are used as speech dialogue text; Using a speech synthesis model, synthesizing the interactive text into speech, and sending the synthesized speech to the calling party; In the conversation, the speech conversation text is processed using a natural language understanding model to obtain the category to which the speech conversation text belongs; When the category to which the voice conversation text belongs is an abnormal category, updating the voiceprint as an abnormal voiceprint to the abnormal voiceprint library; Providing the user with the text of the voice conversation and the category to which the text of the voice conversation belongs; Receive adjustment information of the category, and use the adjusted category corresponding to the adjustment information and the corresponding voice dialogue text as samples for optimizing a natural language understanding model for obtaining the category to which the voice dialogue text belongs.

41. A computing device comprising a memory and a processor; The memory is used to store computer programs; The processor is used to execute the computer program to: when a communication voice request is received and the source of the communication voice request is not a preset source, obtain the voiceprint of the communication voice of the caller through the smart speaker, and block receiving the user voice and blocking providing the communication voice to the user; In the case where the voiceprint does not belong to the abnormal voiceprint library, obtaining the voice text corresponding to the communication voice of the caller through the smart speaker; Determine the semantics of the voice text through the smart speaker, search for interactive text corresponding to the semantics of the voice text from the dialogue set, and the voice text and its corresponding interactive text are used as voice dialogue text; Using a speech synthesis model through a smart speaker, the interactive text is synthesized into speech, and the synthesized speech is sent to the caller; Using a natural language understanding model to process the voice conversation text during the conversation through the smart speaker, and obtaining the category to which the voice conversation text belongs; When the category to which the voice conversation text belongs is an abnormal category, updating the voiceprint as an abnormal voiceprint to the abnormal voiceprint library through the smart speaker; Providing the voice conversation text and the category to which the voice conversation text belongs to the user through the smart speaker; The adjustment information of the category is received through the smart speaker, and the adjusted category corresponding to the adjustment information and the corresponding voice dialogue text are used as samples for optimizing the natural language understanding model for obtaining the category to which the voice dialogue text belongs.

42. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by one or more processors, the one or more processors are caused to implement the steps of the method according to any one of claims 1 to 34.

Images