Voice information processing method, device, equipment, storage medium and program product

Abstract

The present disclosure provides a voice information processing method, device, equipment, storage medium and program product, artificial intelligence, computer technology field, especially in the voice processing, deep learning technical field. The specific implementation scheme of the voice information processing method is: determining a target audio feature associated with an object identifier according to the received voice information, wherein the target audio feature represents the audio feature of the target object indicated by the object identifier; in the artificial question and answer mode, detecting whether the voice information matches the target audio feature to obtain a target object detection result.

Description

Technical Field

[0001] This disclosure relates to the fields of artificial intelligence and computer technology, and in particular to the fields of speech processing and deep learning technology, specifically to a speech information processing method, apparatus, device, storage medium, and program product. Background Technology

[0002] Speech processing, as an important branch of artificial intelligence, can be applied to various applications. However, in some application scenarios, the efficiency of speech information processing is not high and cannot meet the actual needs of the corresponding application scenarios. Summary of the Invention

[0003] This disclosure provides a voice information processing method, apparatus, device, storage medium, and program product.

[0004] According to one aspect of this disclosure, a voice information processing method is provided, comprising: determining target audio features associated with an object identifier based on an object identifier associated with received voice information, wherein the target audio features characterize the audio features of a target object indicated by the object identifier; and detecting whether the voice information matches the target audio features in a human question-and-answer mode to obtain a target object detection result.

[0005] According to another aspect of this disclosure, a voice information processing apparatus is provided, comprising: a target audio feature determination module and a target object detection result determination module. The target audio feature determination module is used to determine target audio features associated with an object identifier based on an object identifier associated with received voice information, wherein the target audio features characterize the audio features of the target object indicated by the object identifier; the target object detection result determination module is used to detect whether the voice information matches the target audio features in a human question-and-answer mode, thereby obtaining a target object detection result.

[0006] According to another aspect of this disclosure, an electronic device is provided, comprising: at least one processor and a memory communicatively connected to the at least one processor. The memory stores instructions executable by the at least one processor, which, when executed by the at least one processor, enables the at least one processor to perform the methods of embodiments of this disclosure.

[0007] According to another aspect of this disclosure, a non-transitory computer-readable storage medium is provided storing computer instructions for causing a computer to perform the methods of embodiments of this disclosure.

[0008] According to another aspect of this disclosure, a computer program product is provided, including a computer program stored on at least one of a readable storage medium and an electronic device, wherein the computer program, when executed by a processor, implements the methods of embodiments of this disclosure.

[0009] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0010] The accompanying drawings are provided to better understand this solution and do not constitute a limitation of this disclosure. Wherein:

[0011] Figure 1 A schematic diagram illustrating the system architecture of a voice information processing method and apparatus according to embodiments of the present disclosure is shown.

[0012] Figure 2 A flowchart illustrating a voice information processing method according to an embodiment of the present disclosure is shown schematically.

[0013] Figure 3 A schematic diagram of a voice information processing method according to another embodiment of the present disclosure is shown;

[0014] Figure 4 The diagram illustrates a method for processing voice information according to yet another embodiment of the present disclosure, showing how to obtain the target object detection result.

[0015] Figure 5 A schematic diagram of a voice information processing method according to yet another embodiment of the present disclosure is shown.

[0016] Figure 6 A schematic diagram of a voice information processing method according to yet another embodiment of the present disclosure is shown.

[0017] Figure 7 A block diagram of a voice information processing apparatus according to an embodiment of the present disclosure is schematically shown; and

[0018] Figure 8 A block diagram of an electronic device that can implement the voice information processing method of the embodiments of the present disclosure is shown schematically. Detailed Implementation

[0019] The exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments to aid understanding, and should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this disclosure. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0020] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit this disclosure. The terms “comprising,” “including,” etc., as used herein indicate the presence of the stated features, steps, operations, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, or components.

[0021] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein are to be interpreted in a manner consistent with the context of this specification, and not in an idealized or overly rigid way.

[0022] When using expressions such as "at least one of A, B, and C", they should generally be interpreted in accordance with the meaning that is commonly understood by a person skilled in the art (e.g., "a system having at least one of A, B, and C" should include, but is not limited to, a system having A alone, a system having B alone, a system having C alone, a system having A and B, a system having A and C, a system having B and C, and / or a system having A, B, and C, etc.).

[0023] Voice processing, as an important branch of artificial intelligence, can be applied to various scenarios. The following description uses the voice information processing method of this disclosure as an example applied to a call center scenario. A call center can be understood as a service organization located in a relatively centralized location, composed of a group of customer service personnel (i.e., customer service representatives). It typically utilizes computer communication technology to handle telephone inquiries from businesses and customers. Call centers have the ability to handle a large number of incoming calls simultaneously, display caller ID numbers, automatically assign calls to customer service personnel with the appropriate skills, and record and store all incoming call information.

[0024] In a call center setting, multiple customer service seats can be set up, each with a call terminal. Each customer service representative is assigned to at least one customer service seat's call terminal to answer incoming calls.

[0025] In actual call center scenarios, the following situations exist.

[0026] 1) With the development of computer and internet technologies, chatbots can replace human customer service representatives in some communication interactions. However, chatbots are usually programmed with a generic voice and their voices sound mechanical, which reduces the fluency and efficiency of the call. For example, customers may reject chatbots as a result.

[0027] 2) For example, customer service agent A is matched with customer service personnel a, and customer service agent B is matched with customer service personnel b. For some reason, incoming calls to customer service agent A's terminal are answered by customer service personnel b, causing confusion in incoming call answering among customer service agents. This confusion not only affects the management of the call center scenario, but also affects the call experience of the service recipients.

[0028] It should be noted that the collection, storage, use, processing, transmission, provision, and disclosure of user personal information, such as voice recordings, involved in the technical solutions disclosed herein comply with relevant laws and regulations and do not violate public order and good morals.

[0029] Figure 1 The schematic illustration shows the system architecture of a voice information processing method and apparatus according to an embodiment of the present disclosure. It should be noted that... Figure 1 The examples shown are merely examples of system architectures that can be applied to the embodiments of this disclosure, in order to help those skilled in the art understand the technical content of this disclosure, but do not mean that the embodiments of this disclosure cannot be used in other devices, systems, environments or scenarios.

[0030] like Figure 1 As shown, the system architecture 100 according to an embodiment of this disclosure may include communication terminals 101, 102, and 103, a network 104, and a server 105. The network 104 serves as a medium for providing communication links between the communication terminals 101, 102, and 103 and the server 105. The network 104 may include various connection types, such as wired or wireless communication links, or fiber optic cables, etc.

[0031] The calling terminals 101, 102, and 103 can be electronic devices with calling capabilities, such as landline telephones. The target user can use the calling terminals 101, 102, and 103 to make calls, and the obtained voice information can be sent to the server 105 for processing. For example, in a call center scenario, the target user can be a customer service representative, who can use the calling terminals to communicate with the service recipient.

[0032] Server 105 can be a server that provides various services, such as processing voice information from calls made by the target object using terminals 101, 102, and 103. Additionally, server 105 can also be a cloud server, meaning it has cloud computing capabilities.

[0033] For example, the system architecture 100 according to an embodiment of this disclosure may also include a client 106.

[0034] Client 106 can be various electronic devices with a display screen and web browsing support, including but not limited to smartphones, tablets, laptops, and desktop computers. In this embodiment, client 106 can, for example, run an application. Target object detection results, key content, etc., obtained after processing by server 105 can be sent and displayed on client 106.

[0035] It should be noted that the voice information processing method provided in this embodiment can be executed by server 105. Correspondingly, the voice information processing device provided in this embodiment can be located in server 105. The voice information processing method provided in this embodiment can also be executed by a server or server cluster that is different from server 105 and capable of communicating with call terminals 101, 102, 103 and / or server 105. Correspondingly, the voice information processing device provided in this embodiment can also be located in a server or server cluster that is different from server 105 and capable of communicating with call terminals 101, 102, 103 and / or server 105.

[0036] It should be understood that Figure 1 The number of calling terminals, networks, servers, and clients shown is merely illustrative. Depending on implementation needs, any number of calling terminals, networks, servers, and clients can be included.

[0037] This disclosure provides a voice information processing method, which will be described below in conjunction with... Figure 1 The system architecture, referencing Figures 2-6 This describes a voice information processing method according to exemplary embodiments of the present disclosure. The voice information processing method of the embodiments of the present disclosure may, for example, be derived from... Figure 1 The server 105 shown is used to execute this.

[0038] Figure 2 A flowchart illustrating a voice information processing method according to an embodiment of the present disclosure is shown schematically.

[0039] like Figure 2 As shown, the voice information processing method 200 of this embodiment may include, for example, operations S210 to S220.

[0040] In operation S210, the target audio features associated with the object identifier are determined based on the object identifier associated with the received voice information.

[0041] For example, the calling terminal can send voice information of the call, the server can receive the voice information sent by the calling terminal, the server can also determine the corresponding calling terminal based on the voice information, and determine the object identifier that matches the calling terminal based on the corresponding calling terminal.

[0042] It is understood that voice calls using a calling terminal involve two parties, one of whom is the target audience. The voice information processing method of this disclosure embodiment targets the voice information of the target audience. In a call center scenario, among the two parties involved in a voice call, the "target audience" is the customer service representative, and the other party is the service recipient. The following will illustrate the application of the voice information processing method of this disclosure embodiment to a call center scenario.

[0043] The object identifier can be, for example, an agent identifier that matches the call terminal or an identifier of a customer service representative. If the object identifier is an agent identifier that matches the call terminal, the corresponding customer service representative can also be matched based on the agent identifier.

[0044] For example, an object identifier matching the calling terminal can be determined based on the calling terminal that sends the voice information.

[0045] The target audio features characterize the audio features of the target object indicated by the object identifier.

[0046] For example, the target audio features can be predetermined. For instance, target audio features corresponding to each target object can be pre-registered to obtain a set of target audio features. Upon receiving voice information, for example, the target audio features associated with the object identifier can be determined from the set of target audio features based on the object identifier associated with the voice information.

[0047] When operating the S220 in human question-and-answer mode, it detects whether the voice information matches the target audio features to obtain the target object detection result.

[0048] The human-assisted question-and-answer mode can be understood as a mode in which a human conducts the call interaction.

[0049] There is an association between the object identifier and the target object it points to. In some cases, the voice information and the object identifier are confused, which will cause a mismatch between the voice information and the target object.

[0050] Each target object has its own specific target audio features, which can be used as a basis for detecting whether speech information matches the target object.

[0051] According to the voice information processing method of this disclosure, by detecting whether the voice information matches the target audio features, the obtained target object detection result can realize the automated detection of whether the incoming call is chaotic. Since the target object has its specific target audio features, by detecting whether the voice information matches the target audio features, an accurate target object detection result can be obtained.

[0052] For example, the target object detection results can also be sent. These results can be sent to relevant personnel, allowing them to determine whether the speech information matches the target audio features, and thus whether the speech information and the target object are confused. If the target object detection results indicate a mismatch between the object identifier associated with the speech information and the target audio features, further processing can be initiated.

[0053] Figure 3 A schematic diagram of a voice information processing method according to another embodiment of the present disclosure is shown.

[0054] like Figure 3 As shown, the voice information processing method 300 according to an embodiment of the present disclosure may further include operation S330.

[0055] In operation S330, if the speech information represented by the target object detection result does not match the target audio features, the first target interactive speech to be output is determined based on the speech information and the target audio features.

[0056] When the target object detection result represents speech information that does not match the target audio features, it indicates that the speech information and the target object are mismatched. Taking a call center application scenario as an example, the mismatch between speech information and target audio features in the target object detection result indicates that customer service personnel are working different shifts, such as customer service personnel A answering a call from customer service personnel B.

[0057] For example, speech information can be processed based on target audio features using speech synthesis models or similar methods, ensuring that the content of the first target interactive speech matches the speech information and that the first target interactive speech possesses the target audio features. It is understood that each target object has its specific audio features, which may reflect, for example, the target object's pitch, frequency, and other deeper characteristics. The first target interactive speech obtained after processing the speech information based on the target audio features can possess these target audio features, thus restoring the audio characteristics of the target object matching the object identifier.

[0058] According to the voice information processing method of this disclosure, by determining a first target interactive voice for output based on voice information and target audio features, it can address the situation where the voice information and the target object are confused, ensuring that the output first target interactive voice possesses target audio features and restores the audio characteristics of the target object matching the voice information. Since the service recipient hears the first target interactive voice, the call experience for the service recipient can be improved.

[0059] exist Figure 3 The example also schematically illustrates operations S310 to S320. Operations S310 and S320 are similar to operations S210 and S220 described above, respectively, and will not be repeated here.

[0060] exist Figure 3 The example illustrates a specific instance where service object 301 and target object 303-S1 conduct a conversation via call terminal 302-S2 in human question-and-answer mode, obtaining voice information 304. Figure 3 In the example, the object identifier 305 of the target object matched with the call terminal 302-S2 is S2. The voice information 304 sent by the call terminal 302-S2 is associated with the object identifier 305, and the target audio feature F-S2 associated with it can be determined through the object identifier 305. The voice information 304 is obtained from the call between the target object 303-S1 and the service object 301. Therefore, the voice information 304 reflects the voice of the target object S1. By detecting the voice information 304 and the target audio feature F-S2, it can be found that the target object S1 corresponding to the voice information 304 does not match the target object S2 corresponding to the target audio feature F-S2. Therefore, the target object detection result 306 indicates that the voice information 304 and the target audio feature F-S2 do not match. At this time, the first target interactive voice 307 for output can be determined based on the voice information 304 and the target audio feature F-S2. The content of the first target interactive voice 307 can be consistent with the voice information 304, and the first target interactive voice 307 can have the target audio feature F-S2.

[0061] exist Figure 3 The example also schematically illustrates, for example, a call center scenario, N target objects 303 and a total of N object identifiers S1 to SN of the target objects 303. Figure 3 It also schematically illustrates how each target object is associated with a corresponding target audio feature using an object identifier, for example, object identifier S1 is associated with target audio feature F-S1.

[0062] The voice information processing method according to another embodiment of the present disclosure may further include: determining an audio feature set based on the target object indicated by each object identifier.

[0063] For any object identifier, the audio feature set includes multiple candidate audio features associated with the object identifier, and these multiple candidate audio features represent multiple pronunciation states of the target object.

[0064] Target audio features represent the unique audio characteristics of the corresponding target object. As the basis for identifying whether voice information matches the target object, target audio features need to accurately represent the audio characteristics of the target object. However, the target object, as the person answering the call, may exhibit significant changes in their vocalization state in certain situations, making it difficult for the target audio features to accurately match the corresponding target object. For example, the customer service representative, as the target object, may have a different vocalization state compared to a normal emotional state when experiencing emotions such as excitement or frustration.

[0065] According to the speech information processing method of the present disclosure, the audio features of a target object can be accurately and comprehensively characterized by candidate audio features in multiple pronunciation states, thereby improving the matching accuracy between speech information and the target object.

[0066] Figure 4 The diagram illustrates a method for obtaining target object detection results according to yet another embodiment of the present disclosure for voice information processing.

[0067] like Figure 4 As shown, for example, the following embodiment can be used to implement operation S420 in human question-and-answer mode, detecting whether the voice information matches the target audio features, and obtaining the target object detection result.

[0068] In operation S421, the voice features 406 are determined based on the voice information 404.

[0069] For example, the Mel-Cepstral Coefficients of the speech information can be determined, and the Mel-Cepstral Coefficients of the speech information can be used as spectral features. The spectral features can be processed, for example, by convolution, to extract speech features.

[0070] Mel-scale frequency cepstral coefficients, or MFCCs for short, are used to determine parameters that exhibit better robustness, better match the auditory characteristics of the human ear, and maintain good recognition performance even when the signal-to-noise ratio decreases.

[0071] For example, this can be achieved through: pre-emphasis → framing → windowing → Fast Fourier Transform → triangular bandpass filter → Mel-frequency filter bank → calculating the logarithmic energy of each filter bank's output → obtaining MFCC via discrete cosine transform. Pre-emphasis can be implemented by passing the target speech data through a high-pass filter. Pre-emphasis can boost the high-frequency components, flattening the signal spectrum and maintaining the signal across the entire frequency band from low to high frequencies. Pre-emphasis can also eliminate the effects of the vocal cords and lips during phonation, compensating for the high-frequency components of the speech signal suppressed by the vocal system and highlighting high-frequency formants.

[0072] In operation S422, the similarity between the speech features and multiple candidate audio features of the target object is determined, and multiple candidate similarities are obtained.

[0073] For example, parameters such as cosine similarity or spatial distance between speech features and multiple candidate audio features of the target object can be used to characterize the similarity between speech features and multiple candidate audio features of the target object, thereby obtaining multiple candidate similarities.

[0074] In operation S423, the target object detection result 407 is determined based on multiple candidate similarities.

[0075] For example, a target similarity can be preset, and the numerical values ​​of multiple candidate similarities can be compared with the target similarity. If any of the candidate similarities has a value greater than or equal to the target similarity, it can be determined that the speech information represented by the target object detection result matches the target audio features. If the values ​​of multiple candidate similarities are all less than the target similarity, it can be determined that the speech information represented by the target object detection result does not match the target audio features.

[0076] The speech information processing method according to the embodiments of this disclosure determines the similarity between the speech features of the speech information and multiple candidate audio features of the target object, and uses the obtained multiple candidate similarities as a reference for determining the target object detection result, which has higher accuracy.

[0077] exist Figure 4 The example illustrates a specific instance where service object 401 and target object 403-S1 conduct a conversation via call terminal 402-S2 in human question-and-answer mode, obtaining voice information 404. Figure 4 In the example, the object identifier 405 of the target object matched with the call terminal 402-S2 is S2. The voice information 404 sent through the call terminal 402-S2 is associated with the object identifier 405. M candidate audio features associated with the object identifier 405 can be determined. The M candidate audio features are candidate audio features CF1-S2 to candidate audio features CFM-S2.

[0078] exist Figure 4 In the example, speech features 406 can be obtained from speech information 404. For instance, the similarity between speech features 406 and M candidate audio features associated with object identifier 405 can be calculated to obtain M candidate similarities, including candidate similarity FS1 to candidate similarity FSM. Candidate similarity FS1 can be represented, for example, by using the cosine similarity or spatial distance between speech features 406 and the corresponding candidate audio features CF1-S2.

[0079] Figure 5 A schematic diagram of a voice information processing method according to yet another embodiment of the present disclosure is shown.

[0080] like Figure 5 As shown, the voice information processing method 500 according to another embodiment of the present disclosure may further include operation S540.

[0081] In operation of S540, in machine question-and-answer mode, the second target interactive voice 507 for output is determined based on the standard interactive voice 506 corresponding to the machine question-and-answer mode and the target audio features.

[0082] Machine-based question-and-answer (CTO) mode can be understood as a mode of communication where machines conduct the interaction. For example, in some situations, in addition to human interaction, chatbots or similar tools can be used to handle routine question-and-answer interactions. Figure 5 The example illustrates a specific instance where a standard interactive voice 506 is determined by an electronic device, computer 503.

[0083] In machine question-and-answer mode, parameters such as timbre and pronunciation interval of the standard interactive voice 506 output by the robot are set, which makes the standard interactive voice relatively mechanical. For example, service recipients often have a negative attitude towards the standard interactive voice.

[0084] For example, a speech synthesis model or similar method can be used to process the standard interactive speech based on the target audio features, so that the content of the second target interactive speech is consistent with the standard interactive speech, and the second target interactive speech has the target audio features.

[0085] According to the voice information processing method of this disclosure, in machine question-and-answer mode, by comparing standard interactive voice with target audio features, the determined second target interactive voice possesses target audio features, and the audio characteristics of the target object matching the voice information are restored. The service recipient hears the second target interactive voice, which can improve the service recipient's call experience.

[0086] For example, operation S540 may be performed before operation S210 or operation S310.

[0087] exist Figure 5 The example illustrates a specific instance of obtaining voice information 504 by communicating with a service object 501 via a call terminal 502-S2 in machine question-and-answer mode. It also illustrates a specific instance of determining the target audio feature F-S2 associated with the object identifier 505 (specifically S2) in operation S510 based on the object identifier 505 associated with the received voice information 504. Refer to the description in the above embodiments for further details, which will not be repeated here.

[0088] According to another embodiment of the voice information processing method of this disclosure, the machine question-and-answer mode may include a first machine question-and-answer stage and a second machine question-and-answer stage, and the standard interactive voice includes a first standard interactive voice and a second standard interactive voice. The first standard interactive voice corresponding to the first machine question-and-answer stage can be executed by a robot, and the second standard interactive voice corresponding to the second machine question-and-answer stage can be executed by a robot or a target object indicated by an object identifier. The historical response frequency of the first standard interactive voice is higher than the historical response frequency of the second standard interactive voice.

[0089] To alleviate the pressure on human operators, robots can perform standard interactive voice commands. In some business scenarios, standard interactive voice commands are more adaptable and typically involve answering routine questions. While some business questions are answered manually, in real-world scenarios, it is impossible to manually answer a large number of incoming calls with business-related questions.

[0090] For example, historical call records can be analyzed to identify frequently asked business questions that require human intervention. For each frequently asked business question, a corresponding standard interactive voice is determined, which is the second standard interactive voice for the second machine-based question-and-answer phase. When a large number of subsequent calls involve frequently asked business questions, the robot can execute the second standard interactive voice to answer, reducing the workload of human operators.

[0091] When a human can answer business-related calls, a second standard interactive voice can also be used by a human, offering greater flexibility.

[0092] Figure 6 A schematic diagram of a voice information processing method according to yet another embodiment of the present disclosure is shown.

[0093] like Figure 6 As shown, the voice information processing method 600 according to another embodiment of the present disclosure may include, for example, operations S650 to S680.

[0094] In operation S650, in response to question-and-answer mode adjustment instruction 601, the machine question-and-answer mode M1 corresponding to the current object identifier is switched to human question-and-answer mode M2.

[0095] For example, the question-and-answer mode adjustment instruction can be determined by the target object. The voice information processing method according to embodiments of this disclosure can support question-and-answer mode adjustment and flexibly adapt to different scenarios.

[0096] For example, a target object corresponds to two calling terminals, P1 and P2. When both calling terminals receive a call, the target object can answer the call L1 from calling terminal P1. The call L2 from calling terminal P2 can be handled by a robot in machine question-and-answer mode. When the target object finishes answering the call L1, the machine question-and-answer mode corresponding to the call L2 can be switched to human question-and-answer mode.

[0097] In operation S660, the standard interactive voice 602 of the machine question-and-answer mode M1 and the response content 603 corresponding to the standard interactive voice 602 are determined.

[0098] Understandably, in machine question-and-answer mode, the call interaction involves both the robot and the service recipient. The standard question-and-answer voice 602 is executed by the robot, while the corresponding response is made by the service recipient. The response content corresponding to the standard voice can be determined based on the service recipient's response, thus obtaining the complete call interaction process. This complete call interaction process reflects the service recipient's intentions and achieves the call's purpose.

[0099] In operation S670, keywords are extracted from response content 603 to obtain key content 604.

[0100] When operating S680, and switching from machine question-and-answer mode M1 to human question-and-answer mode M2, key information 604 is fed back to the target object 605.

[0101] In real-world applications, when switching from machine-based question-and-answer mode to human-based question-and-answer mode for a specific incoming call, the target user is unaware of the specific details of the conversation and interaction that has already taken place with the service recipient in machine-based question-and-answer mode.

[0102] According to the voice information processing method of this disclosure, when switching from machine-based question-and-answer mode to human-based question-and-answer mode, the specific details of the current call interaction can be obtained by determining the standard interactive voice and corresponding response content in the machine-based question-and-answer mode. By extracting keywords from the response content and feeding back the obtained key content to the target object, the target object can obtain key information about the current call interaction, facilitating efficient, fast, and smooth call interaction with the service recipient in human-based question-and-answer mode.

[0103] For example, key content can be sent to the client corresponding to the target object, and the key content can be displayed on the client of the target object, thus realizing a specific example of feeding back key content to the target object.

[0104] The voice information processing method according to another embodiment of this disclosure may further include, for example,: determining the evaluation value of key content based on evaluation parameters; and determining the question-answering mode switching priority of the corresponding machine question-answering mode based on the evaluation value of the key content.

[0105] The question-and-answer mode switching priority indicates the priority of switching from machine question-and-answer mode to human question-and-answer mode.

[0106] Evaluation parameters include the ratio of key content to the full range of standard interactive speech.

[0107] The ratio of key content to the full amount of standard interactive speech can characterize the progress of the robot's conversational interaction in the current machine question-and-answer mode.

[0108] For example, call answering efficiency can be improved by having one target object correspond to multiple calling terminals.

[0109] When multiple calls come in simultaneously, the target user can determine which call needs to be answered manually. The remaining calls can be set to machine question-and-answer mode. When the current call ends, the call that needs to be answered manually can be selected from the remaining calls in machine question-and-answer mode.

[0110] The voice information processing method according to the embodiments of this disclosure can be adapted to the above-mentioned scenario. By evaluating parameters, the evaluation value of key content can be used to evaluate the necessity of switching the machine question-and-answer mode. The evaluation value of key content can be used to determine the question-and-answer mode switching priority of the corresponding machine question-and-answer mode.

[0111] For example, the target audience can selectively answer multiple incoming calls in machine question-and-answer mode based on the priority switching of the question-and-answer mode, resulting in higher call efficiency.

[0112] Figure 7 A block diagram of a voice information processing apparatus according to an embodiment of the present disclosure is shown schematically.

[0113] like Figure 7 As shown, the voice information processing apparatus 700 of this embodiment includes, for example, a target audio feature determination module 710 and a target object detection result determination module 720.

[0114] The target audio feature determination module 710 is used to determine the target audio features associated with the object identifier based on the object identifier associated with the received voice information.

[0115] The target audio feature characterizes the audio features of the target object indicated by the object identifier.

[0116] The target object detection result determination module 720 is used to detect whether the voice information matches the target audio features in the human question-and-answer mode, and obtain the target object detection result.

[0117] The voice information processing apparatus according to an embodiment of the present disclosure further includes: a first target interactive voice determination module, configured to determine a first target interactive voice for output based on the voice information and the target audio features when the voice information representing the target object detection result does not match the target audio features.

[0118] The voice information processing apparatus according to an embodiment of the present disclosure further includes: an audio feature set determination module, configured to determine an audio feature set based on a target object indicated by each object identifier, wherein, for any object identifier, the audio feature set includes multiple candidate audio features associated with the object identifier, and the multiple candidate audio features represent multiple pronunciation states of the target object.

[0119] According to an embodiment of this disclosure, the target object detection result determination module includes: a speech feature determination submodule, a candidate similarity determination submodule, and a target object detection result determination submodule.

[0120] The speech feature determination submodule is used to determine speech features based on speech information.

[0121] The candidate similarity determination submodule is used to determine the similarity between speech features and multiple candidate audio features of the target object, and obtain multiple candidate similarities.

[0122] The target object detection result determination submodule is used to determine the target object detection result based on multiple candidate similarities.

[0123] The voice information processing apparatus according to an embodiment of the present disclosure further includes: a second target interactive voice determination module, configured to determine, in machine question-and-answer mode, a second target interactive voice for output based on the standard interactive voice corresponding to the machine question-and-answer mode and the target audio features.

[0124] According to the voice information processing apparatus of this disclosure, the machine question-and-answer mode includes a first machine question-and-answer stage and a second machine question-and-answer stage, and the standard interactive voice includes a first standard interactive voice and a second standard interactive voice; the first standard interactive voice corresponding to the first machine question-and-answer stage is executed by a robot, and the second standard interactive voice corresponding to the second machine question-and-answer stage is executed by a robot or a target object indicated by an object identifier; the historical response frequency of the first standard interactive voice is higher than the historical response frequency of the second standard interactive voice.

[0125] The voice information processing apparatus according to embodiments of the present disclosure further includes: a question-and-answer mode switching module, a response content determination module, a key content determination module, and a key content feedback module.

[0126] The question-and-answer mode switching module is used to switch the machine question-and-answer mode corresponding to the current object identifier to the human question-and-answer mode in response to the question-and-answer mode adjustment command.

[0127] The response content determination module is used to determine the standard interactive voice of the machine question-and-answer mode and the response content corresponding to the standard interactive voice.

[0128] The key content identification module is used to extract keywords from the response content to obtain key content.

[0129] The key content feedback module is used to provide key content feedback to the target audience when the machine question-and-answer mode is switched to the human question-and-answer mode.

[0130] The voice information processing apparatus according to embodiments of the present disclosure further includes: an evaluation value determination module and a question-and-answer mode switching priority determination module.

[0131] The evaluation value determination module is used to determine the evaluation value of key content based on evaluation parameters.

[0132] Evaluation parameters include the ratio of key content to the full range of standard interactive speech.

[0133] The question-and-answer mode switching priority determination module is used to determine the question-and-answer mode switching priority of the corresponding machine question-and-answer mode based on the evaluation value of key content.

[0134] The question-and-answer mode switching priority indicates the priority of switching from machine question-and-answer mode to human question-and-answer mode.

[0135] It should be understood that the embodiments of the apparatus portion of this disclosure correspond to the same or similar embodiments of the method portion of this disclosure, and the technical problems solved and the technical effects achieved are also the same or similar. This disclosure will not repeat them here.

[0136] According to embodiments of this disclosure, this disclosure also provides an electronic device, a readable storage medium, and a computer program product.

[0137] Figure 8A schematic block diagram of an example electronic device 800 that can be used to implement embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0138] like Figure 8 As shown, device 800 includes a computing unit 801, which can perform various appropriate actions and processes based on a computer program stored in read-only memory (ROM) 802 or a computer program loaded from storage unit 808 into random access memory (RAM) 803. RAM 803 may also store various programs and data required for the operation of device 800. The computing unit 801, ROM 802, and RAM 803 are interconnected via bus 804. Input / output (I / O) interface 805 is also connected to bus 804.

[0139] Multiple components in device 800 are connected to I / O interface 805, including: input unit 806, such as keyboard, mouse, etc.; output unit 807, such as various types of monitors, speakers, etc.; storage unit 808, such as disk, optical disk, etc.; and communication unit 809, such as network card, modem, wireless transceiver, etc. Communication unit 809 allows device 800 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0140] The computing unit 801 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 801 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 801 performs the various methods and processes described above, such as voice information processing methods. For example, in some embodiments, the voice information processing method may be implemented as a computer software program tangibly contained in a machine-readable medium, such as storage unit 808. In some embodiments, part or all of the computer program may be loaded and / or installed on device 800 via ROM 802 and / or communication unit 809. When the computer program is loaded into RAM 803 and executed by the computing unit 801, one or more steps of the voice information processing method described above may be performed. Alternatively, in other embodiments, the computing unit 801 may be configured to perform the voice information processing method by any other suitable means (e.g., by means of firmware).

[0141] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0142] The program code used to implement the methods of this disclosure may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0143] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0144] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0145] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with embodiments of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0146] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other.

[0147] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this disclosure can be achieved, and this is not limited herein.

[0148] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A method for processing voice information, comprising: Based on an object identifier associated with the received voice information, a target audio feature associated with the object identifier is determined, wherein the target audio feature characterizes the audio features of the target object indicated by the object identifier; In the human question-and-answer mode, the system detects whether the voice information matches the target audio features to obtain the target object detection result. In response to the question-and-answer mode adjustment instruction, the mode corresponding to the current object identifier is switched from machine question-and-answer mode to human question-and-answer mode; Determine the response content corresponding to the standard interactive voice in the machine question-and-answer mode; Extract keywords from the response content to obtain key content; and The key information is fed back to the target object.

2. The method according to claim 1, further comprising: If the target object detection result indicates that the speech information does not match the target audio features, a first target interactive speech for output is determined based on the speech information and the target audio features.

3. The method according to claim 1, further comprising: An audio feature set is determined based on the target object indicated by each of the object identifiers, wherein, for any object identifier, the audio feature set includes multiple candidate audio features associated with the object identifier, and the multiple candidate audio features characterize multiple pronunciation states of the target object.

4. The method according to claim 3, wherein, In the human question-and-answer mode, detecting whether the voice information matches the target audio features to obtain the target object detection result includes: Based on the spoken information, determine the spoken features; Determine the similarity between the speech feature and multiple candidate audio features of the target object to obtain multiple candidate similarities; and The target object detection result is determined based on the multiple candidate similarities.

5. The method according to any one of claims 1-4, further comprising: In machine question-and-answer mode, a second target interactive voice for output is determined based on the standard interactive voice corresponding to the machine question-and-answer mode and the target audio features.

6. The method according to claim 5, wherein, The machine question-and-answer mode includes a first machine question-and-answer stage and a second machine question-and-answer stage. The standard interactive voice includes a first standard interactive voice and a second standard interactive voice. The first standard interactive voice corresponding to the first machine question-and-answer stage is executed by the robot, and the second standard interactive voice corresponding to the second machine question-and-answer stage is executed by the robot or the target object indicated by the object identifier. The historical response frequency of the first standard interactive voice is higher than the historical response frequency of the second standard interactive voice.

7. The method according to claim 5, further comprising: Determine the standard interactive voice for the machine question-and-answer mode.

8. The method according to claim 7, further comprising: The evaluation value of the key content is determined based on the evaluation parameters, wherein the evaluation parameters include the ratio of the key content to the full amount of the standard interactive speech; Based on the evaluation value of the key content, the corresponding question-answering mode switching priority of the machine question-answering mode is determined, and the question-answering mode switching priority represents the priority of switching from the machine question-answering mode to the human question-answering mode.

9. A voice information processing device, comprising: The target audio feature determination module is used to determine target audio features associated with an object identifier based on an object identifier associated with received voice information, wherein the target audio features characterize the audio features of the target object indicated by the object identifier; The target object detection result determination module is used to detect whether the voice information matches the target audio features in the human question-and-answer mode, and obtain the target object detection result; The question-and-answer mode switching module is used to switch the mode corresponding to the current object identifier from machine question-and-answer mode to human question-and-answer mode in response to the question-and-answer mode adjustment instruction. The response content determination module is used to determine the response content corresponding to the standard interactive voice in the machine question-and-answer mode; The key content determination module is used to extract keywords from the response content to obtain key content; and The key content feedback module is used to provide feedback of the key content to the target object.

10. The apparatus according to claim 9, further comprising: The first target interactive voice determination module is used to determine the first target interactive voice for output based on the voice information and the target audio features when the target object detection result indicates that the voice information and the target audio features do not match.

11. The apparatus according to claim 9, further comprising: An audio feature set determination module is used to determine an audio feature set based on the target object indicated by each object identifier, wherein, for any object identifier, the audio feature set includes multiple candidate audio features associated with the object identifier, and the multiple candidate audio features characterize multiple pronunciation states of the target object.

12. The apparatus according to claim 11, wherein, The target object detection result determination module includes: The speech feature determination submodule is used to determine speech features based on the speech information; A candidate similarity determination submodule is used to determine the similarity between the speech feature and multiple candidate audio features of the target object, thereby obtaining multiple candidate similarities; and The target object detection result determination submodule is used to determine the target object detection result based on the multiple candidate similarities.

13. The apparatus according to any one of claims 9-12, further comprising: The second target interactive voice determination module is used to determine the second target interactive voice for output based on the standard interactive voice corresponding to the machine question-and-answer mode and the target audio features in machine question-and-answer mode.

14. The apparatus according to claim 13, wherein, The machine question-and-answer mode includes a first machine question-and-answer stage and a second machine question-and-answer stage. The standard interactive voice includes a first standard interactive voice and a second standard interactive voice. The first standard interactive voice corresponding to the first machine question-and-answer stage is executed by the robot, and the second standard interactive voice corresponding to the second machine question-and-answer stage is executed by the robot or the target object indicated by the object identifier. The historical response frequency of the first standard interactive voice is higher than the historical response frequency of the second standard interactive voice.

15. The apparatus according to claim 13, wherein, The response content determination module is also used to determine the standard interactive voice of the machine question-and-answer mode.

16. The apparatus of claim 15, further comprising: An evaluation value determination module is used to determine the evaluation value of the key content based on evaluation parameters, wherein the evaluation parameters include the ratio of the key content to the full amount of the standard interactive speech; The question-and-answer mode switching priority determination module is used to determine the question-and-answer mode switching priority of the corresponding machine question-and-answer mode based on the evaluation value of the key content. The question-and-answer mode switching priority represents the priority of switching from the machine question-and-answer mode to the human question-and-answer mode.

17. An electronic device comprising: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-8.

18. A non-transitory computer-readable storage medium storing computer instructions, wherein, The computer instructions are used to cause the computer to perform the method according to any one of claims 1-8.

19. A computer program product comprising a computer program stored on at least one of a readable storage medium and an electronic device, the computer program implementing the method according to any one of claims 1-8 when executed by a processor.

Images