CN108877792A - For handling method, apparatus, electronic equipment and the computer readable storage medium of voice dialogue - Google Patents

Abstract

According to example embodiments of the present disclosure, there are provided a method, an apparatus, an electronic device, and a computer-readable storage medium for processing a voice dialog. The method includes providing a first reply to the first speech in response to receiving the first speech from the user, wherein the first reply is generated based on a first recognition result of the first speech. The method also includes receiving a second speech from the user, wherein the second speech is used to correct or supplement the first recognition result. In addition, the method further includes generating a second reply to be provided to the user based on the first voice and the second voice, wherein the second reply is more in line with the user's intention than the first reply. According to the embodiments of the present disclosure, when the chatbot cannot accurately recognize the voice content of the user due to abnormal voice recognition, the user can use further voice dialogues to actively correct or supplement, thereby solving the abnormal voice recognition.

Description

Method, device, electronic device and computer readable storage for processing speech dialogue storage medium

technical field

Embodiments of the present disclosure generally relate to the field of artificial intelligence, and more particularly relate to methods, apparatuses, electronic devices, and computer-readable storage media for processing speech dialogues.

Background technique

In recent years, the concept of "Conversation as a Platform" has become increasingly popular, and more and more network products and applications have begun to use conversational human-computer interaction. A chat robot refers to a computer program or software that can realize human-computer interaction through text, voice or pictures, etc., which can understand the content sent by the user and automatically respond. To a certain extent, chatbots can replace real people for conversations, and they can be integrated into dialogue systems as automatic online assistants for scenarios such as intelligent chat, customer service, and information inquiries.

Voice dialogue is a common form of human-computer interaction. Compared with text dialogue, voice dialogue also involves the processing of voice content, such as front-end recognition, speech recognition, and speech synthesis. Since the dialogue system works based on speech recognition content, it has high requirements on the accuracy of speech recognition. The application scenarios of voice dialogue can include intelligent voice assistants, smart speakers, car navigation, etc.

Contents of the invention

According to example embodiments of the present disclosure, there are provided a method, an apparatus, an electronic device, and a computer-readable storage medium for processing a voice dialog.

In a first aspect of the present disclosure, a method for processing a voice dialog is provided. The method includes: in response to receiving a first speech from a user, providing a first reply to the first speech, wherein the first reply is generated based on a first recognition result of the first speech; receiving a second speech from the user , wherein the second voice is used to correct or supplement the first recognition result; and based on the first voice and the second voice, generate a second reply to be provided to the user, wherein the second reply is more in line with the user's intention than the first reply.

In a second aspect of the present disclosure, an apparatus for processing a voice conversation is provided. The apparatus includes: a first providing module configured to provide a first reply to the first speech in response to receiving the first speech from the user, wherein the first reply is generated based on a first recognition result of the first speech The speech receiving module is configured to receive a second speech from the user, wherein the second speech is used to correct or supplement the first recognition result; and the second providing module is configured to generate the following speech based on the first speech and the second speech A second reply provided to the user, where the second reply is more in line with the user's intent than the first reply.

In a third aspect of the present disclosure, there is provided an electronic device comprising one or more processors; and storage means for storing one or more programs. One or more programs, when executed by one or more processors, cause the electronic device to implement the method or process according to the embodiments of the present disclosure.

In a fourth aspect of the present disclosure, a computer-readable medium is provided, on which a computer program is stored, and when the program is executed by a processor, the method or process according to the embodiments of the present disclosure is implemented.

It should be understood that what is described in the Summary of the Invention is not intended to limit the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will be readily understood through the following description.

Description of drawings

The above and other features, advantages and aspects of the various embodiments of the present disclosure will become more apparent with reference to the following detailed description when taken in conjunction with the accompanying drawings. In the drawings, identical or similar reference numerals denote identical or similar elements, wherein:

Figure 1 shows a schematic diagram of an example environment in which embodiments of the present disclosure can be implemented;

FIG. 2 shows a flowchart of a method for processing a voice dialogue according to an embodiment of the present disclosure;

FIG. 3 shows a schematic diagram of a process for processing voice messages according to an embodiment of the present disclosure;

FIG. 4 shows a flowchart of a method for solving character recognition errors through dialogue according to an embodiment of the present disclosure;

FIG. 5 shows a flow chart of a method for solving digit recognition errors through dialogue according to an embodiment of the present disclosure;

FIG. 6 shows a flow chart of a method for supplementing recognition results through dialogue according to an embodiment of the present disclosure;

FIG. 7 shows a block diagram of an apparatus for processing a voice dialog according to an embodiment of the present disclosure; and

FIG. 8 shows a block diagram of an electronic device capable of implementing various embodiments of the present disclosure.

Detailed ways

Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present disclosure are shown in the drawings, it should be understood that the disclosure may be embodied in various forms and should not be construed as limited to the embodiments set forth herein; For a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are for exemplary purposes only, and are not intended to limit the protection scope of the present disclosure.

In the description of the embodiments of the present disclosure, the term "comprising" and its similar expressions should be interpreted as an open inclusion, that is, "including but not limited to". The term "based on" should be understood as "based at least in part on". The term "one embodiment" or "the embodiment" should be read as "at least one embodiment". Other definitions, both express and implied, may also be included below.

In the scene of voice dialogue, due to the presence of environmental noise or user's accent, abnormal voice recognition (such as voice recognition error or failure to recognize) often occurs. In order to solve the problem of abnormal speech recognition, one improvement method is to improve the accuracy of speech recognition itself, and another improvement method is to increase the error tolerance rate of semantics. However, even if the above two improved methods are used, there will still be scenarios where accurate conversations cannot be made due to abnormal speech recognition. Generally speaking, semantic understanding will be based on the recognition results of speech, and when the chat robot cannot recognize or misidentify the user's intention, it may cause unexpected consequences.

Embodiments of the present disclosure propose a solution for processing voice dialogues. According to an embodiment of the present disclosure, when the chat robot misrecognizes or fails to recognize the user's voice content due to abnormal voice recognition, the user can use further voice dialogues to actively correct or supplement. Therefore, the semantic understanding platform according to the embodiments of the present disclosure can solve the problem of abnormal speech recognition, thereby improving user experience during chatting. Some example embodiments of the present disclosure will be described in detail below with reference to FIGS. 1-8 .

FIG. 1 shows a schematic diagram of an example environment 100 in which embodiments of the present disclosure can be implemented. In the example environment 100, a user 110 is engaged in a voice conversation with a chatbot 120 (also referred to as a "chat engine"). Optionally, the user 110 may be local to the chatbot 120 , that is, the user 110 may directly have a conversation with the chatbot 120 . Alternatively, the user 110 can also use his local device (such as a laptop computer, a desktop computer, a smart phone, a tablet computer, etc.) to conduct a voice conversation with the chat robot 120 through the network. It should be understood that the chatbot 120 can be deployed in a local electronic device, it can also be deployed in the cloud or distributed.

Referring to FIG. 1 , a user 110 sends a voice 121 (referred to as a "first voice") to a chatbot 120, and the chatbot 120 processes the voice 121 and provides a corresponding reply 122 to the user 110 (referred to as a "first reply"). So far, the first round of dialogue between the user 110 and the chatbot 120 has been completed. In some embodiments, the text of the voice 121 can be displayed on the display of the user equipment at the same time, so that the user can understand the current conversation content more clearly.

In an embodiment of the present disclosure, if the reply 122 fails to meet the needs of the user 110 (for example, there is a recognition error in the voice 121 so that the chat robot 120 cannot accurately recognize the intention of the user 110), the user can send a further voice 123 to the chat robot 120 ( called “second voice”) for correction or supplementation, chatbot 120 processes voice 123 and provides corresponding reply 124 (called “second reply”) to user 110 . According to an embodiment of the present disclosure, since the voice 123 is a correction or supplement to the recognition result of the voice 121 , by combining the voices 121 and 123 , the chatbot 120 can more accurately recognize the intention of the user 110 . Table 1 below shows an example of a dialog with the speech corrected by the first speech recognition result.

Table 1

The recognition result of the user's voice: Find Wang Feng's contact information Response from the chatbot: Checking Wang Feng's contact information, please wait The recognition result of the user's voice: It's a bountiful harvest Response from the chatbot: Checking Wang Feng's contact information, please wait

For example, the recognition result of the voice 121 of the user 110 is "looking for contact information of Feng Wang", and the corresponding reply 122 generated by the chatbot 120 is "inquiring about the contact information of Feng Wang, please wait". Since user 110 intended to "find Wang Feng's contact information", he corrected the recognition result of voice 121 in voice 123 and said "it's a good harvest". The chat robot 120 generates a reply 124 "inquiring about Wang Feng's contact information, please wait later" according to the corrected content. In this way, the chatbot 120 can accurately identify the real intention of the user 110 .

Fig. 2 shows a flow chart of a method 200 for processing a voice dialog according to an embodiment of the present disclosure. It should be understood that the method 200 may be performed by the chatbot 120 described above with reference to FIG. 1 .

At block 202, in response to receiving a first speech from a user, a first reply to the first speech is provided, wherein the first reply is generated based on a first recognition result of the first speech. For example, after the chatbot 120 receives the voice 121 from the user 110 , it provides a corresponding reply 122 to the user 110 . In the embodiment of the present disclosure, the first reply fails to accurately recognize the user's intention due to abnormal speech recognition, for example, may cause a wrong reply, or prompts "unrecognizable" and requires the user to say it again.

In some embodiments, reply 122 may be provided to user 110 by voice only. In some other embodiments, in order to allow the user to understand the recognition result of the voice more intuitively, the recognition result of the voice 121 and the text form of the reply 122 may also be presented in a visual way (for example, through a display device).

At block 204, a second speech from the user is received, wherein the second speech is used to correct or supplement the first recognition result. For example, since the reply 122 (ie, the first reply) fails to correctly recognize the user's intention, the chatbot 120 further receives a correction or supplementary speech 123 from the user 110 . That is to say, when the chatbot 120 misrecognizes or fails to recognize the voice content of the user 110 due to a voice recognition error, the user 110 can actively clarify through further voice dialogue. For example, the user can actively correct or actively supplement, including correcting one or more characters and/or numbers through voice, or supplement through voice.

In block 206, based on the first voice and the second voice, a second reply to be provided to the user is generated, wherein the second reply is more consistent with the user's intention than the first reply. For example, the chatbot 120 provides a reply 124 to the user 110 based on the recognition result of the voice 123 and the recognition result of the voice 121 . Since the speech 123 corrects or supplements the recognition result of the speech 121, the chat robot 120 can further understand the intention of the user 110, so the reply 124 (i.e. the second reply) is more in line with the intention of the user 110 than the reply 122 (the first reply), thus Solve the abnormality in speech recognition and improve the user experience of the chat process.

In some embodiments, the action associated with the second reply may be performed if the second reply has correctly identified the user's intent. For example, since the second voice is a correction or supplement to the first recognition result of the first voice, the chat robot can accurately recognize the user's intention, so the chat robot can perform or instruct to perform the action associated with the second reply, such as calling call, map navigation, etc. In some embodiments, if no further voice is received from the user within a threshold time after the second reply is provided, it may be assumed that the second reply has met the user's intention by default, and then the actions associated with the second reply may be directly performed . It should be understood that the actions associated with the second reply may also be directly performed while or before and after the second reply is generated, regardless of whether the second reply already meets the user's intention.

In some embodiments, if the second reply still does not recognize the user's intent, a third voice from the user may be received. Next, based at least in part on the third speech, a third reply is provided to the user. For example, although the second reply is more in line with the user's intention than the first reply, it may still not fully satisfy the user's needs. In this case, the user can initiate further speech to continue to correct or supplement the previous speech recognition results.

FIG. 3 shows a schematic diagram of a process 300 for processing voice messages according to an embodiment of the present disclosure. It should be appreciated that process 300 may be performed by chatbot 120 as described above with reference to FIG. 1 , and that process 300 may be an example implementation of providing replies based on received speech as described above with reference to FIG. 2 .

At block 302, speech from a user is input, and at block 304, the input speech is converted to text by automatic speech recognition (ASR). At block 306, the text is converted into a computer-recognizable representation by natural language understanding (NLU). At block 308, the intents and word slots in the text are extracted and integrated with the historical dialog state through dialog state maintenance (DST). In block 310, according to the current dialog state, an action that best matches the current state is selected through action candidate ranking. After the actions are obtained, at block 312 , natural language is generated (NLG), and at block 314 the generated natural language is subjected to speech synthesis (TTS). Then, at block 316, the speech is output for presentation to the user. In process 300, blocks 302, 304, 314, and 316 relate to speech processing, while blocks 306, 308, 310, and 312 relate to natural language processing, where dialog state maintenance and action candidate ranking constitute dialog management , which can generate actions to be performed based on the semantic representation of speech and the current context, and update the context.

In some embodiments, in order to provide the user with a better interactive experience, the speech recognition result may be displayed on a display device (such as a display of the user equipment). In this case, the voice recognition result and its reply can be presented simultaneously through the display device, so that the user can know the voice recognition result.

FIG. 4 shows a flow chart of a method 400 for solving character recognition errors through dialogue according to an embodiment of the present disclosure. It should be understood that the method 400 may be performed by the chatbot 120 described above with reference to FIG. 1 , and block 402 may be implemented as an example of block 204 described above with reference to FIG. An example implementation of block 206 of .

In block 402, a second speech for correcting one or more characters in a first recognition result is received, wherein a character recognition error exists in the first recognition result of the first speech. At block 404, the first recognition result is corrected using the second recognition result for the second speech. At block 406, a second reply is presented via the display device based on the corrected first recognition result. For example, the following Tables 2-3 show examples of speech dialogues for correcting one or more text errors of the first speech recognition result.

Table 2

The recognition result of the user's voice: I'm going to Xierqi Response from the chatbot: sorry i didn't understand what you mean The recognition result of the user's voice: It's Xi'er Banner Response from the chatbot: Searching for the navigation route to Xierqi

table 3

The recognition result of the user's voice: I'm going to learn Erqi Response from the chatbot: sorry i didn't understand what you mean The recognition result of the user's voice: is the west of the thing, the flag of the flag Response from the chatbot: Searching for the navigation route to Xierqi

In the example of Table 2, the user corrected a single typo "odd" in the recognition result of the first speech in the second speech; in the example of Table 3, the user corrected the typo in the recognition result of the first speech in the second speech Multiple typos of "Xi" and "Qi". Next, the chatbot provides a second reply based on the corrected speech recognition result, for example, in the examples in Table 2 and Table 3, the corrected recognition result is "I'm going to Xierqi". It should be understood that although Chinese characters are corrected in the above embodiments of Tables 2-3, the embodiments of the present disclosure can also be used to correct speech recognition errors in other languages.

FIG. 5 shows a flowchart of a method 500 for resolving digit recognition errors through dialogue according to an embodiment of the present disclosure. It should be understood that the method 500 may be performed by the chatbot 120 described above with reference to FIG. 1 , and block 502 may be implemented as an example of block 204 described above with reference to FIG. An example implementation of block 206 of .

At block 502, a second speech for correcting one or more digits in a first recognition result is received, wherein a digit recognition error exists for the first recognition result of the first speech. At block 504, the first recognition result is corrected using one or more digits in the second recognition result for the second speech. At block 506, a second reply is presented via the display device based on the corrected first recognition result. For example, Table 4 below shows an example of a speech dialogue for correcting a numerical error of the first speech recognition result.

Table 4

The recognition result of the user's voice: Call 13511652271 Response from the chatbot: Calling 13511652271 The recognition result of the user's voice: is 110 Response from the chatbot: Calling 13511052271

In the example in Table 4, the user corrects a single wrong number "6" in the recognition result of the first voice in the second voice, that is, the sixth digit in the phone number should be 0 instead of 6. In some embodiments, during the process of the user correcting numbers, the part of the number that needs to be corrected by the user may be determined based on the maximum match between the recognition results of the second voice and the first voice. For example, in the example in Table 4, the number "110" matches "116" in the recognition result to the greatest extent, so it can be determined that the number "110" is used to correct "116" in the previous round of dialogue. In some embodiments, the user can also correct both erroneous letters and numbers in the speech recognition results at the same time.

The inventors of the present application have found that some speech recognition errors can only be found through visual display (for example, text recognition errors due to homophony), while some speech recognition errors can also be found through speech, such as in phone numbers. One or more digits of was misidentified. In some embodiments, instead of displaying the recognition result of the first voice and the first reply through the display device, the first reply is only provided through voice. In this case, the user may issue a voice for correcting one or more digits in the set of numbers based on the reading of the set of numbers in the first reply. Next, the chatbot provides a second reply that is more in line with the user's intent based on the corrected set of numbers. It should be understood that the scene where the recognition result is not displayed may be a scene without a display (for example, a smart speaker without a display), or a scene with a display but the display is not used to display the speech recognition result (for example, a black screen on a smart phone state voice dialogue).

FIG. 6 shows a flow chart of a method 600 for supplementing recognition results through dialogue according to an embodiment of the present disclosure. It should be understood that the method 600 may be performed by the chat robot 120 described above with reference to FIG. 1 , and block 602 may be implemented as an example of block 204 described above with reference to FIG. An example implementation of block 206 of .

At block 602, a second speech supplementing a first recognition result is received. For example, the first recognition result for the first voice does not fully reflect the needs of the user. At block 604, the first recognition result and the second recognition result for the second speech are combined to generate a third recognition result in response to the content of the second speech semantically complementing the content of the first speech.

Various means can be used to determine whether the content of the two speeches has a supplementary or explanatory relationship. For example, in some embodiments, it may be determined whether the content of the second voice can semantically have a succession relationship with the content of the first voice. For example, if the content of two speeches can be analyzed together as a whole, it can be determined that the two have a semantic succession relationship, and accordingly it can be determined that the content of the second speech is a supplement to the first speech. Alternatively or additionally, in some embodiments, it may be determined whether the probability that the contents of the two speeches appear simultaneously is greater than a predetermined threshold. For example, if the content of the two voices occurs at the same time most of the time, it can be determined that the content of the second voice is a supplement to the first voice.

At block 606, a second reply is presented via the display device based on the third recognition result. For example, Table 5 below shows an example of a dialog supplementing the first speech recognition result.

table 5

In the example in Table 5, the user's intention is to go to the west gate of Peking University, but the speech recognition system has already started to recognize after receiving "Peking University", so it cannot recognize the user's accurate intention. In this case, the user can add information through natural language dialogue, and then combine the recognition results of the two voices to generate a new recognition result "I am going to the West Gate of Peking University", and then generate a corresponding reply based on the new recognition result. In some embodiments, for the incomplete recognition result, the user can directly add information without waiting for the return of the recognition result and analysis result, and without repeating the previous dialogue content.

Fig. 7 shows a block diagram of an apparatus 700 for processing a voice dialogue according to an embodiment of the present disclosure. As shown in FIG. 7 , the device 700 includes a first providing module 710 , a voice receiving module 720 and a second providing module 730 . The first providing module 710 is configured to provide a first reply to the first speech in response to receiving the first speech from the user, wherein the first reply is generated based on a first recognition result of the first speech. The speech receiving module 720 is configured to receive a second speech from the user, wherein the second speech is used to correct or supplement the first recognition result. The second providing module 730 is configured to generate a second reply to be provided to the user based on the first voice and the second voice, wherein the second reply is more in line with the user's intention than the first reply.

In some embodiments, the first providing module 710 includes a first presentation module, and the first presentation module is configured to present the first recognition result and the first reply through a display device.

In some embodiments, the second providing module 730 includes a first correcting module, and the first correcting module is configured to use the second recognition result of the second voice to correct character recognition errors in the first recognition result.

In some embodiments, wherein the second providing module 730 includes a second correction module, the second correction module is configured to use one or more digits in the second recognition result for the second voice to correct the digits in the first recognition result Number recognition error.

In some embodiments, wherein the second speech is used to complement the first recognition result, and the second providing module 730 includes: a combining module configured to, in response to determining that the content of the second speech semantically complements the content of the first speech, Combining the first recognition result and the second recognition result for the second voice to generate a third recognition result; and a second presentation module configured to present a second reply through the display device based on the third recognition result.

In some embodiments, wherein the first providing module 710 includes a first speech providing module, the first speech providing module is configured to provide the first reply only by speech, wherein the first reply includes a set of digits from the first recognition result.

In some embodiments, wherein the second voice is used to correct a set of numbers, and the second providing module 730 includes: a third correction module configured to use one or more of the second recognition results for the second voice correcting a set of digits; and a second voice providing module configured to provide a second reply by voice based on the corrected set of digits.

In some embodiments, the apparatus 700 further includes an action execution module configured to execute an action associated with the second reply in response to the second reply having identified the user's intention.

In some embodiments, the apparatus 700 further includes: a third voice receiving module, configured to receive a third voice from the user in response to the second reply not recognizing the user's intention; and a third providing module, configured to at least Based in part on the third speech, a third reply is provided to the user.

It should be understood that the first providing module 710 , the voice receiving module 720 and the second providing module 730 shown in FIG. 7 may be included in the chat robot 120 described with reference to FIG. 1 . Moreover, it should be understood that the modules shown in FIG. 7 may execute steps or actions in the methods or processes referring to the embodiments of the present disclosure.

FIG. 8 shows a schematic block diagram of an example device 800 that may be used to implement embodiments of the present disclosure. It should be understood that the device 800 may be used to implement the apparatus 700 for processing voice dialogue or the chat robot 120 described in this disclosure. As shown, device 800 includes a central processing unit (CPU) 801 that can be programmed according to computer program instructions stored in read only memory (ROM) 802 or loaded from storage unit 808 into random access memory (RAM) 803 program instructions to perform various appropriate actions and processes. In the RAM 803, various programs and data necessary for the operation of the device 800 can also be stored. The CPU 801 , ROM 802 , and RAM 803 are connected to each other via a bus 804 . An input/output (I/O) interface 805 is also connected to the bus 804 .

Multiple components in the device 800 are connected to the I/O interface 805, including: an input unit 806, such as a keyboard, a mouse, etc.; an output unit 807, such as various types of displays, speakers, etc.; a storage unit 808, such as a magnetic disk, an optical disk, etc. ; and a communication unit 809, such as a network card, a modem, a wireless communication transceiver, and the like. The communication unit 809 allows the device 800 to exchange information/data with other devices over a computer network such as the Internet and/or various telecommunication networks.

The processing unit 801 executes the various methods and processes described above, such as the method 200 , the process 300 , the method 400 , the method 500 and the method 600 . For example, in some embodiments, method 200 , process 300 , method 400 , method 500 , and method 600 may be implemented as a computer software program tangibly embodied on 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 onto the device 800 via the ROM 802 and/or the communication unit 809 . When the computer program is loaded into RAM 803 and executed by CPU 801 , one or more actions or steps of method 200 , process 300 , method 400 , method 500 and method 600 described above may be performed. Alternatively, in other embodiments, the CPU 801 may be configured to execute the method 200 , the process 300 , the method 400 , the method 500 and the method 600 in any other suitable manner (for example, by means of firmware).

The functions described above herein may be performed at least in part by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: Field Programmable Gate Array (FPGA), Application Specific Integrated Circuit (ASIC), Application Specific Standard Product (ASSP), System on a Chip (SOC), load programmable logic device (CPLD), etc.

Program codes for implementing the methods of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special purpose computer, or other programmable data processing devices, so that the program codes, when executed by the processor or controller, make the functions/functions specified in the flow diagrams and/or block diagrams Action is implemented. The program code may execute entirely on the machine, partly on the machine, as a stand-alone software package partly on the machine and partly on a remote machine or entirely on the remote machine or server.

In the context of the present disclosure, a machine-readable medium may 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 may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, 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 would include one or more wire-based electrical connections, portable computer discs, hard drives, random access memory (RAM), read only memory (ROM), erasable programmable read only memory (EPROM or flash memory), optical fiber, compact disk read only memory (CD-ROM), optical storage, magnetic storage, or any suitable combination of the foregoing.

Furthermore, although acts or steps are depicted in a particular order, this should be understood as requiring that such acts or steps be performed in the particular order shown or in a sequential order, or that all illustrated acts or steps should be performed to achieve desired result. Under certain circumstances, multitasking and parallel processing may be advantageous. Likewise, while the above discussion contains several specific implementation details, these should not be construed as limitations on the scope of the disclosure. Certain features that are described in the context of separate embodiments can also be implemented in combination in a single implementation. Conversely, various features that are described in the context of a single implementation can also be implemented in multiple implementations separately or in any suitable subcombination.

Although the embodiments of the present disclosure have been described in language specific to structural features and/or methodological acts, it is to be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or acts described above. Rather, the specific features and acts described above are merely example forms of implementing the claims.

Claims (20)

1. A method for processing a speech dialogue, comprising: in response to receiving a first speech from a user, providing a first reply to the first speech, the first reply being generated based on a first recognition result for the first speech; receiving a second voice from the user, the second voice being used to correct or supplement the first recognition result; and Based on the first voice and the second voice, a second reply to be provided to the user is generated, the second reply is more in line with the user's intention than the first reply. 2. The method of claim 1, wherein providing a first reply to the first speech comprises: Presenting the first recognition result and the first reply through a display device. 3. The method of claim 2, wherein generating the second reply comprises: Using the second recognition result of the second speech to correct the character recognition error in the first recognition result. 4. The method of claim 2, wherein generating the second reply comprises: Using one or more digits in the second recognition result for the second speech to correct digit recognition errors in the first recognition result. 5. The method of claim 2, wherein generating the second reply comprises: In response to determining that the content of the second speech semantically complements the content of the first speech, combining the first recognition result with a second recognition result for the second speech to generate a third recognition result; and Presenting the second reply through the display device based on the third recognition result. 6. The method of claim 1, wherein providing a first reply to the first speech comprises: The first reply is provided by voice, the first reply includes a set of digits from the first recognition result. 7. The method of claim 6, wherein generating the second reply comprises: correcting the set of digits using one or more digits in the second recognition result for the second speech; and The second reply is provided by voice based on the corrected set of digits. 8. The method of any one of claims 1-7, further comprising: An action associated with the second reply is performed in response to the second reply having identified the intent of the user. 9. The method of any one of claims 1-7, further comprising: receiving a third speech from the user in response to the second reply not yet recognizing the intent of the user; and A third reply is provided to the user based at least in part on the third speech. 10. An apparatus for processing a voice conversation, comprising: A first providing module configured to, in response to receiving a first voice from a user, provide a first reply to the first voice, the first reply being selected based on a first recognition result for the first voice generate; A speech receiving module configured to receive a second speech from the user, the second speech is used to correct or supplement the first recognition result; and The second providing module is configured to generate a second reply to be provided to the user based on the first voice and the second voice, the second reply is more in line with the user's preference than the first reply intention. 11. The apparatus of claim 10, wherein the first providing module comprises: The first presentation module is configured to present the first recognition result and the first reply through a display device. 12. The apparatus of claim 11, wherein the second providing module comprises: The first correction module is configured to use the second recognition result of the second speech to correct the character recognition error in the first recognition result. 13. The apparatus of claim 11, wherein the second providing module comprises: The second correction module is configured to use one or more digits in the second recognition result for the second speech to correct digit recognition errors in the first recognition result. 14. The apparatus of claim 11, wherein the second providing module comprises: a combination module configured to combine the first recognition result and the second recognition result for the second speech to generate a second recognition result in response to determining that the content of the second speech semantically complements the content of the first speech 3. Identification results; and The second presentation module is configured to present the second reply through the display device based on the third recognition result. 15. The apparatus of claim 10, wherein the first providing module comprises: The first speech providing module is configured to provide the first reply by speech, the first reply includes a set of numbers from the first recognition result. 16. The apparatus of claim 15, wherein the second providing module comprises: A third correction module configured to use one or more digits in the second recognition result for the second speech to correct the set of digits; and The second voice providing module is configured to provide the second reply only by voice based on the corrected set of digits. 17. The apparatus according to any one of claims 10-16, further comprising: An action execution module configured to execute an action associated with the second reply in response to the second reply having identified the user's intention. 18. The apparatus according to any one of claims 10-16, further comprising: A third voice receiving module configured to receive a third voice from the user in response to the second reply not recognizing the user's intention; and A third providing module configured to provide a third reply to the user based at least in part on the third voice. 19. An electronic device comprising: one or more processors; and A storage device for storing one or more programs, and when the one or more programs are executed by the one or more processors, the electronic device implements the electronic device according to any one of claims 1-9 method. 20. A computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the method according to any one of claims 1-9 is implemented.