Information processing method and device, equipment, storage medium and program product
By obtaining user voice input information in the intelligent voice navigation system, identifying the candidate criteria with the highest confidence level and adding it as an extended question, the problem of intent recognition errors is solved, and the system's intent recognition accuracy and user experience is improved.
Patent Information
- Application Number
- CN202411216919.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-02
- Publication Date
- 2024-12-06
AI Technical Summary
In the process of user interaction, the intelligent voice navigation system reduces the user experience due to the diversity of user expressions and logical entanglement.
By obtaining the user's voice input information, generating text information, and determining the candidate criteria with the highest confidence score from the preset semantic knowledge base, the user selects the target criteria and sends target criteria and inquiry information to the semantic knowledge base. And add the query information as an extended question to increase the confidence of the next identification.
Effectively correct intent misrecognition, improve the intent recognition accuracy of intelligent voice navigation, improve user experience, timely optimize the semantic knowledge base, and reduce manual labeling costs.
Smart Images

Figure CN119089994A_ABST
Abstract
Description
Technical Field
[0001] The present application belongs to the field of artificial intelligence technology, and in particular, relates to an information processing method, device, equipment, storage medium and program product. Background Art
[0002] During the interaction between the intelligent voice navigation system and the user, the system can recognize the user's intention based on the user's voice, and the conversation management system can answer questions or perform corresponding business processing based on the intention recognition results.
[0003] However, due to the diversity of user expressions and logical entanglement between nodes, errors in intent recognition may occur, resulting in a decline in user experience. Summary of the invention
[0004] The embodiments of the present application provide an information processing method, apparatus, device and computer storage medium, which can improve the accuracy of intent recognition.
[0005] In a first aspect, an embodiment of the present application provides an information processing method, including:
[0006] Obtaining user query information, where the query information includes text information generated according to the user's voice input information;
[0007] Determine N candidate standard questions corresponding to the query information from a preset semantic knowledge base, the semantic knowledge base includes multiple preset standard questions, the N candidate standard questions are the first N standard questions with the highest confidence scores among the multiple preset standard questions, and N is an integer greater than 1;
[0008] Obtain the user's selection results for N candidate criteria;
[0009] In the case where the selection result includes a first selection result, determining a type of the target standard question, wherein the first selection result includes determining the target standard question from N candidate standard questions;
[0010] When the target standard question is not the first standard question, the target standard question and query information are sent to the semantic knowledge base, and the first standard question is the candidate standard question with the highest confidence score among the N candidate standard questions;
[0011] Add the query information as an extension of the target standard question.
[0012] In a possible implementation, after obtaining the user's selection result for the N candidate standard questions, the method further includes:
[0013] In a case where the selection results include the first selection result, a target operation corresponding to the target criterion is performed.
[0014] In a possible implementation, determining N candidate standard questions corresponding to the query information from a preset semantic knowledge base includes:
[0015] Perform semantic recognition on the query information to obtain a semantic recognition result;
[0016] Calculating the similarity between a question set corresponding to a plurality of preset standard questions and the semantic recognition result respectively, and obtaining a plurality of similarity results corresponding to the plurality of preset standard questions, wherein the question set includes the preset standard questions and the extended questions corresponding to the preset standard questions;
[0017] Calculating confidence scores for each of the multiple preset standard questions according to the similarity results corresponding to each of the multiple preset standard questions;
[0018] The first N standard questions with the highest confidence scores among the multiple preset standard questions are determined as N candidate standard questions.
[0019] In a possible implementation, obtaining the user's selection result for N candidate standard questions includes:
[0020] Play a prompt tone corresponding to the first candidate standard question, where the first candidate standard question is the standard question with the highest confidence score among the N candidate standard questions;
[0021] Obtaining the user's voice selection input for the first candidate standard question;
[0022] In the case where the voice selection input is the target input, the first candidate standard question is determined as the target standard question to obtain a first selection result;
[0023] When the voice selection input is not the target input, the standard question with the highest confidence score among the remaining candidate standard questions is used as the new first candidate standard question, and the prompt sound corresponding to the first candidate standard question is played back until the number of remaining candidate standard questions is 0, and a second selection result is obtained, which includes determining that the target standard question is not included in the N candidate standard questions.
[0024] In a possible implementation, after obtaining the user's selection result for the N candidate standard questions, the method further includes:
[0025] In the case where the selection result includes the second selection result, prompt information is played to prompt the user to perform an operation through the touch control system.
[0026] In a possible implementation, obtaining the user's query information includes:
[0027] When the intelligent voice navigation system is in the target state, collecting the user's voice input information;
[0028] Perform voice recognition on the voice input information to obtain inquiry information.
[0029] In a second aspect, an embodiment of the present application provides an information processing device, including:
[0030] An acquisition module, used to acquire the user's inquiry information, the inquiry information including text information generated according to the user's voice input information;
[0031] A determination module, used to determine N candidate standard questions corresponding to the query information from a preset semantic knowledge base, the semantic knowledge base includes multiple preset standard questions, the N candidate standard questions are the first N standard questions with the highest confidence scores among the multiple preset standard questions, and N is an integer greater than 1;
[0032] The acquisition module is also used to obtain the user's selection results for N candidate standard questions;
[0033] The determination module is further used to determine the type of the target standard question when the selection result includes a first selection result, wherein the first selection result includes determining the target standard question from N candidate standard questions;
[0034] a sending module, configured to send the target standard question and query information to the semantic knowledge base when the target standard question is not the first standard question, and the first standard question is the candidate standard question with the highest confidence score among the N candidate standard questions;
[0035] Add module to add query information as an extension of the target standard question.
[0036] In a third aspect, an embodiment of the present application provides an information processing device, the device comprising:
[0037] A processor and a memory storing computer program instructions; when the processor executes the computer program instructions, any one of the above information processing methods is implemented.
[0038] In a fourth aspect, an embodiment of the present application provides a computer storage medium, on which computer program instructions are stored. When the computer program instructions are executed by a processor, any one of the above-mentioned information processing methods is implemented.
[0039] In a fifth aspect, an embodiment of the present application provides a computer program product. When instructions in the computer program product are executed by a processor of an electronic device, the electronic device is enabled to perform any one of the above-mentioned information processing methods.
[0040] The information processing method, apparatus, device and computer storage medium of the embodiment of the present application can obtain the user's query information, which includes text information generated according to the user's voice input information. Then, the N candidate standard questions corresponding to the query information are determined from the preset semantic knowledge base. The N candidate standard questions are the first N standard questions with the highest confidence scores among the multiple preset standard questions in the semantic knowledge base, and N is an integer greater than 1. In this way, multiple standard questions closest to the user's true intention can be used as candidate standard questions for the user to select. Then the user's selection result for the N candidate standard questions can be obtained. When the selection result includes the first selection result, that is, when the user determines the target standard question from the N candidate standard questions, the type of the target standard question is determined. When the target standard question is not the first standard question, the target standard question and the query information are sent to the semantic knowledge base, and the query information is added as an extended question of the target standard question. The first standard question is the candidate standard question with the highest confidence score among the N candidate standard questions. In this way, the next time the query information is recognized for intent, the confidence score of the target standard question can be improved because the query information is included in the extended question of the target standard question. In this way, the target standard question and query information can be automatically used for optimization and training of the semantic knowledge base, increasing the probability of hitting the correct intent the next time the user expresses it with the same words. In this way, it is possible to effectively correct misidentification of intent, and timely and efficiently perform self-learning optimization of the semantic knowledge base, thereby improving the accuracy of intent recognition of intelligent voice navigation, and thus improving user experience. BRIEF DESCRIPTION OF THE DRAWINGS
[0041] In order to more clearly illustrate the technical solution of the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.
[0042] Figure 1 It is a flowchart of an information processing method provided by an embodiment of the present application;
[0043] Figure 2 is a structural diagram of an information processing device provided by another embodiment of the present application;
[0044] Figure 3 It is a structural diagram of an information processing device provided by yet another embodiment of the present application. DETAILED DESCRIPTION
[0045] The features and exemplary embodiments of various aspects of the present application will be described in detail below. In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only intended to explain the present application, rather than to limit the present application. For those skilled in the art, the present application can be implemented without the need for some of these specific details. The following description of the embodiments is only to provide a better understanding of the present application by illustrating the examples of the present application.
[0046] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the statement "include..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.
[0047] It should be noted that the acquisition, storage, use and processing of data in the embodiments of the present application are in compliance with the relevant provisions of national laws and regulations.
[0048] It should be noted that in the embodiments of the present application, certain software, components, models and other existing solutions in the industry may be mentioned, and they should be regarded as exemplary. Their purpose is only to illustrate the feasibility of implementing the technical solution of the present application, but it does not mean that the applicant has or will necessarily use the solution.
[0049] As described in the background art, during the interaction between the intelligent voice navigation system and the user, intent recognition errors may occur, resulting in a decrease in user experience.
[0050] The relevant technology involves reducing the probability of misidentification of intent through manual annotation. Specifically, it is necessary to define annotation rules, and then collect voice interaction data. The annotators annotate the data according to the annotation rules to obtain the annotation results. The data is clearly screened according to the annotation results for iterative optimization of the semantic knowledge base.
[0051] However, the above method consumes a lot of time and manpower. In addition, manual labeling has a lag, and it is necessary to collect voice interaction data first, then label it, and then optimize it. Therefore, the above solution cannot correct the misidentification of intent in time, resulting in a poor user experience.
[0052] In order to solve the problems of the prior art, the embodiments of the present application provide an information processing method, apparatus, device and computer storage medium.
[0053] The following first introduces the information processing method of the embodiment of the present application with reference to the drawings.
[0054] Figure 1 FIG. 1 is a flow chart showing an information processing method provided by an embodiment of the present application. Figure 1 As shown, the method may include the following steps S110 to S160.
[0055] S110, obtaining user query information, where the query information includes text information generated according to the user's voice input information.
[0056] In step S110, the inquiry information may include business inquiry information initiated to the intelligent voice navigation system. The business inquiry information may be used to obtain certain business information, or to request the business system to handle corresponding business for the user. In one example, the user may call an interactive voice response (IVR) and call the intelligent voice navigation system through the IVR. The above inquiry information may be obtained through the voice recognition module of the intelligent voice navigation system.
[0057] S120, determining N candidate standard questions corresponding to the query information from a preset semantic knowledge base, where the semantic knowledge base includes multiple preset standard questions, and the N candidate standard questions are first N standard questions with the highest confidence scores among the multiple preset standard questions, where N is an integer greater than 1.
[0058] In step S120, the confidence score of the preset standard question can represent the probability that the preset standard question is the true intention of the user corresponding to the query information. Exemplarily, the confidence score can be determined based on the semantic similarity between the preset standard question and the query information, or can be determined based on a preset artificial intelligence algorithm, etc., which is not limited here. In the embodiment of the present application, the value of N can be selected according to actual needs. For example, N can be 2, 3, 4, 5... and so on, which is not limited in the embodiment of the present application.
[0059] S130, obtaining the user's selection results for N candidate standard questions.
[0060] In step S130, the selection result may include the selection result of the user selecting N candidate standard questions in sequence during the voice interaction with the intelligent voice navigation system. Exemplarily, the intelligent voice navigation system may play the prompt tone of the N candidate standard questions, and the user may select the N candidate standard questions by voice. Based on the user's selection, the intelligent voice navigation system may obtain the corresponding selection result.
[0061] S140: When the selection result includes a first selection result, determine the type of the target standard question, wherein the first selection result includes determining the target standard question from N candidate standard questions.
[0062] S150: When the target standard question is not the first standard question, send the target standard question and query information to the semantic knowledge base, and the first standard question is the candidate standard question with the highest confidence score among the N candidate standard questions.
[0063] S160: Add the query information as an extended question of the target standard question.
[0064] The information processing method of the embodiment of the present application can obtain the user's query information, which includes text information generated according to the user's voice input information. Then, N candidate standard questions corresponding to the query information are determined from the preset semantic knowledge base. The N candidate standard questions are the first N standard questions with the highest confidence scores among the multiple preset standard questions in the semantic knowledge base, and N is an integer greater than 1. In this way, multiple standard questions closest to the user's true intention can be used as candidate standard questions for the user to select. Then the user's selection result for the N candidate standard questions can be obtained. In the case where the selection result includes the first selection result, that is, the user determines the target standard question from the N candidate standard questions, the type of the target standard question is determined. In the case where the target standard question is not the first standard question, the target standard question and the query information are sent to the semantic knowledge base, and the query information is added as an extended question of the target standard question. The first standard question is the candidate standard question with the highest confidence score among the N candidate standard questions. In this way, when the query information is recognized for the next time, the confidence score of the target standard question can be improved because the query information is included in the extended question of the target standard question. In this way, the target standard question and query information can be automatically used for optimization and training of the semantic knowledge base, increasing the probability of hitting the correct intent the next time the user uses the same words to express. As a result, it is possible to effectively correct misidentification of intent, and timely and efficiently perform self-learning optimization of the semantic knowledge base, thereby improving the accuracy of intent recognition of intelligent voice navigation, and thus improving user experience.
[0065] In one embodiment, after obtaining the user's selection result for N candidate criteria, the method may further include:
[0066] In a case where the selection results include the first selection result, a target operation corresponding to the target criterion is performed.
[0067] In the above implementation, the target operation may include handling corresponding services or providing corresponding answers to target standard questions.
[0068] According to the above implementation, the intention recognition result can be determined by the user's autonomous selection, and the operation corresponding to the intention recognition result can be performed. In this way, the probability of intention recognition error can be reduced, the accuracy of intelligent voice navigation can be improved, and the user experience can be improved.
[0069] In one embodiment, determining N candidate standard questions corresponding to the query information from a preset semantic knowledge base may specifically include:
[0070] Perform semantic recognition on the query information to obtain a semantic recognition result.
[0071] The similarities between the question sets corresponding to the plurality of preset standard questions and the semantic recognition results are respectively calculated to obtain a plurality of similarity results corresponding to the plurality of preset standard questions, wherein the question sets include the preset standard questions and the extended questions corresponding to the preset standard questions.
[0072] The confidence scores of the plurality of preset standards are calculated respectively according to the similarity results corresponding to the plurality of preset standards.
[0073] The first N standard questions with the highest confidence scores among the multiple preset standard questions are determined as N candidate standard questions.
[0074] In the above implementation, the similarity can be calculated by a variety of methods, for example, it can be calculated by a method based on a semantic dictionary, a method based on deep learning, a method based on a knowledge graph, word vectors, edit distance, word mover's distance (WMD) or BERT (Bidirectional Encoder Representations from Transformers) similarity, etc., without limitation here.
[0075] Exemplarily, calculating the similarity between the question set corresponding to the preset standard question and the semantic recognition result may include: calculating the similarity between the preset standard question and the semantic recognition result, and respectively calculating the similarity between each extended question corresponding to the preset standard question and the semantic recognition result, to obtain the similarity result of the preset standard question.
[0076] Exemplarily, calculating the confidence score of the preset standard question according to the similarity result corresponding to the preset standard question may include: calculating the confidence score according to the preset calculation rule based on the similarity between the preset standard question and the semantic recognition result and the similarity between each extended question and the semantic recognition result. The preset calculation rule may be set according to the needs of those skilled in the art. As an example, the calculation rule may include calculating the average value of the similarity between the preset standard question and the semantic recognition result and the similarity between each extended question and the semantic recognition result as the confidence score. As another example, the weights of the similarity between the preset standard question and the semantic recognition result and the similarity between each extended question and the semantic recognition result may also be set respectively, and the calculation rule may also include calculating the weighted average value of the similarity between the preset standard question and the semantic recognition result and the similarity between each extended question and the semantic recognition result as the confidence score. As another example, the calculation rule may also include calculating the average value of the similarity between the preset standard question and the semantic recognition result and the similarity between each extended question and the semantic recognition result, and processing the average value according to the preset confidence score mapping rule to obtain the confidence score.
[0077] According to the above implementation, the confidence score can be calculated by the similarity between the semantic recognition result and the standard question and the extended question with different standard question words. In this way, the confidence score of the preset standard question can be comprehensively measured by the similarity between the semantic recognition and different words. In this way, when the user uses a different expression method from the standard question for voice interaction, the intention recognition can also be performed more accurately, thereby improving the accuracy of the intention recognition.
[0078] In one embodiment, obtaining the user's selection result for N candidate criteria may specifically include:
[0079] Play the prompt tone corresponding to the first candidate standard question, where the first candidate standard question is the standard question with the highest confidence score among the N candidate standard questions.
[0080] Obtain the user's voice selection input for the first candidate standard question.
[0081] When the voice selection input is the target input, the first candidate standard question is determined as the target standard question to obtain a first selection result.
[0082] When the voice selection input is not the target input, the standard question with the highest confidence score among the remaining candidate standard questions is used as the new first candidate standard question, and the prompt sound corresponding to the first candidate standard question is played back until the number of remaining candidate standard questions is 0, and a second selection result is obtained, which includes determining that the target standard question is not included in the N candidate standard questions.
[0083] According to the above implementation, the prompt sounds corresponding to the candidate standard questions can be played in order from high to low confidence scores for the user to select. When the user determines the first candidate standard question as the target standard question, the first selection result is obtained, or when the number of remaining candidate standard questions is 0, the second selection result is obtained. In this way, the candidate standard questions that are more likely to be the user's true intention can be selected first, thereby improving the efficiency of intent recognition and enhancing the user's experience in voice interaction.
[0084] In one embodiment, after obtaining the user's selection result for N candidate criteria, the method may further include:
[0085] In the case where the selection result includes the second selection result, prompt information is played to prompt the user to perform an operation through the touch control system.
[0086] In the above implementation, the touch control system may include a touch control system of an interactive device, for example, may include a button menu.
[0087] According to the above implementation, when none of the N candidate criteria can express the user's true intention, the user can be prompted to operate through the touch interaction method. In this way, when the intention recognition fails, other interaction methods can be provided in time for the user to operate, thereby solving the user's needs in time and improving the user experience.
[0088] In one embodiment, obtaining the user's query information may specifically include:
[0089] When the intelligent voice navigation system is in the target state, the user's voice input information is collected.
[0090] Perform voice recognition on the voice input information to obtain inquiry information.
[0091] In the above implementation, the target state may include the intelligent voice navigation system being in a state of receiving user voice input, rather than being in other states such as voice playback.
[0092] According to the above implementation, the user's voice input information can be recognized when the intelligent voice navigation system is in the target state. In this way, the probability of erroneously collecting the user's non-interactive voice can be reduced. For example, while waiting for the intelligent voice navigation system to perform intent recognition, the user may talk to others, and the voice recognition of the conversation content will not be performed at this time. In this way, not only can the unnecessary voice recognition processing volume be reduced, but also the user's interactive experience can be improved.
[0093] For ease of understanding, based on the above implementation, a specific example is given below to illustrate the information processing method of the embodiment of the present application.
[0094] Exemplarily, for the sake of convenience, N may be 3. The information processing method may include the following steps S210 to S280.
[0095] S210, the user makes a call to enter the IVR, and the IVR calls up the intelligent voice navigation system.
[0096] S220, the intelligent voice navigation system plays a menu prompt tone and recognizes the user's voice input information, transcribes the user's audio into text through voice recognition, and obtains a transcribed text of the user's voice.
[0097] S230, parsing the transcribed text of the user's speech through a semantic knowledge base to obtain three semantic results, which are standard questions with the first, second and third confidence scores.
[0098] S240, interacting with the user and playing a standard service prompt tone with the highest confidence score.
[0099] S250, if the user expresses denial to the service prompt tone of the standard question with the first confidence score, then play the service prompt tone of the standard question with the second confidence score, and continue to execute step S260; if the user expresses confirmation, then handle the corresponding service of the standard question with the first confidence score or answer the corresponding question for the user.
[0100] S260, if the user expresses denial to the service prompt tone of the standard question with the second highest confidence score, the service prompt tone of the standard question with the third highest confidence score is played, and step S270 is continued; if the user expresses confirmation, the corresponding service of the standard question with the second highest confidence score is processed for the user or the corresponding question is answered.
[0101] S270, if the user expresses denial to the service prompt tone of the standard question with the third highest confidence score, a prompt tone is played to prompt the user to switch to the button menu and no longer provide intelligent voice navigation service; if the user expresses confirmation, the corresponding service of the standard question with the third highest confidence score is processed for the user or the corresponding question is answered.
[0102] S280, if the user confirms the service prompt tone of the standard question ranked second or third in confidence, the transcribed text of the user's voice and the confirmed standard question are sent to the semantic knowledge base, and the semantic knowledge base adds the voice transcribed text to the extended question of the standard question to improve the score of the re-hit.
[0103] For example, when a user uses the intelligent voice navigation system, he or she can input "I have withdrawn money from my savings card and I want to check the balance" by voice. The transcribed text of the user's voice is parsed through the semantic knowledge base to obtain three candidate standard questions with confidence scores from high to low: 1. Unknown transaction; 2. Savings card status; 3. Balance inquiry.
[0104] The intelligent application navigation system can first play the service prompt tone "You will be able to process unknown transactions", at which point the user can express a negative response. Then, the intelligent application navigation system can continue to play the service prompt tone "You will be able to process savings card status inquiry", at which point the user can express a negative response. Then, the intelligent application navigation system can continue to play the service prompt tone "You will be able to process balance inquiry", at which point the user can express a positive response. The intelligent voice navigation system can process the balance inquiry service for the user through the balance inquiry node of the IVR and return the corresponding service processing result.
[0105] Furthermore, the intelligent application navigation can automatically send the corpus "My savings card has been withdrawn, I check the balance" and the standard question "Balance query" to the semantic knowledge base, so as to add the corpus "My savings card has been withdrawn, I check the balance" to the extended question of the "Balance query" standard question node. In this way, the confidence score of triggering the "Balance query" standard question can be improved the next time the user expresses "My savings card has been withdrawn, I check the balance" again.
[0106] Based on the information processing method provided in the above embodiments, the present application also provides a specific implementation of the information processing device. Please refer to the following embodiments.
[0107] See first Figure 2 The information processing device 300 provided in the embodiment of the present application includes an acquisition module 301, a determination module 302, a sending module 303 and an adding module 304.
[0108] The acquisition module 301 is used to acquire the user's inquiry information, where the inquiry information includes text information generated according to the user's voice input information.
[0109] The determination module 302 is used to determine N candidate standard questions corresponding to the query information from a preset semantic knowledge base, where the semantic knowledge base includes multiple preset standard questions, and the N candidate standard questions are the first N standard questions with the highest confidence scores among the multiple preset standard questions, where N is an integer greater than 1.
[0110] The acquisition module 301 is also used to obtain the selection result of the user on the N candidate standard questions.
[0111] The determination module 302 is further configured to determine the type of the target standard question when the selection result includes a first selection result, wherein the first selection result includes determining the target standard question from N candidate standard questions.
[0112] The sending module 303 is used to send the target standard question and the query information to the semantic knowledge base when the target standard question is not the first standard question, and the first standard question is the candidate standard question with the highest confidence score among the N candidate standard questions.
[0113] The adding module 304 is used to add the query information as an extended question of the target standard question.
[0114] The information processing device of the embodiment of the present application can obtain the user's query information, which includes text information generated according to the user's voice input information. Then, N candidate standard questions corresponding to the query information are determined from the preset semantic knowledge base. The N candidate standard questions are the first N standard questions with the highest confidence scores among the multiple preset standard questions in the semantic knowledge base, and N is an integer greater than 1. In this way, multiple standard questions closest to the user's true intention can be used as candidate standard questions for the user to select. Then the user's selection result for the N candidate standard questions can be obtained. In the case where the selection result includes the first selection result, that is, the user determines the target standard question from the N candidate standard questions, the type of the target standard question is determined. In the case where the target standard question is not the first standard question, the target standard question and the query information are sent to the semantic knowledge base, and the query information is added as an extended question of the target standard question. The first standard question is the candidate standard question with the highest confidence score among the N candidate standard questions. In this way, when the query information is recognized for the next time, the confidence score of the target standard question can be improved because the query information is included in the extended question of the target standard question. In this way, the target standard question and query information can be automatically used for optimization and training of the semantic knowledge base, increasing the probability of hitting the correct intent the next time the user uses the same words to express. As a result, it is possible to effectively correct misidentification of intent, and timely and efficiently perform self-learning optimization of the semantic knowledge base, thereby improving the accuracy of intent recognition of intelligent voice navigation, and thus improving user experience.
[0115] In one embodiment, the apparatus may further include:
[0116] The execution module is used to execute the target operation corresponding to the target standard when the selection result includes the first selection result.
[0117] In one embodiment, the determination module may be specifically used to:
[0118] Perform semantic recognition on the query information to obtain a semantic recognition result.
[0119] The similarities between the question sets corresponding to the plurality of preset standard questions and the semantic recognition results are respectively calculated to obtain a plurality of similarity results corresponding to the plurality of preset standard questions, wherein the question sets include the preset standard questions and the extended questions corresponding to the preset standard questions.
[0120] The confidence scores of the plurality of preset standard questions are calculated respectively based on the similarity results corresponding to the plurality of preset standard questions.
[0121] The first N standard questions with the highest confidence scores among the multiple preset standard questions are determined as N candidate standard questions.
[0122] In one embodiment, the acquisition module is specifically used for:
[0123] Play the prompt tone corresponding to the first candidate standard question, where the first candidate standard question is the standard question with the highest confidence score among the N candidate standard questions.
[0124] Obtain the user's voice selection input for the first candidate standard question.
[0125] When the voice selection input is the target input, the first candidate standard question is determined as the target standard question to obtain a first selection result.
[0126] When the voice selection input is not the target input, the standard question with the highest confidence score among the remaining candidate standard questions is used as the new first candidate standard question, and the prompt sound corresponding to the first candidate standard question is played back until the number of remaining candidate standard questions is 0, and a second selection result is obtained, which includes determining that the target standard question is not included in the N candidate standard questions.
[0127] In one embodiment, the apparatus may further include:
[0128] The playing module is used to play prompt information after obtaining the selection results of the user on the N candidate standard questions, if the selection results include the second selection result, to prompt the user to operate through the touch control system.
[0129] In one embodiment, the acquisition module is specifically used for:
[0130] When the intelligent voice navigation system is in the target state, the user's voice input information is collected.
[0131] Perform voice recognition on the voice input information to obtain inquiry information.
[0132] Each module of the information processing device provided in the embodiment of the present application can realize Figure 1 The functions of each step of the information processing method are provided, and the corresponding technical effects can be achieved. For the sake of concise description, they will not be repeated here.
[0133] Figure 3 A schematic diagram of the hardware structure of an information processing device provided in an embodiment of the present application is shown.
[0134] The information processing device may include a processor 401 and a memory 402 storing computer program instructions.
[0135] Specifically, the processor 401 may include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or may be configured to implement one or more integrated circuits of the embodiments of the present application.
[0136] Memory 402 may include a large capacity memory for data or instructions. By way of example and not limitation, memory 402 may include a hard disk drive (HDD), a floppy disk drive, a flash memory, an optical disk, a magneto-optical disk, a magnetic tape, or a universal serial bus (USB) drive or a combination of two or more of these. In appropriate cases, memory 402 may include a removable or non-removable (or fixed) medium. In appropriate cases, memory 402 may be inside or outside of an integrated gateway disaster recovery device. In a specific embodiment, memory 402 is a non-volatile solid-state memory.
[0137] The memory may include read-only memory (ROM), random access memory (RAM), magnetic disk storage media devices, optical storage media devices, flash memory devices, electrical, optical or other physical / tangible memory storage devices. Thus, typically, the memory includes one or more tangible (non-transitory) computer-readable storage media (e.g., memory devices) encoded with software including computer-executable instructions, and when the software is executed (e.g., by one or more processors), it is operable to perform the operations described with reference to the method according to an aspect of the present disclosure.
[0138] The processor 401 implements any one of the information processing methods in the above embodiments by reading and executing computer program instructions stored in the memory 402 .
[0139] In one example, the information processing device may further include a communication interface 403 and a bus 410. Figure 3 As shown, the processor 401, the memory 402, and the communication interface 403 are connected via a bus 410 and communicate with each other.
[0140] The communication interface 403 is mainly used to implement communication between various modules, devices, units and / or equipment in the embodiments of the present application.
[0141] Bus 410 includes hardware, software or both, and the parts of online data flow billing equipment are coupled to each other. For example, but not limitation, bus may include accelerated graphics port (AGP) or other graphics bus, enhanced industrial standard architecture (EISA) bus, front-side bus (FSB), hypertransport (HT) interconnection, industrial standard architecture (ISA) bus, infinite bandwidth interconnection, low pin count (LPC) bus, memory bus, micro channel architecture (MCA) bus, peripheral component interconnection (PCI) bus, PCI-Express (PCI-X) bus, serial advanced technology attachment (SATA) bus, video electronics standard association local (VLB) bus or other suitable bus or two or more of these combinations. In appropriate cases, bus 410 may include one or more buses. Although the present application embodiment describes and shows a specific bus, the present application considers any suitable bus or interconnection.
[0142] The information processing device can execute the information processing method in the embodiment of the present application, thereby realizing the combination Figure 1 and Figure 2 The information processing method and device described herein. The information processing method in the above embodiments can be implemented by providing a computer storage medium in the present application embodiment. The computer storage medium stores computer program instructions; when the computer program instructions are executed by a processor, any one of the information processing methods in the above embodiments is implemented.
[0143] An embodiment of the present application also provides a computer program product, including a computer program, which implements any one of the information processing methods in the above embodiments when the computer program is processed and executed.
[0144] It should be clear that the present application is not limited to the specific configuration and processing described above and shown in the figures. For the sake of simplicity, a detailed description of the known method is omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present application is not limited to the specific steps described and shown, and those skilled in the art can make various changes, modifications and additions, or change the order between the steps after understanding the spirit of the present application.
[0145] The functional blocks shown in the above-described block diagram can be implemented as hardware, software, firmware or a combination thereof. When implemented in hardware, it can be, for example, an electronic circuit, an application specific integrated circuit (ASIC), appropriate firmware, a plug-in, a function card, etc. When implemented in software, the elements of the present application are programs or code segments that are used to perform the required tasks. The program or code segment can be stored in a machine-readable medium, or transmitted on a transmission medium or a communication link by a data signal carried in a carrier wave. "Machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, optical fiber media, radio frequency (RF) links, etc. The code segment can be downloaded via a computer network such as the Internet, an intranet, etc.
[0146] It should also be noted that the exemplary embodiments mentioned in this application describe some methods or systems based on a series of steps or devices. However, this application is not limited to the order of the above steps, that is, the steps can be performed in the order mentioned in the embodiment, or in a different order from the embodiment, or several steps can be performed simultaneously.
[0147] Aspects of the present disclosure are described above with reference to the flowchart and / or block diagram of the method, device (system) and computer program product according to the embodiment of the present disclosure. It should be understood that each box in the flowchart and / or block diagram and the combination of each box in the flowchart and / or block diagram can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device to produce a machine so that these instructions executed by the processor of the computer or other programmable data processing device enable the implementation of the function / action specified in one or more boxes of the flowchart and / or block diagram. Such a processor can be, but is not limited to, a general-purpose processor, a special-purpose processor, a special application processor, or a field programmable logic circuit. It can also be understood that each box in the block diagram and / or flowchart and the combination of boxes in the block diagram and / or flowchart can also be implemented by dedicated hardware that performs a specified function or action, or can be implemented by a combination of dedicated hardware and computer instructions.
[0148] The above are only specific implementation methods of the present application. Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working processes of the systems, modules and units described above can refer to the corresponding processes in the aforementioned method embodiments, and will not be repeated here. It should be understood that the protection scope of the present application is not limited to this. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed in this application, and these modifications or replacements should be included in the protection scope of this application.
Claims
1. An information processing method, characterized in that: include: Acquire user's inquiry information, wherein the inquiry information includes text information generated according to the user's voice input information; Determine N candidate standard questions corresponding to the query information from a preset semantic knowledge base, wherein the semantic knowledge base includes a plurality of preset standard questions, and the N candidate standard questions are the first N standard questions with the highest confidence scores among the plurality of preset standard questions, where N is an integer greater than 1; Obtaining the user's selection result for the N candidate criteria; In a case where the selection result includes a first selection result, determining a type of a target standard question, wherein the first selection result includes determining the target standard question from the N candidate standard questions; In a case where the target standard question is not the first standard question, sending the target standard question and the query information to the semantic knowledge base, wherein the first standard question is the candidate standard question with the highest confidence score among the N candidate standard questions; The query information is added as an extended question of the target standard question.
2. The information processing method according to claim 1, characterized in that: After obtaining the user's selection result for the N candidate criteria, the method further includes: In a case where the selection results include the first selection result, a target operation corresponding to the target criterion is performed.
3. The information processing method according to claim 1, characterized in that: The determining N candidate standard questions corresponding to the query information from a preset semantic knowledge base includes: Performing semantic recognition on the query information to obtain a semantic recognition result; Respectively calculating similarities between a question set corresponding to the plurality of preset standard questions and the semantic recognition result to obtain a plurality of similarity results corresponding to the plurality of preset standard questions, wherein the question set includes preset standard questions and extended questions corresponding to the preset standard questions; Calculating the confidence scores of the plurality of preset standards respectively according to the similarity results corresponding to the plurality of preset standards; The first N standard questions with the highest confidence scores among the plurality of preset standard questions are determined as the N candidate standard questions.
4. The method according to claim 1, characterized in that: The obtaining of the user's selection result for the N candidate standard questions includes: Playing a prompt tone corresponding to a first candidate standard question, where the first candidate standard question is a standard question with the highest confidence score among the N candidate standard questions; Obtaining a user's voice selection input for the first candidate standard question; In a case where the voice selection input is a target input, determining the first candidate standard question as the target standard question to obtain the first selection result; In the case that the voice selection input is not the target input, the standard question with the highest confidence score among the remaining candidate standard questions is used as the new first candidate standard question, and the prompt sound corresponding to the first candidate standard question is returned until the number of remaining candidate standard questions is 0, and a second selection result is obtained, wherein the second selection result includes determining that the target standard question is not included in the N candidate standard questions.
5. The method according to claim 4, characterized in that After obtaining the user's selection result for the N candidate criteria, the method further includes: In a case where the selection result includes the second selection result, prompt information is played to prompt the user to perform an operation through the touch control system.
6. The information processing method according to claim 1, characterized in that: The obtaining of the user's inquiry information includes: When the intelligent voice navigation system is in the target state, collecting the user's voice input information; Perform voice recognition on the voice input information to obtain the inquiry information.
7. An information processing device, characterized in that: include: An acquisition module, used to acquire the user's inquiry information, wherein the inquiry information includes text information generated according to the user's voice input information; A determination module, configured to determine N candidate standard questions corresponding to the query information from a preset semantic knowledge base, wherein the semantic knowledge base includes a plurality of preset standard questions, and the N candidate standard questions are the first N standard questions with the highest confidence scores among the plurality of preset standard questions, where N is an integer greater than 1; The acquisition module is further used to obtain the user's selection result on the N candidate standard questions; The determining module is further configured to determine a type of a target standard question when the selection result includes a first selection result, wherein the first selection result includes determining the target standard question from the N candidate standard questions; a sending module, configured to send the target standard question and the query information to the semantic knowledge base if the target standard question is not the first standard question, wherein the first standard question is the candidate standard question with the highest confidence score among the N candidate standard questions; An adding module is used to add the query information as an extended question of the target standard question.
8. An information processing device, characterized in that: The device comprises: a processor and a memory storing computer program instructions; when the processor executes the computer program instructions, the information processing method according to any one of claims 1 to 6 is implemented.
9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer program instructions, and when the computer program instructions are executed by a processor, the information processing method according to any one of claims 1 to 6 is implemented.
10. A computer program product, characterized in that When the instructions in the computer program product are executed by a processor of an electronic device, the electronic device is enabled to execute the information processing method according to any one of claims 1 to 6.
Citation Information
Patent Citations
Updating method and device of question answer sequencing model for automatic question answer system
CN104572998A
Question and answer knowledge base updating method and device
CN110928991A
Automatic question and answer processing method and device, computer equipment and storage medium
CN111858859A
Information display method and device, equipment and storage medium
CN113111155A
Question and answer method and device, electronic equipment and storage medium
CN113157867A
