Intelligent dialogue reverse question clarification implementation method and related equipment

Abstract

The invention provides an intelligent dialogue anti-question clarification implementation method. The method comprises the steps that a query statement input by a user is received; key information needing to be clarified in the query statement is identified based on a preset candidate index set, and candidate indexes corresponding to the key information needing to be clarified are determined; outputting the query statement in response to determining that the query statement does not comprise key information needing to be clarified; in response to determining that the query statement comprises at least one piece of key information needing to be clarified, generating at least one problem needing to be clarified based on the at least one piece of key information needing to be clarified and a candidate index corresponding to the key information, and feeding back the problem to the user; and rewriting the query statement based on a clarification result of the user for supplementing the at least one question needing to be clarified to obtain a rewritten query statement. The invention also provides a related device for realizing intelligent dialogue reverse question clarification.

Description

Technical Field

[0001] The present disclosure relates to the field of computer technologies, and in particular, to a method for implementing intelligent dialogue rhetorical question clarification and related devices. Background Art

[0002] Nowadays, natural language processing is one of the most important branches in the field of artificial intelligence. Intelligent dialogue is one of the very classic application scenarios in the field of natural language processing and is also one of the commonly used means of human-computer interaction. An intelligent dialogue can understand a user's question described in natural language and generate a reply content containing an answer or execute a task issued by the user. Rhetorical question clarification is a common dialogue strategy, which can help an intelligent dialogue system more accurately understand the user's intention. However, existing intelligent dialogue systems often lack this ability of rhetorical question clarification, resulting in poor intelligent dialogue effects. Summary of the Invention

[0003] In view of this, embodiments of the present disclosure provide a method for implementing intelligent dialogue rhetorical question clarification and related devices, which can provide an intelligent dialogue system with an efficient ability of rhetorical question clarification, thereby improving the intelligent dialogue efficiency and the user's dialogue experience.

[0004] The method for implementing intelligent dialogue rhetorical question clarification according to the embodiments of the present disclosure may include: receiving a query statement input by a user; identifying key information to be clarified in the query statement based on a pre-set candidate index set, and determining candidate indexes corresponding to the key information to be clarified; in response to determining that the query statement does not include key information to be clarified, outputting the query statement; in response to determining that the query statement includes at least one piece of key information to be clarified, generating at least one question to be clarified based on the at least one piece of key information to be clarified and the candidate indexes corresponding to the key information to be clarified, and feeding the at least one question to be clarified back to the user; rewriting the query statement based on the clarification result supplemented by the user for the at least one question to be clarified, and outputting the rewritten query statement.

[0005] In some embodiments of the present disclosure, identifying the key information to be clarified in the query statement and determining the candidate indicators corresponding to the key information to be clarified based on a preset candidate indicator set includes: determining at least one key information in the query statement input by the user; respectively determining whether the at least one key information is the key information to be clarified based on the preset candidate indicator set; wherein, the key information to be clarified includes: the key information for which no corresponding candidate indicator can be found in the preset candidate indicator set, and / or the key information for which multiple corresponding candidate indicators are found in the preset candidate indicator set; and respectively determining the candidate indicators corresponding to the key information determined to be the key information to be clarified; wherein, for the key information to be clarified for which no corresponding candidate indicator can be found in the preset candidate indicator set, determining that its corresponding candidate indicator is an empty set.

[0006] In some embodiments of the present disclosure, generating at least one question to be clarified based on the at least one key information to be clarified and the candidate indicators corresponding to the key information includes: determining whether the key information corresponds to multiple candidate indicators; in response to determining that the key information corresponds to multiple candidate indicators, obtaining the multiple candidate indicators corresponding to the key information, obtaining an information mapping group of the key information and the multiple candidate indicators, generating a question to be clarified based on the key information in the information mapping group, and generating a checkbox including multiple options based on the multiple candidate indicators in the information mapping group; and in response to determining that the candidate indicator corresponding to the key information is an empty set, generating a question to be clarified based on the key information and generating an input box for the user to input a clarification result.

[0007] Some embodiments of the present disclosure further include: when generating a checkbox including multiple options based on the multiple candidate indicators in the information mapping group, generating an input box for the user to input a clarification result.

[0008] In some embodiments of the present disclosure, rewriting the query statement based on the clarification results supplemented by the user for the at least one question to be clarified includes: respectively obtaining the clarification results for each key information to be clarified, obtaining at least one clarification result information pair composed of the key information to be clarified and the clarification results; and inputting the query statement and the at least one clarification result information pair into a large language model, and the large language model rewrites the query statement based on the at least one clarification result information pair to obtain a rewritten query statement.

[0009] Some embodiments of the present disclosure further include: before identifying the key information to be clarified in the query statement based on a pre-trained query statement understanding model, searching for auxiliary information matching the query statement input by the user from stored general knowledge and / or historical information; and rewriting the query statement based on the auxiliary information.

[0010] In some embodiments of the present disclosure, rewriting the query statement based on the auxiliary information includes: inputting the query statement and the auxiliary information into a large language model, and having the large language model complete the rewriting of the query statement.

[0011] Some embodiments of the present disclosure further include: after obtaining the clarification results supplemented by the user for at least one question to be clarified, storing the obtained clarification result information pairs composed of the key information to be clarified and the clarification results as the historical information.

[0012] Some embodiments of the present disclosure further include: after obtaining the rewritten query statement, returning to the step of identifying the key information to be clarified in the query statement based on a pre-trained query statement understanding model.

[0013] Some embodiments of the present disclosure further include: initializing the clarification count parameter to 0; after obtaining the rewritten query statement, incrementing the clarification count parameter by 1; determining whether the clarification count parameter reaches a pre-set target round of rhetorical clarification; in response to determining that the target round of rhetorical clarification has not been reached, returning to the step of identifying the key information to be clarified in the query statement based on a pre-trained query statement understanding model; and in response to determining that the target round of rhetorical clarification has been reached, outputting the rewritten query statement.

[0014] Corresponding to the above method for realizing intelligent dialogue rhetorical clarification, some embodiments of the present disclosure also disclose an apparatus for realizing intelligent dialogue rhetorical clarification, including:

[0015] A user interface module, configured to receive a query statement input by a user;

[0016] A query statement understanding module, configured to identify the key information to be clarified in the query statement based on a pre-set candidate index set and determine the candidate index corresponding to the key information to be clarified;

[0017] An output module, configured to output the query statement in response to determining that the query statement does not include key information to be clarified;

[0018] A rhetorical clarification module, configured to generate at least one question to be clarified based on the at least one key information to be clarified and the candidate index corresponding to the key information and feedback it to the user in response to determining that the query statement includes at least one key information to be clarified; and

[0019] A rewriting module, configured to rewrite the query statement based on the clarification results supplemented by the user for the at least one question to be clarified and output the rewritten query statement.

[0020] Some embodiments of the present disclosure further include a pre - rewriting module in the intelligent dialogue rhetorical question clarification implementation device. The pre - rewriting module is configured to, before inputting the query statement output by the user interface module into the query statement understanding module, find auxiliary information that matches the query statement input by the user from the stored general knowledge and / or historical information, and rewrite the query statement based on the auxiliary information.

[0021] In addition, an embodiment of the present disclosure further provides an electronic device, including: a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, the above - mentioned intelligent dialogue rhetorical question clarification implementation method is implemented.

[0022] An embodiment of the present disclosure also provides a non - transitory computer - readable storage medium storing computer instructions for causing a computer to execute the above - mentioned intelligent dialogue rhetorical question clarification implementation method.

[0023] An embodiment of the present disclosure also provides a computer program product including computer program instructions. When the computer program instructions run on a computer, the computer is caused to execute the above - mentioned intelligent dialogue rhetorical question clarification implementation method.

[0024] In the above - mentioned intelligent dialogue rhetorical question clarification implementation method and related devices, for a query statement input by a user, key information to be clarified and its corresponding candidate indicators in the query statement can be first identified based on a query statement understanding model, and a question to be clarified can be generated based on the identified information. Then, after obtaining the clarification result supplemented by the user for the question to be clarified, the query statement is rewritten based on the obtained clarification result, so as to provide the ability of rhetorical question clarification for the intelligent dialogue system and improve the intelligent dialogue effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the technical solutions in the present disclosure or related technologies, the following will briefly introduce the drawings required for use in the description of the embodiments or related technologies. Obviously, the following - described drawings are only embodiments of the present disclosure. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0026] Figure 1 Shows the implementation process of the intelligent dialogue rhetorical question clarification implementation method according to some embodiments of the present disclosure.

[0027] Figure 2 Shows the method flow of identifying key information to be clarified in a query statement based on a pre - set candidate indicator set and determining the candidate indicators corresponding to the key information to be clarified according to an embodiment of the present disclosure.

[0028] Figure 3 Shows the method flow of generating questions to be clarified for each key piece of information to be clarified described in the embodiments of the present disclosure.

[0029] Figure 4 Shows an example of the questions to be clarified generated in the embodiments of the present disclosure.

[0030] Figure 5 Shows the method flow of rewriting the query statement based on the clarification results supplemented by the user for at least one of the above questions to be clarified described in the embodiments of the present disclosure.

[0031] Figure 6 Shows the method flow of rewriting the query statement based on the stored general knowledge and / or historical information described in the embodiments of the present disclosure.

[0032] Figure 7 Shows the internal structure of the intelligent dialogue rhetorical question clarification implementation device described in some embodiments of the present disclosure.

[0033] Figure 8 Shows a more specific schematic diagram of the hardware structure of an electronic device described in some embodiments of the present disclosure. Detailed implementation manners

[0034] To make the objectives, technical solutions and advantages of the present disclosure more clear and understandable, the present disclosure will be further described in detail below with reference to specific embodiments and the accompanying drawings.

[0035] It should be noted that unless otherwise defined, the technical terms or scientific terms used in the embodiments of the present disclosure should have the ordinary meanings understood by those of ordinary skill in the art to which the present disclosure belongs. The "first", "second" and similar terms used in the embodiments of the present disclosure do not denote any order, quantity or importance, but are only used to distinguish different components. The terms such as "including" or "comprising" mean that the elements or items appearing before this word cover the elements or items listed after this word and their equivalents, without excluding other elements or items. The terms such as "connected" or "coupled" are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. The terms such as "upper", "lower", "left" and "right" are only used to represent relative positional relationships, and when the absolute position of the object being described changes, the relative positional relationship may also change accordingly.

[0036] It can be understood that before using the technical solutions of the various embodiments of the present disclosure, the types, usage scopes, usage scenarios, etc. of the personal information involved will be informed to the user in an appropriate manner and the user's authorization will be obtained.

[0037] For example, when receiving an active request from a user, a prompt message is sent to the user to clearly prompt the user that the operation requested by the user will require obtaining and using the user's personal information. Thus, the user can autonomously choose whether to provide personal information to software or hardware such as an electronic device, an application program, a server, or a storage medium that performs the operations of the present disclosure technical solution according to the prompt message.

[0038] As an optional but non-limiting implementation manner, the manner of sending a prompt message to the user in response to receiving an active request from the user may be, for example, in the form of a pop-up window, and the prompt message may be presented in text in the pop-up window. In addition, the pop-up window may also carry a selection control for the user to select "agree" or "disagree" to provide personal information to the electronic device.

[0039] It can be understood that the above notification and obtaining user authorization process is only illustrative and does not limit the implementation manner of the present disclosure. Other manners that meet relevant laws and regulations can also be applied to the implementation manner of the present disclosure.

[0040] As mentioned above, the existing intelligent dialogue system lacks the ability to ask rhetorical questions for clarification, resulting in poor intelligent dialogue effects. For this reason, the embodiments of the present disclosure provide an intelligent dialogue rhetorical question clarification implementation method, which is executed by an intelligent dialogue system and can provide the intelligent dialogue system with the ability to ask rhetorical questions for clarification, thereby improving the intelligent dialogue effect of the intelligent dialogue system.

[0041] Figure 1 Shows the implementation process of the intelligent dialogue rhetorical question clarification implementation method described in some embodiments of the present disclosure. As Figure 1 shown, the intelligent dialogue rhetorical question clarification implementation method described in the embodiments of the present disclosure includes the following steps:

[0042] In step 110, a query statement input by the user is received.

[0043] In step 120, based on a pre-set candidate index set, the key information to be clarified in the query statement is identified and the candidate index corresponding to the key information to be clarified is determined.

[0044] In step 130, in response to determining that the query statement does not include key information to be clarified, the query statement is output.

[0045] In step 140, in response to determining that the query statement includes at least one piece of key information to be clarified, at least one question to be clarified is generated based on the at least one piece of key information to be clarified and its corresponding candidate index and fed back to the user.

[0046] In step 150, the query statement is rewritten based on the clarification result supplemented by the user for the at least one question to be clarified, and the rewritten query statement is output.

[0047] It should be noted that in the embodiments of the present disclosure, after performing the above step 150, it is possible to return to the above step 120 to continue the reverse clarification recognition and rewriting of the query statement in the next round until it is determined that the query statement no longer includes the key information to be clarified.

[0048] In addition, as an alternative to the above solution, it is also possible to preset the target round of reverse clarification, for example, 3 times or 5 times. And before executing the above method, initialize a clarification times parameter and set its value to 0. Then, after performing the above step 150, increment the above clarification times parameter by 1 to record the round of query statement rewriting that has been performed. Then, before returning to the above step 120, first determine whether the value of the above clarification times parameter reaches the preset target round of reverse clarification. In response to determining that the preset target round of reverse clarification has not been reached, continue to return to the above step 120 to perform the reverse clarification recognition and rewriting of the query statement in the next round; while in response to determining that the preset target round of reverse clarification has been reached, the rewritten query statement of this round can be output. It can be seen that the purpose of setting the above target round of reverse clarification in the above embodiments is to limit the maximum number of rounds of reverse clarification, thereby improving the execution efficiency and effect of reverse clarification, and avoiding reverse clarifying the user too many times, which may affect the user's dialogue experience.

[0049] In addition, in the above method for realizing reverse clarification in intelligent dialogue, when outputting the query statement input by the user or the rewritten query statement, the query statement can be first converted into a domain-specific language (DSL), and then output.

[0050] The following will further illustrate each step of the above method for realizing reverse clarification in intelligent dialogue in detail with specific examples.

[0051] In the above step 110, the query statement input by the user usually refers to the actual question or task input by the user. In the field of intelligent dialogue, the above query statement is also called a Query. Usually, the query statement or Query input by the user may contain some fuzzy or ambiguous semantics. Therefore, if the query statement input by the user is directly used for information retrieval, a large amount of inaccurate information may be obtained, thereby reducing the accuracy of the intelligent dialogue output and resulting in a poor dialogue experience for the user. Therefore, when the semantics expressed by the user is unclear, reverse clarification can be performed to clarify the user's intention.

[0052] Regarding the above step 120, in an embodiment of the present disclosure, the above-mentioned pre-set candidate index set usually includes a number of pre-set candidate indexes. These candidate indexes usually represent the expressions with clear meanings of the key information that users are concerned about in certain specific business fields, that is, the standard names of the indexes that users usually care about in certain specific business fields. For example, in the e-commerce field, the candidate indexes related to the number of people may include: the number of transactions, the number of refunds, and the number of payments, etc. For another example, in the e-commerce field, the candidate indexes related to the amount may include: the transaction amount, the refund amount, and the payment amount, etc. Using the above-mentioned pre-set candidate index set as a benchmark can assist in identifying whether the query statement input by the user contains key information that needs to be clarified, that is, it can determine whether the semantics of the query statement input by the user is clear.

[0053] Specifically, Figure 2 shows the method for identifying the key information that needs to be clarified in the query statement based on the pre-set candidate index set described in the embodiment of the present disclosure and determining the candidate index corresponding to the key information that needs to be clarified. That is Figure 2 shows the implementation method of the above step 120. As Figure 2 shown, the above step 120 may include:

[0054] In step 210, determine at least one piece of key information in the query statement input by the user.

[0055] It can be understood that not all content in the query statement input by the user contains key information. Therefore, in order to help the computer understand the content of the user's query and find relevant information in the database or information resources, it is very important to extract key information from the query statement. Usually, the extraction of key information in the query statement is an important application of natural language processing (NLP) and can be implemented through NLP technology.

[0056] In step 220, based on the pre-set candidate index set, determine whether each of the at least one piece of key information is key information that needs to be clarified.

[0057] Specifically, in an embodiment of the present disclosure, the above-mentioned key information that needs to be clarified usually may include: the key information for which no corresponding candidate index can be found in the above-mentioned pre-set candidate index set; and / or the key information for which multiple corresponding candidate indexes are found in the above-mentioned pre-set candidate index set.

[0058] For example, the query statement entered by the user is "How about the overview of the data dashboard tonight?". Among them, by understanding the query statement, one of the key information in the above query statement can be obtained as "overview of the dashboard". If no corresponding candidate indicator for "overview of the dashboard" can be found in the above pre-set candidate indicator set, then "overview of the dashboard" belongs to the key information that needs to be clarified. At this time, a rhetorical clarification is required to ask the user to supplement what specific indicator "overview of the dashboard" refers to. Another example, the query statement entered by the user is "Query the amount in the past three months". Among them, by understanding the query statement, one of the key information in the above query statement can be obtained as "amount". If multiple candidate indicators corresponding to "amount" can be found in the above pre-set candidate indicator set, such as including: "transaction amount", "refund amount", "sales amount", etc., then "amount" also belongs to the key information that needs to be clarified. At this time, a rhetorical clarification is also required to ask the user to select the target candidate indicator corresponding to "amount" from the above multiple corresponding indicators. Another example, the query statement entered by the user is "Query the sales amount in the past three months". Among them, by understanding the query statement, one of the key information in the above query statement can be obtained as "sales amount" or ask the user to supplement what specific indicator "amount" refers to. If there is exactly one candidate indicator corresponding to "sales amount" in the above pre-set candidate indicator set, that is, "sales amount" itself, it means that the meaning of the key information in the query statement is clear, so no rhetorical clarification is required.

[0059] In step 230, for the key information determined to need clarification, their corresponding candidate indicators are respectively determined.

[0060] Specifically, in the embodiments of the present disclosure, for the key information that needs to be clarified and for which no corresponding candidate indicator can be found in the above pre-set candidate indicator set, its corresponding candidate indicator can be determined to be an empty set; while for the key information that needs to be clarified and for which multiple corresponding candidate indicators are found in the above pre-set candidate indicator set, its corresponding candidate indicators can be determined to be the above-found multiple corresponding candidate indicators.

[0061] In some embodiments of the present disclosure, the above Figure 2 The shown method can be implemented by a trained query statement understanding model. The above query statement understanding model can be a large language model trained specifically, mainly used to identify the key information that needs to be clarified in the above query statement based on a pre-set candidate indicator set. In these embodiments, the advantages of the large language model in natural language understanding can be utilized to identify and analyze the key information of the query statement, so as to accurately identify the key information that needs to be clarified in the query statement.

[0062] Regarding the above step 140, in some embodiments of the present disclosure, Figure 3It shows the method flow of generating questions to be clarified for each key piece of information to be clarified described in the embodiments of the present disclosure. As Figure 3 shown, the method of generating questions to be clarified for each key piece of information to be clarified can specifically perform the following steps:

[0063] In step 310, determine whether the key information corresponds to multiple candidate indicators.

[0064] In step 320, in response to determining that the above key information corresponds to multiple candidate indicators, determine the information mapping group corresponding to the above key information.

[0065] Wherein, the above information mapping group includes the mapping relationship between the key information and the candidate indicators.

[0066] In step 330, generate a question to be clarified based on the key information in the above information mapping group, and generate a checkbox including multiple options based on the multiple candidate indicators in the above information mapping group.

[0067] In the above step 330, in addition to the above checkbox, an input box for the user to input the clarification result can be further generated to suit the situation where all candidate indicators do not meet the user's requirements.

[0068] In step 340, in response to determining that the candidate indicator corresponding to the above key information is an empty set, generate a question to be clarified based on the above key information, and generate an input box for the user to input the clarification result.

[0069] Based on the foregoing example, assume that the key information to be clarified in the query statement obtained through the query statement understanding model is "overall market overview". Since "overall market overview" cannot find a corresponding candidate indicator in the above pre-set candidate indicator set, in the above step 340, a question to be clarified "May I ask what 'overall market overview' represents?" can be generated based on the key information to be clarified as "overall market overview". At the same time, an input box for the user to input the clarification result is generated. Another example, assume that the key information to be clarified in the query statement obtained through the query statement understanding model is "amount". Since "amount" corresponds to multiple candidate indicators: transaction amount, refund amount, sales amount. Then in the above step 330, a question to be clarified "May I ask what 'amount' represents?" can be generated based on the key information to be clarified as "amount". Further, in the above step 330, a checkbox including three options of "transaction amount", "refund amount", and "sales amount" is generated based on the multiple candidate indicators corresponding to "amount": "transaction amount", "refund amount", and "sales amount". Furthermore, an input box for the user to input the clarification result is also generated.

[0070] Figure 4 It shows an example of the question to be clarified generated in the embodiments of the present disclosure. In Figure 4In the example shown, there are a total of two issues to be clarified: Issue 410 and Issue 420. Among them, Issue 410 includes: the content 411 of Issue 410 and the input box 412 for the user to input the clarification result; Issue 420 includes: the content 421 of Issue 420, the checkbox 422 for the user to select the clarification result, and the input box 423 for the user to input the clarification result.

[0071] Regarding the above step 150, Figure 5 It shows the method flow for rewriting a query statement based on the clarification results supplemented by the user for the above at least one issue to be clarified according to the embodiments of the present disclosure. As Figure 5 shown, the method for rewriting the query statement may specifically include the following steps:

[0072] Step 510, respectively obtain the clarification results for each key piece of information to be clarified, and obtain at least one pair of clarification result information composed of the key piece of information to be clarified and the clarification result.

[0073] Step 520, input the query statement and the above at least one pair of clarification result information into a large language model, and the large language model rewrites the query statement based on the above at least one pair of clarification result information to obtain the rewritten query statement.

[0074] In the above embodiment, the advantages of the large language model in natural language processing can be utilized to rewrite the key piece of information to be clarified in the query statement based on the input pair of clarification result information, so that the key piece of information with ambiguous semantics in the query statement can be rewritten into a more accurate expression.

[0075] Based on the foregoing example, assuming that for the key piece of information to be clarified "amount", and the candidate index corresponding to "amount" selected by the user in the checkbox is "transaction amount", then in the above step 510, a pair of clarification result information "amount" - "transaction amount" will be established. Then in step 520, input the query statement "Query the amount in the past three months" and the pair of clarification result information "amount" - "transaction amount" into the large language model, and the large language model rewrites the query statement, so that the rewritten query statement "Query the transaction amount in the past three months" can be obtained. It can be seen that the semantics of the rewritten query statement is clearer. In this way, more accurate results can be obtained by recalling information with the modified query statement, thereby greatly improving the efficiency and effect of intelligent dialogue and enhancing the user's dialogue experience.

[0076] It can be seen that in the embodiments of the present disclosure, for a query statement input by a user, the key information to be clarified in the query statement and its corresponding candidate metrics can be first identified based on a query statement understanding model, and a question to be clarified can be generated based on the identified information. Then, after obtaining the clarification result supplemented by the user for the question to be clarified, the query statement can be rewritten based on the obtained clarification result, so as to provide the intelligent dialogue system with the ability of efficient rhetorical clarification and improve the intelligent dialogue effect.

[0077] To further enhance the implementation effect of rhetorical clarification, before performing step 120 above to identify the key information to be clarified in the query statement based on a pre-trained query statement understanding model, the embodiments of the present disclosure may further include: rewriting the query statement based on stored general knowledge and / or historical information to obtain a rewritten query statement.

[0078] Specifically, Figure 6 shows the method flow of rewriting a query statement based on stored general knowledge and / or historical information according to the embodiments of the present disclosure. As Figure 6 shown, the method of rewriting a query statement may specifically include the following steps:

[0079] In step 610, auxiliary information matching the query statement input by the user is found from the stored general knowledge and / or historical information.

[0080] In step 620, the query statement input by the user is rewritten based on the above auxiliary information.

[0081] Specifically, in the embodiments of the present disclosure, the above general knowledge usually may include the corresponding relationships between common sayings (also known as industry jargon) and standard terms in a specific business field, question-answer data pairs (also known as QA pairs) in a specific business field, and public business knowledge such as documents.

[0082] In the embodiments of the present disclosure, the above historical information may include: historical information related to the user stored in the intelligent dialogue system. For example, the historical intelligent dialogue information of the user stored in the intelligent dialogue system and the clarification results submitted by the user for rhetorical clarification in the past, etc.

[0083] In this way, in the above step 610, based on a similarity matching strategy, auxiliary information with a high similarity to the query statement input by the user can be obtained from the stored general knowledge and / or historical information. For example, in an application, one or more pieces of information with a similarity greater than a pre-set threshold to the query statement input by the user or one or more pieces of information with the highest similarity ranking can be recalled from the stored general knowledge and / or historical information through an information recall method as the above auxiliary information.

[0084] After obtaining one or more pieces of auxiliary information, in step 620 above, the query statement input by the user and the one or more pieces of auxiliary information can be input into the large language model, and the large language model can complete the rewriting of the query statement input by the user. It should be noted that the large language model mentioned above can be the same large language model as the one used in step 150 above.

[0085] In a specific example, assume that the query statement input by the user is "Query the transaction situation in the past 3 months", and in the general knowledge, it is stored that "transaction situation" is a common saying in a certain business field, and its corresponding standard term is "transaction amount and number of transactions". In this way, through step 610 above, the information recalled from the general knowledge corresponding to "transaction situation" is "transaction amount and number of transactions", and the recalled information is used as the above-mentioned auxiliary information. Thus, based on the above-mentioned auxiliary information, the large language model can modify the query statement input by the user to "Query the transaction amount and number of transactions in the past 3 months".

[0086] In another specific example, assume that the query statement input by the user is "What about the month before last?". In this way, through step 610 above, the intelligent dialogue information of the user's history stored in the intelligent dialogue system can be queried using the above query statement, and the historical record that is close in time and has the highest similarity - "Query the sales data for this month" - is obtained as the above-mentioned auxiliary information. Thus, based on the above-mentioned auxiliary information, the large language model can modify the query statement input by the user to "Query the sales data for the month before last".

[0087] In yet another specific example, assume that the query statement input by the user is "Query the amount in the past three months". In this way, through step 610 above, the clarification result information pair obtained through rhetorical clarification with the highest similarity - "amount" - "transaction amount" - can be found by querying the clarification results submitted for rhetorical clarification stored in the intelligent dialogue system's history using the above query statement, and it is used as the above-mentioned auxiliary information. Thus, based on the above-mentioned auxiliary information, the large language model can modify the query statement input by the user to "Query the transaction amount in the past three months".

[0088] It can be seen from this that by implementing the method of steps 610 - 620 above, before performing rhetorical clarification, the query statement input by the user can be automatically rewritten in advance according to general knowledge and historical information, so as to effectively utilize general knowledge and historical information to automatically eliminate some semantically ambiguous or ambiguous content in the query statement, thereby improving the quality and effectiveness of the query statement, effectively reducing the number of key information that needs to be rhetorically clarified or the number of rounds of rhetorical clarification, improving the service efficiency of the intelligent dialogue system, and at the same time improving the user experience of intelligent dialogue.

[0089] It can be seen that in order to achieve the aboveFigure 6 The method shown, in some embodiments of the present disclosure, at Figure 1 In the method shown, after obtaining the clarification results supplemented by the user for at least one question to be clarified, it is possible to further store the obtained pair of clarification result information composed of the key information to be clarified and the clarification results. In this way, the above pair of clarification result information will be used as historical information to automatically rewrite the query statement input by the user during subsequent intelligent conversations. The above storage step can be executed before or after the above step 150, and the embodiments of the present disclosure do not make any limitations.

[0090] Corresponding to the above intelligent conversation rhetorical clarification implementation method, some embodiments of the present disclosure also disclose an intelligent conversation rhetorical clarification implementation device. Figure 7 Shows the internal structure of the intelligent conversation rhetorical clarification implementation device described in some embodiments of the present disclosure. As Figure 7 Described above, the above intelligent conversation rhetorical clarification implementation device may include:

[0091] A user interface module 710, configured to receive a query statement input by a user;

[0092] A query statement understanding module 720, configured to identify the key information to be clarified in the query statement based on a pre-set candidate index set and determine the candidate index corresponding to the key information to be clarified;

[0093] An output module 730, configured to output the query statement in response to determining that the query statement does not include key information to be clarified;

[0094] A rhetorical clarification module 740, configured to generate at least one question to be clarified based on the at least one key information to be clarified and the candidate index corresponding to the key information and feedback it to the user in response to determining that the query statement includes at least one key information to be clarified; and

[0095] A rewriting module 750, configured to rewrite the query statement based on the clarification results supplemented by the user for the at least one question to be clarified to obtain a rewritten query statement.

[0096] Furthermore, in some embodiments of the present disclosure, the above intelligent conversation rhetorical clarification implementation device may further include: a pre-rewriting module 760, configured to search for auxiliary information matching the query statement input by the user from stored general knowledge and / or historical information before inputting the query statement output by the user interface module into the query statement understanding module, and rewrite the query statement based on the auxiliary information.

[0097] It can be seen that by using the above intelligent dialogue rhetorical question clarification implementation device, for the query statement input by the user, the key information to be clarified and its corresponding candidate metrics in the query statement can be first identified based on the query statement understanding model, and a question to be clarified can be generated based on the identified information. Then, after obtaining the clarification result supplemented by the user for the question to be clarified, the query statement can be rewritten based on the obtained clarification result, so as to provide the intelligent dialogue system with the ability of rhetorical question clarification and improve the intelligent dialogue effect.

[0098] Based on the same inventive concept, corresponding to the method of any of the above embodiments, the present disclosure further provides an electronic device, including: a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein when the processor executes the program, it implements the intelligent dialogue rhetorical question clarification implementation method of any of the above embodiments.

[0099] Figure 8 FIG. shows a schematic hardware structure diagram of a more specific electronic device provided in this embodiment. The device may include: a processor 2010, a memory 2020, an input / output interface 2030, a communication interface 2040, and a bus 2050. Among them, the processor 2010, the memory 2020, the input / output interface 2030, and the communication interface 2040 are communicatively connected to each other inside the device through the bus 2050.

[0100] The processor 2010 may be implemented in a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, etc., and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this specification.

[0101] The memory 2020 may be implemented in the form of a ROM (Read Only Memory), a RAM (Random Access Memory), a static storage device, a dynamic storage device, etc. The memory 2020 may store an operating system and other application programs. When implementing the technical solutions provided in the embodiments of this specification through software or firmware, the relevant program codes are stored in the memory 2020 and are called and executed by the processor 2010.

[0102] The input / output interface 2030 is used to connect input / output devices to achieve information input and output. Among them, the input / output devices can be configured as components in the device or externally connected to the device to provide corresponding functions. The input devices can include microphones, various sensors, etc., and the output devices can include displays, speakers, vibrators, indicator lights, etc.

[0103] The communication interface 2040 is used to connect a communication module (not shown in the figure) to achieve communication interaction between this device and other devices. The communication module can achieve communication through wired means (such as USB, network cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.).

[0104] The bus 2050 includes a path for transmitting information between various components of the device (such as the processor 2010, the memory 2020, the input / output interface 2030, and the communication interface 2040).

[0105] It should be noted that although the above device only shows the processor 2010, the memory 2020, the input / output interface 2030, the communication interface 2040, and the bus 2050, in the specific implementation process, the device may also include other components necessary for normal operation. In addition, those skilled in the art can understand that the above device may also only include the components necessary to implement the solution of the embodiments of this specification, and do not have to include all the components shown in the figure.

[0106] The electronic device in the above embodiment is used to implement the corresponding intelligent dialogue rhetorical question clarification implementation method in any of the foregoing embodiments, and has the beneficial effects of the corresponding method embodiments, which will not be elaborated here.

[0107] Based on the same inventive concept, corresponding to the method in any of the above embodiments, the present disclosure also provides a non-transitory computer-readable storage medium storing computer instructions for causing the computer to execute the intelligent dialogue rhetorical question clarification implementation method as described in any of the foregoing embodiments.

[0108] The computer-readable medium of this embodiment includes permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. The information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassette tapes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information that can be accessed by a computing device.

[0109] The computer instructions stored in the storage medium of the above embodiment are used to cause the computer to execute the task processing method described in any of the above embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be elaborated here.

[0110] Those of ordinary skill in the art should understand that: the discussion of any of the above embodiments is only exemplary and is not intended to imply that the scope of the present disclosure (including the claims) is limited to these examples; under the concept of the present disclosure, the technical features in the above embodiments or different embodiments can also be combined, the steps can be implemented in any order, and there are many other variations in different aspects of the embodiments of the present disclosure as described above, which are not provided in detail for the sake of brevity.

[0111] In addition, for simplicity of description and discussion, and so as not to make the embodiments of the present disclosure difficult to understand, the well-known power / ground connections to integrated circuit (IC) chips and other components may or may not be shown in the provided drawings. In addition, the devices may be shown in block diagram form so as not to make the embodiments of the present disclosure difficult to understand, and this also takes into account the fact that the details of the implementation of these block diagram devices are highly dependent on the platform on which the embodiments of the present disclosure are to be implemented (i.e., these details should be completely within the understanding of those skilled in the art). In the case where specific details (such as circuits) are set forth to describe the exemplary embodiments of the present disclosure, it will be apparent to those skilled in the art that the embodiments of the present disclosure can be implemented without these specific details or with variations of these specific details. Therefore, these descriptions should be considered illustrative rather than restrictive.

[0112] Although the present disclosure has been described in connection with specific embodiments thereof, many alternatives, modifications, and variations of these embodiments will be apparent to those of ordinary skill in the art in light of the foregoing description. For example, other memory architectures (e.g., dynamic RAM (DRAM)) may be used with the embodiments discussed.

[0113] Embodiments of the present disclosure are intended to cover all such alternatives, modifications, and variations that fall within the broad scope of the appended claims. Accordingly, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the embodiments of the present disclosure shall be included within the protection scope of the present disclosure.

Claims

1. A method for realizing intelligent dialogue question clarification, comprising: Receive query statements input by users; Identify key information to be clarified in the query statement based on a preset candidate indicator set and determine candidate indicators corresponding to the key information to be clarified; In response to determining that the query statement does not include key information that needs to be clarified, outputting the query statement; In response to determining that the query statement includes at least one key information to be clarified, generating at least one question to be clarified based on the at least one key information to be clarified and a candidate indicator corresponding to the key information and feeding back the question to the user; as well as The query statement is rewritten based on the clarification result supplemented by the user for the at least one question requiring clarification, and the rewritten query statement is output.

2. The method according to claim 1, wherein: Identifying the key information to be clarified in the query statement based on a preset candidate indicator set and determining the candidate indicators corresponding to the key information to be clarified include: Determining at least one key information in the query statement input by the user; Determining whether the at least one key information is key information to be clarified based on a preset candidate indicator set; wherein the key information to be clarified includes: key information for which a corresponding candidate indicator cannot be found in the preset candidate indicator set, and / or key information for which a plurality of corresponding candidate indicators are found in the preset candidate indicator set; and For the key information determined to be to be clarified, the corresponding candidate indicators are determined respectively; wherein, for the key information to be clarified for which the corresponding candidate indicator cannot be found in the preset candidate indicator set, the corresponding candidate indicator is determined to be an empty set.

3. The method according to claim 2, wherein: Generating at least one question to be clarified based on the at least one key information to be clarified and the candidate indicator corresponding to the key information includes: Determining whether the key information corresponds to multiple candidate indicators; In response to determining that the key information corresponds to a plurality of candidate indicators, acquiring a plurality of candidate indicators corresponding to the key information, obtaining an information mapping group of the key information and the plurality of candidate indicators, generating a question to be clarified based on the key information in the information mapping group, and generating a check box including a plurality of options based on the plurality of candidate indicators in the information mapping group; and In response to determining that the candidate indicators corresponding to the key information are an empty set, generating a question to be clarified based on the key information and generating an input box for a user to input a clarification result.

4. The method according to claim 3, wherein: Further including: When a check box including a plurality of options is generated based on a plurality of candidate indicators in the information mapping group, an input box for a user to input a clarification result is generated.

5. The method according to claim 1, wherein: Rewriting the query statement based on the clarification result supplemented by the user for the at least one question to be clarified includes: respectively obtaining clarification results for each key information to be clarified, and obtaining at least one clarification result information pair consisting of the key information to be clarified and the clarification result; and The query statement and the at least one clarification result information pair are input into a large language model, and the large language model rewrites the query statement based on the at least one clarification result information pair to obtain a rewritten query statement.

6. The method according to claim 1, further comprising: Before identifying key information to be clarified in the query sentence based on the pre-trained query sentence understanding model, searching for auxiliary information matching the query sentence input by the user from the stored general knowledge and / or historical information; as well as The query statement is rewritten based on the auxiliary information.

7. The method according to claim 6, wherein: Rewriting the query statement based on the auxiliary information includes: inputting the query statement and the auxiliary information into a large language model, and using the large language model to complete the rewriting of the query statement.

8. The method according to claim 7, further comprising: After obtaining the clarification result supplemented by the user for at least one question to be clarified, the obtained clarification result information pair consisting of the key information to be clarified and the clarification result is stored as the historical information.

9. The method according to claim 1, further comprising: After obtaining the rewritten query statement, the process returns to the step of identifying key information to be clarified in the query statement based on the pre-trained query statement understanding model.

10. The method according to claim 1, further comprising: Initialize the clarification times parameter to 0; After obtaining the rewritten query statement, the clarification times parameter is increased by 1; Determine whether the clarification number parameter reaches a preset target round of counter-question clarification; In response to determining that the counter-question clarification target round has not been reached, returning to the step of identifying key information to be clarified in the query sentence based on the pre-trained query sentence understanding model; as well as In response to determining that the counter-question clarification target round has been reached, the rewritten query statement is output.

11. An intelligent dialogue questioning and clarification implementation device, comprising: A user interface module, used for receiving a query statement input by a user; A query statement understanding module, used to identify key information to be clarified in the query statement based on a preset candidate indicator set and determine candidate indicators corresponding to the key information to be clarified; An output module, configured to output the query statement in response to determining that the query statement does not include key information to be clarified; a counter-question clarification module, configured to, in response to determining that the query statement includes at least one key information to be clarified, generate at least one question to be clarified based on the at least one key information to be clarified and a candidate indicator corresponding to the key information, and feed back the question to the user; as well as The rewriting module is used to rewrite the query statement based on the clarification result supplemented by the user for the at least one question to be clarified, and output the rewritten query statement.

12. The apparatus according to claim 11, further comprising: A pre-rewriting module is used to search for auxiliary information matching the query statement input by the user from the stored general knowledge and / or historical information before inputting the query statement output by the user interface module into the query statement understanding module, and rewrite the query statement based on the auxiliary information.

13. An electronic device, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method for realizing intelligent dialogue question clarification as described in any one of claims 1 to 10 is implemented.

14. A non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to enable a computer to execute the intelligent dialogue question clarification implementation method as described in any one of claims 1-10.

15. A computer program product, comprising computer program instructions, which, when executed on a computer, enable the computer to execute the method for implementing intelligent dialogue questioning and clarification as claimed in any one of claims 1 to 10.

Images