Method, apparatus and record generation system for assisting in generating a record

Abstract

The application provides a method, device and record generation system for assisting in generating a record. The method comprises: obtaining historical information in a historical interrogation process, the target historical information comprising a historical object element, a historical objective element, a historical subject element and a historical subjective element; constructing an artificial intelligence model; obtaining current information in a current interrogation process; extracting target current information from the current information by using the artificial intelligence model, and generating a record according to the target current information, the target current information comprising a current object element, a current objective element, a current subject element and a current subjective element. In the scheme, the historical information and the corresponding target historical information are trained by constructing the artificial intelligence model, the target current information can be automatically extracted in the current interrogation process by using the artificial intelligence model, the structured expression of the target current information is formed, the record is generated, and the credibility of the generated record is high.

Description

Technical Field

[0001] This application relates to the fields of speech processing and natural language processing, and more specifically, to a method, apparatus, and system for assisting in the generation of transcripts. Background Technology

[0002] The transcript mainly records the statements of the respondent and is an important basis for case description and characterization. However, the current transcript system cannot automatically extract target information, nor can it perform structured analysis of the transcript content according to the target information. When cross-referencing and predicting the cause of action, it relies on the wording, word frequency, or word vector of the transcript text, without utilizing the effective information in the transcript. It can only assist the questioner in manually recording through fixed prompts. Due to the inability to automatically extract target information, the credibility of the transcript is low. Summary of the Invention

[0003] The main purpose of this application is to provide a method, apparatus and system for assisting in the generation of transcripts, so as to solve the problem that the target information cannot be automatically extracted and the credibility of transcripts is low in the prior art.

[0004] According to one aspect of the present invention, a method for assisting in the generation of a transcript is provided, comprising: acquiring historical information during a historical inquiry process, wherein the historical information includes historical voice information and / or historical text information, and the historical voice information and / or the historical text information includes at least target historical information, wherein the target historical information includes historical object elements, historical objective elements, historical subject elements, and historical subjective elements; constructing an artificial intelligence model, wherein the artificial intelligence model is trained using multiple sets of training data, each set of training data including the historical information and the target historical information corresponding to the historical information; acquiring current information during the current inquiry process, wherein the current information includes current voice information and / or current text information; using the artificial intelligence model to extract target current information from the current information, and generating a transcript based on the target current information, wherein the target current information includes current object elements, current objective elements, current subject elements, and current subjective elements.

[0005] Optionally, constructing an artificial intelligence model includes: extracting the target historical information from the historical information, training the model using the historical information and the target historical information, and obtaining multiple artificial intelligence models, wherein each artificial intelligence model is trained using multiple sets of training data, and each set of training data includes the historical information of a case of a certain case type and the target historical information corresponding to the historical information.

[0006] Optionally, the process of extracting target current information from the current information using the artificial intelligence model and generating a record based on the target current information includes: determining the case type corresponding to the current information; determining the artificial intelligence model corresponding to the current information based on the case type; extracting the target current information from the current information using the artificial intelligence model corresponding to the current information; and generating the record based on the target current information.

[0007] Optionally, when there are multiple pieces of target current information, generating a record based on the target current information includes: acquiring multiple pieces of target current information; sorting the multiple pieces of target current information in chronological order to obtain sorted multiple pieces of target current information; converting the sorted multiple pieces of target current information to obtain target text information; adding the target text information to a target template to generate the record.

[0008] Optionally, after converting the sorted current information of multiple targets to obtain target text information, adding the target text information to the target template, and generating the record, the method further includes: determining whether the record conforms to a predetermined format; if the record does not conform to the predetermined format, generating a first prompt message; determining whether the four elements of the record are complete; if the four elements of the record are incomplete, generating a second prompt message.

[0009] Optionally, there are multiple records, and each record is generated from the current information of a target object during the current query process. After converting the sorted current information of multiple targets to obtain target text information, adding the target text information to the target template, and generating the record, the method further includes: obtaining the similarity between multiple records; generating a third prompt message if the similarity is greater than or equal to a similarity threshold; and generating a fourth prompt message if the similarity is less than the similarity threshold.

[0010] Optionally, after obtaining the current information in the current query process, the method further includes: determining the language of the current information; if the language is not the target language, performing speech recognition on the current speech information in the current information, performing machine translation on the current text information in the current information, and converting the current information in the non-target language into the current information in the target language.

[0011] Optionally, after generating the record based on the target current information, the method further includes: sending the record to a display device so that the display device displays the record.

[0012] According to another aspect of the present invention, an apparatus for assisting in the generation of transcripts is also provided, comprising: a first acquisition unit, configured to acquire historical information during a historical inquiry process, the historical information including historical voice information and / or historical text information, wherein the historical voice information and / or the historical text information includes at least target historical information, the target historical information including historical object elements, historical objective elements, historical subject elements, and historical subjective elements; a construction unit, configured to construct an artificial intelligence model, wherein the artificial intelligence model is trained using multiple sets of training data, each set of training data including the historical information and the target historical information corresponding to the historical information; a second acquisition unit, configured to acquire current information during the current inquiry process, the current information including current voice information and / or current text information; and a first processing unit, configured to extract target current information from the current information using the artificial intelligence model, and generate a transcript based on the target current information, the target current information including current object elements, current objective elements, current subject elements, and current subjective elements.

[0013] According to another aspect of the present invention, a computer-readable storage medium is also provided, the computer-readable storage medium including a stored program, wherein, when the program is executed, it controls the device where the computer-readable storage medium is located to perform any of the methods described.

[0014] According to another aspect of the present invention, a processor is also provided, the processor being configured to run a program, wherein the program, when running, executes any one of the methods described.

[0015] According to another aspect of the present invention, a transcript generation system is also provided, comprising: a microphone, a display device, and an apparatus for assisting in transcript generation, wherein the apparatus for assisting in transcript generation communicates with the microphone and the display device respectively, and the apparatus for assisting in transcript generation is used to perform any of the methods described.

[0016] In this embodiment of the invention, historical information from previous query processes is first acquired, then an artificial intelligence model is constructed, followed by the acquisition of current information from the current query process. Finally, the artificial intelligence model is used to extract target current information from the current information, and a transcript is generated based on the target current information. In this scheme, by constructing an artificial intelligence model and training it on historical information and corresponding target historical information, the target current information can be automatically extracted during the current query process, forming a structured representation of the target current information and generating a transcript with high reliability. Attached Figure Description

[0017] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings:

[0018] Figure 1 A flowchart illustrating a method for assisting in the generation of transcripts according to an embodiment of this application is shown;

[0019] Figure 2 A schematic diagram of the structure of an apparatus for assisting in the generation of transcripts according to an embodiment of this application is shown. Detailed Implementation

[0020] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0021] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0022] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate for the embodiments of this application described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0023] It should be understood that when an element (such as a layer, film, region, or substrate) is described as being "on" another element, the element may be directly on the other element, or there may be an intermediate element present. Furthermore, in the specification and claims, when an element is described as being "connected" to another element, the element may be "directly connected" to the other element, or "connected" to the other element via a third element.

[0024] As mentioned in the background section, existing technologies cannot automatically extract target information, resulting in low credibility of transcripts. To address these issues, one embodiment of this application provides a method, apparatus, and transcript generation system for assisting in transcript generation.

[0025] According to an embodiment of this application, a method for assisting in the generation of transcripts is provided. Figure 1 This is a flowchart of a method for generating a record according to an embodiment of this application. For example... Figure 1 As shown, the method includes the following steps:

[0026] Step S101: Obtain historical information during the historical query process. The historical information includes historical voice information and / or historical text information. The historical voice information and / or the historical text information includes at least target historical information. The target historical information includes historical object elements, historical objective elements, historical subject elements, and historical subjective elements.

[0027] Step S102: Construct an artificial intelligence model, wherein the artificial intelligence model is trained using multiple sets of training data, and each set of training data includes the aforementioned historical information and the aforementioned target historical information corresponding to the aforementioned historical information.

[0028] Specifically, artificial intelligence models can be linear regression models, logistic regression models, decision tree models, leaf Bayes models, deep neural network models, and so on.

[0029] Step S103: Obtain current information during the current query process, including current voice information and / or current text information;

[0030] Step S104: Use the above-mentioned artificial intelligence model to extract target current information from the above-mentioned current information, and generate a record based on the above-mentioned target current information. The above-mentioned target current information includes current object elements, current objective elements, current subject elements, and current subjective elements.

[0031] The method described above first acquires historical information from past inquiry processes, then constructs an artificial intelligence model, subsequently acquires current information from the current inquiry process, and finally uses the artificial intelligence model to extract target current information from the current information, generating a transcript based on this target current information. This approach trains the artificial intelligence model on historical information and its corresponding target historical information, enabling the automatic extraction of target current information during the current inquiry process. This results in a structured representation of the target current information, generating a transcript with high credibility.

[0032] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0033] In one embodiment of this application, constructing an artificial intelligence model includes: extracting the target historical information from the aforementioned historical information; training the model using the aforementioned historical information and the target historical information to obtain multiple artificial intelligence models. Each of these artificial intelligence models is trained using multiple sets of training data, and each set of training data includes the aforementioned historical information of a case type and the target historical information corresponding to that historical information. In this embodiment, corresponding artificial intelligence models can be constructed for cases of different case types. This allows training on the historical information and the corresponding target historical information of cases of different case types, and subsequently, the corresponding artificial intelligence models can be used to extract the target current information for cases of different case types.

[0034] In addition, the AI ​​model needs to be pre-trained. Human experts can manually annotate multiple real transcripts from different case types to identify the target information for each case. This annotated data is then labeled in a way that machine learning algorithms can read. The annotated data is divided into training and test sets, which have no overlap. These two datasets are input into the AI ​​model for training. The machine learning algorithm performs supervised learning on the training set to form a trained model, and then extracts the target information from the test set. Compared to manual annotation, the accuracy of the model's extraction results exceeds a preset threshold. The trained model can then be used as a working model to extract target information.

[0035] In another embodiment of this application, the aforementioned artificial intelligence model is used to extract target current information from the current information, and a record is generated based on the target current information. This includes: determining the case type corresponding to the current information; determining the artificial intelligence model corresponding to the current information based on the case type; extracting the target current information from the current information using the artificial intelligence model corresponding to the current information; and generating the record based on the target current information. In this embodiment, for current information, the corresponding artificial intelligence model can be determined from multiple artificial intelligence models based on the case type of the current case. This allows different artificial intelligence models to be used to extract target current information for different case types, thereby generating the corresponding record more efficiently and accurately, further ensuring the high credibility of the generated record.

[0036] In another embodiment of this application, when there are multiple pieces of the aforementioned target current information, generating a record based on the target current information includes: acquiring multiple pieces of the aforementioned target current information; sorting the multiple pieces of the aforementioned target current information in chronological order to obtain sorted multiple pieces of target current information; converting the sorted multiple pieces of target current information to obtain target text information; adding the aforementioned target text information to a target template to generate the aforementioned record. In this embodiment, for the target current information, the specific event actually has a timeline. Therefore, the target current information can be sorted in chronological order, which makes the timeline of the target current information clearer and the generated record more accurate.

[0037] For example, if the respondent's current information includes information A, information B, and information C, where information A is "I went to the supermarket at 10 o'clock," information B is "I got up at 8 o'clock today," and information C is "I paid at 11 o'clock," the order of the current information should be information B, information A, and information C. This would help the questioner to organize the case.

[0038] In order to detect the format and completeness of the transcript and further improve its credibility, so as to provide a quantitative basis for consistency judgment for the questioner, in a specific embodiment of this application, after converting the sorted multiple target current information to obtain target text information, adding the target text information to the target template to generate the transcript, the method further includes: determining whether the transcript conforms to a predetermined format, and generating a first prompt message if the transcript does not conform to the predetermined format; determining whether the four elements of the transcript are complete, and generating a second prompt message if the four elements of the transcript are incomplete.

[0039] To accurately determine whether there are any discrepancies or contradictions among the statements of multiple target objects in the same case, further improve the credibility of the statements, and provide the questioner with a quantitative basis for consistency, in another specific embodiment of this application, there are multiple statements, each statement is generated from the current information of a target object during the current inquiry process. After converting the sorted current information of multiple targets to obtain target text information, and adding the target text information to the target template to generate the statement, the method further includes: obtaining the similarity between the multiple statements; generating a third prompt message when the similarity is greater than or equal to a similarity threshold; and generating a fourth prompt message when the similarity is less than the similarity threshold.

[0040] Specifically, during the actual interrogation process, a first transcript generated from the voice information spoken by the respondent can be obtained, as can a second transcript generated from the voice information spoken by the questioner, and a third transcript generated from the voice information spoken by the eyewitness. Of course, if there are multiple respondents, there will also be multiple third transcripts. The similarity between the first, second, and third transcripts is obtained. If the similarity is greater than or equal to the similarity threshold, the credibility of the multiple transcripts is considered relatively high. If the similarity is less than the similarity threshold, the consensus among the multiple transcripts is considered low, and there are contradictions. Therefore, the credibility of the multiple transcripts is considered relatively low. In this case, one of the respondents may provide information that does not conform to the truth. It is necessary to generate prompt information to provide to the questioner, so as to provide the interrogator with a quantitative basis for consistency judgment.

[0041] In another specific embodiment of this application, after obtaining the current information during the current query process, the method further includes: determining the language of the current information; if the language is not the target language, performing speech recognition on the current speech information in the current information, performing machine translation on the current text information in the current information, and converting the non-target language current information into the target language current information. In this embodiment, in order to generate transcripts more efficiently, speech recognition can be performed on the current speech information, machine translation can be performed on the current text information, and if the current information is not the target language, it can be converted into the target language current information.

[0042] For example, during an interrogation, the interrogator may use Chinese, while the person being interrogated may speak English, Chinese, French, or other languages. In order to generate transcripts more efficiently and provide a quantitative basis for judging the consistency of the transcript content, the current speech information in the non-target language can be converted into the current text information in the target language.

[0043] In another embodiment of this application, after generating the transcript based on the aforementioned target current information, the method further includes: sending the transcript to a display device so that the display device displays the transcript. In this embodiment, by sending the transcript to the display device, the questioner can more clearly understand the target current information of the case based on the transcript displayed on the display device.

[0044] This application also provides an apparatus for assisting in the generation of transcripts. It should be noted that the apparatus for assisting in the generation of transcripts in this application can be used to execute the method for assisting in the generation of transcripts provided in this application. The apparatus for assisting in the generation of transcripts provided in this application will be described below.

[0045] Figure 2This is a schematic diagram of an apparatus for assisting in the generation of transcripts according to an embodiment of this application. Figure 2 As shown, the device includes:

[0046] The first acquisition unit 10 is used to acquire historical information during the historical inquiry process. The historical information includes historical voice information and / or historical text information. The historical voice information and / or the historical text information includes at least target historical information. The target historical information includes historical object elements, historical objective elements, historical subject elements, and historical subjective elements.

[0047] The construction unit 20 is used to construct an artificial intelligence model, wherein the artificial intelligence model is trained using multiple sets of training data, and each set of training data includes the aforementioned historical information and the aforementioned target historical information corresponding to the aforementioned historical information.

[0048] Specifically, artificial intelligence models can be linear regression models, logistic regression models, decision tree models, leaf Bayes models, deep neural network models, and so on.

[0049] The second acquisition unit 30 is used to acquire current information during the current query process, including current voice information and / or current text information.

[0050] The first processing unit 40 is used to extract target current information from the above-mentioned current information using the above-mentioned artificial intelligence model, and generate a record based on the above-mentioned target current information. The target current information includes current object elements, current objective elements, current subject elements and current subjective elements.

[0051] In the aforementioned device, the first acquisition unit acquires historical information from the historical inquiry process, the construction unit constructs an artificial intelligence model, the second acquisition unit acquires current information from the current inquiry process, and the first processing unit uses the artificial intelligence model to extract target current information from the current information and generates a transcript based on the target current information. In this solution, by constructing an artificial intelligence model to train on historical information and corresponding target historical information, the target current information can be automatically extracted during the current inquiry process using the artificial intelligence model, forming a structured expression of the target current information and generating a transcript with high reliability.

[0052] In one embodiment of this application, the construction unit includes a construction module. The construction module is used to extract the target historical information from the historical information, and to train multiple artificial intelligence models using the historical information and the target historical information. Each of the artificial intelligence models is trained using multiple sets of training data. Each set of training data includes the historical information of a case of a certain case type and the target historical information corresponding to that historical information. In this embodiment, corresponding artificial intelligence models can be constructed for cases of different case types. This allows for training on the historical information and the corresponding target historical information of cases of different case types. Subsequently, the corresponding artificial intelligence models can be used to extract the target current information for cases of different case types.

[0053] In addition, the AI ​​model needs to be pre-trained. Human experts can manually annotate multiple real transcripts from different case types to identify the target information for each case. This annotated data is then labeled in a way that machine learning algorithms can read. The annotated data is divided into training and test sets, which have no overlap. These two datasets are input into the AI ​​model for training. The machine learning algorithm performs supervised learning on the training set to form a trained model, and then extracts the target information from the test set. Compared to manual annotation, the accuracy of the model's extraction results exceeds a preset threshold. The trained model can then be used as a working model to extract target information.

[0054] In another embodiment of this application, the first processing unit includes a first determining module, a second determining module, and a first processing module. The first determining module is used to determine the case type corresponding to the current information; the second determining module is used to determine the artificial intelligence model corresponding to the current information based on the case type; and the first processing module is used to extract the target current information from the current information using the artificial intelligence model corresponding to the current information, and generate the transcript based on the target current information. In this embodiment, for current information, the corresponding artificial intelligence model for the current case type can be determined from multiple artificial intelligence models based on the case type of the current case. This allows different artificial intelligence models to be used to extract the target current information for different case types, thereby generating the corresponding transcript more efficiently and accurately, further ensuring the high credibility of the generated transcript.

[0055] In another embodiment of this application, when there are multiple pieces of the aforementioned target current information, the first processing unit includes an acquisition module and a second processing module. The acquisition module is used to acquire multiple pieces of the aforementioned target current information, sort the multiple pieces of target current information in chronological order, and obtain sorted multiple pieces of target current information. The second processing module is used to convert the sorted multiple pieces of target current information to obtain target text information, add the target text information to the target template, and generate the aforementioned record. In this embodiment, for the target current information, the specific event actually has a timeline. Therefore, the target current information can be sorted in chronological order, which makes the timeline of the target current information clearer and the generated record more accurate.

[0056] For example, if the respondent's current information includes information A, information B, and information C, where information A is "I went to the supermarket at 10 o'clock," information B is "I got up at 8 o'clock today," and information C is "I paid at 11 o'clock," the order of the current information should be information B, information A, and information C. This would help the questioner to organize the case.

[0057] To check the format and completeness of the transcript and further improve its credibility, providing a quantitative basis for consistency judgment for questioners, in a specific embodiment of this application, the device further includes a second processing unit and a third processing unit. The second processing unit is used to convert multiple sorted target current information to obtain target text information, add the target text information to the target template, and generate the transcript. After that, it determines whether the transcript conforms to a predetermined format. If the transcript does not conform to the predetermined format, it generates a first prompt message. The third processing unit is used to determine whether the four elements of the transcript are complete. If the four elements of the transcript are incomplete, it generates a second prompt message.

[0058] To accurately determine whether there are any discrepancies or contradictions among the statements of multiple target objects in the same case, further improve the credibility of the statements, and provide the questioner with a quantitative basis for consistency, in another specific embodiment of this application, there are multiple statements, each statement is generated from the current information of a target object during the current questioning process. The device further includes a third acquisition unit, a first generation unit, and a second generation unit. The third acquisition unit is used to convert the sorted current information of multiple targets to obtain target text information, add the target text information to the target template, generate the statement, and then acquire the similarity between the multiple statements. The first generation unit is used to generate a third prompt message when the similarity is greater than or equal to a similarity threshold. The second generation unit is used to generate a fourth prompt message when the similarity is less than the similarity threshold.

[0059] Specifically, during the actual interrogation process, a first transcript generated from the voice information spoken by the respondent can be obtained, as can a second transcript generated from the voice information spoken by the questioner, and a third transcript generated from the voice information spoken by the eyewitness. Of course, if there are multiple respondents, there will also be multiple third transcripts. The similarity between the first, second, and third transcripts is obtained. If the similarity is greater than or equal to the similarity threshold, the credibility of the multiple transcripts is considered relatively high. If the similarity is less than the similarity threshold, the consensus among the multiple transcripts is considered low, and there are contradictions. Therefore, the credibility of the multiple transcripts is considered relatively low. In this case, one of the respondents may provide information that does not conform to the truth. It is necessary to generate prompt information to provide to the questioner, so as to provide the interrogator with a quantitative basis for consistency judgment.

[0060] In another specific embodiment of this application, the above-mentioned apparatus further includes a determining unit and a fourth processing unit. The determining unit is used to determine the language of the current information after acquiring the current information in the current query process. The fourth processing unit is used to perform speech recognition on the current speech information in the current information and machine translation on the current text information in the current information if the language is not the target language, thereby converting the non-target language current information into the target language current information. In this embodiment, in order to generate transcripts more efficiently, speech recognition can be performed on the current speech information, machine translation can be performed on the current text information, and if the current information is not the target language, it can be converted into the target language current information.

[0061] For example, during an interrogation, the interrogator may use Chinese, while the person being interrogated may speak English, Chinese, French, or other languages. In order to generate transcripts more efficiently and provide a quantitative basis for judging the consistency of the transcript content, the current speech information in the non-target language can be converted into the current text information in the target language.

[0062] In another embodiment of this application, the apparatus further includes a sending unit, which is configured to send the transcript to a display device after generating the transcript based on the target current information, so that the display device displays the transcript. In this embodiment, by sending the transcript to the display device, the questioner can more clearly understand the target current information of the case based on the transcript displayed on the display device.

[0063] The aforementioned device for generating transcripts includes a processor and a memory. The first acquisition unit, the construction unit, the second acquisition unit, and the first processing unit are all stored in the memory as program units, and the processor executes the aforementioned program units stored in the memory to achieve the corresponding functions.

[0064] The processor contains a kernel, which retrieves the corresponding program units from memory. One or more kernels can be configured; by adjusting kernel parameters, target information can be automatically extracted, improving the reliability of the transcript.

[0065] The memory may include non-permanent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0066] This invention provides a computer-readable storage medium storing a program thereon, which, when executed by a processor, implements the above-described method for generating transcripts.

[0067] This invention provides a processor for running a program, wherein the program executes the method for generating transcripts.

[0068] This application also provides a transcript generation system, including a microphone, a display device, and an auxiliary transcript generation device, wherein the auxiliary transcript generation device communicates with the microphone and the display device respectively, and the auxiliary transcript generation device is used to perform any of the above-described methods.

[0069] In the aforementioned system, since any of the methods described above is included, the method first acquires historical information from the historical query process, then constructs an artificial intelligence model, then acquires current information from the current query process, and finally uses the artificial intelligence model to extract target current information from the current information and generates a transcript based on the target current information. In this solution, by constructing an artificial intelligence model and training it on historical information and corresponding target historical information, the model can automatically extract target current information during the current query process, form a structured expression of the target current information, and generate a transcript with high credibility.

[0070] This invention provides a device including a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, it performs at least the following steps:

[0071] Step S101: Obtain historical information during the historical query process. The historical information includes historical voice information and / or historical text information. The historical voice information and / or the historical text information includes at least target historical information. The target historical information includes historical object elements, historical objective elements, historical subject elements, and historical subjective elements.

[0072] Step S102: Construct an artificial intelligence model, wherein the artificial intelligence model is trained using multiple sets of training data, and each set of training data includes the aforementioned historical information and the aforementioned target historical information corresponding to the aforementioned historical information.

[0073] Step S103: Obtain current information during the current query process, including current voice information and / or current text information;

[0074] Step S104: Use the above-mentioned artificial intelligence model to extract target current information from the above-mentioned current information, and generate a record based on the above-mentioned target current information. The above-mentioned target current information includes current object elements, current objective elements, current subject elements, and current subjective elements.

[0075] The devices mentioned in this article can be servers, PCs, tablets, mobile phones, etc.

[0076] This application also provides a computer program product, which, when executed on a data processing device, is suitable for executing an initialization program having at least the following method steps:

[0077] Step S101: Obtain historical information during the historical query process. The historical information includes historical voice information and / or historical text information. The historical voice information and / or the historical text information includes at least target historical information. The target historical information includes historical object elements, historical objective elements, historical subject elements, and historical subjective elements.

[0078] Step S102: Construct an artificial intelligence model, wherein the artificial intelligence model is trained using multiple sets of training data, and each set of training data includes the aforementioned historical information and the aforementioned target historical information corresponding to the aforementioned historical information.

[0079] Step S103: Obtain current information during the current query process, including current voice information and / or current text information;

[0080] Step S104: Use the above-mentioned artificial intelligence model to extract target current information from the above-mentioned current information, and generate a record based on the above-mentioned target current information. The above-mentioned target current information includes current object elements, current objective elements, current subject elements, and current subjective elements.

[0081] In the above embodiments of the present invention, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0082] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units described above can be a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between units or modules may be electrical or other forms.

[0083] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0084] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0085] If the aforementioned integrated units are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0086] As can be seen from the above description, the embodiments of this application achieve the following technical effects:

[0087] 1) The method for assisting in the generation of transcripts in this application first obtains historical information from the historical inquiry process, then constructs an artificial intelligence model, then obtains current information from the current inquiry process, and finally uses the artificial intelligence model to extract target current information from the current information and generates transcripts based on the target current information. In this scheme, by constructing an artificial intelligence model to train on historical information and the target historical information corresponding to the historical information, the artificial intelligence model can automatically extract target current information during the current inquiry process and form a structured expression of the target current information to generate transcripts with high credibility.

[0088] 2) The apparatus for assisting in the generation of transcripts according to this application comprises: a first acquisition unit acquiring historical information from the historical inquiry process; a construction unit constructing an artificial intelligence model; a second acquisition unit acquiring current information from the current inquiry process; and a first processing unit using the artificial intelligence model to extract target current information from the current information and generating a transcript based on the target current information. In this scheme, by constructing an artificial intelligence model to train on historical information and corresponding target historical information, the target current information can be automatically extracted during the current inquiry process using the artificial intelligence model, forming a structured expression of the target current information and generating a transcript with high credibility.

[0089] 3) The transcript generation system of this application, including any of the above-mentioned methods, first acquires historical information from the historical inquiry process, then constructs an artificial intelligence model, then acquires current information from the current inquiry process, and finally uses the artificial intelligence model to extract target current information from the current information and generates a transcript based on the target current information. In this solution, by constructing an artificial intelligence model to train on historical information and the corresponding target historical information, the system can automatically extract target current information during the current inquiry process, form a structured expression of the target current information, and generate a transcript with high credibility.

[0090] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A method for assisting in the generation of electronic transcripts, characterized in that, include: Acquire historical information during the historical query process, the historical information including historical voice information and / or historical text information, the historical voice information and / or the historical text information including at least target historical information, the target historical information including historical object elements, historical objective elements, historical subject elements and historical subjective elements; An artificial intelligence model is constructed, wherein the artificial intelligence model is trained using multiple sets of training data, and each set of training data includes the historical information and the target historical information corresponding to the historical information; Obtain current information during the current query process, including current voice information and / or current text information; The artificial intelligence model is used to extract target current information from the current information, and an electronic record is generated based on the target current information. The target current information includes current object elements, current objective elements, current subject elements, and current subjective elements. When there are multiple pieces of current target information, an electronic record is generated based on the current target information, including: Obtain multiple current target information, sort the multiple current target information according to time sequence, and obtain sorted multiple current target information; The sorted current information of multiple targets is converted to obtain target text information, which is then added to the target template to generate the electronic record. The electronic record may contain multiple entries, each generated from the current information of a target object during the current query process. After converting the sorted current information of multiple targets to obtain target text information, and adding the target text information to the target template to generate the electronic record, the method further includes: Obtain the similarity between multiple electronic records; If the similarity is greater than or equal to the similarity threshold, a third prompt message is generated; If the similarity is less than the similarity threshold, a fourth prompt message is generated.

2. The method according to claim 1, characterized in that, Building artificial intelligence models, including: Extract the target historical information from the historical information, and use the historical information and the target historical information to train multiple artificial intelligence models. Each artificial intelligence model is trained using multiple sets of training data. Each set of training data includes the historical information of a case of a certain case type and the target historical information corresponding to the historical information.

3. The method according to claim 2, characterized in that, The artificial intelligence model is used to extract target current information from the current information, and an electronic record is generated based on the target current information, including: Determine the case type corresponding to the current information; The artificial intelligence model corresponding to the current information is determined based on the case type; The target current information is extracted from the current information using the artificial intelligence model corresponding to the current information, and the electronic record is generated based on the target current information.

4. The method according to claim 1, characterized in that, After converting the sorted current information of multiple targets to obtain target text information, adding the target text information to the target template, and generating the electronic record, the method further includes: Determine whether the electronic record conforms to a predetermined format, and if the electronic record does not conform to the predetermined format, generate a first prompt message; Determine whether the four essential attributes of the electronic record are complete. If the four essential attributes of the electronic record are incomplete, generate a second prompt message.

5. The method according to any one of claims 1 to 4, characterized in that, After obtaining the current information during the current query process, the method further includes: Determine the language of the current information; If the language is not the target language, speech recognition is performed on the current speech information in the current information, and machine translation is performed on the current text information in the current information to convert the current information in the non-target language into the current information in the target language.

6. The method according to any one of claims 1 to 4, characterized in that, After generating an electronic record based on the target's current information, the method further includes: The electronic record is sent to a display device so that the display device can display the electronic record.

7. A device for assisting in the generation of electronic transcripts, characterized in that, include: The first acquisition unit is used to acquire historical information during the historical query process. The historical information includes historical voice information and / or historical text information. The historical voice information and / or the historical text information includes at least target historical information. The target historical information includes historical object elements, historical objective elements, historical subject elements, and historical subjective elements. A construction unit is used to construct an artificial intelligence model, wherein the artificial intelligence model is trained using multiple sets of training data, and each set of training data includes the historical information and the target historical information corresponding to the historical information; The second acquisition unit is used to acquire current information during the current query process, wherein the current information includes current voice information and / or current text information; The first processing unit is used to extract target current information from the current information using the artificial intelligence model, and generate an electronic record based on the target current information, wherein the target current information includes current object elements, current objective elements, current subject elements and current subjective elements; The device that assists in generating electronic transcripts is also used to perform the following steps: Obtain multiple current target information, sort the multiple current target information according to time sequence, and obtain sorted multiple current target information; The sorted current information of multiple targets is converted to obtain target text information, which is then added to the target template to generate the electronic record. There are multiple electronic records. Each electronic record is generated from the current information of a target object during the current query process. After converting the sorted current information of multiple targets to obtain target text information, the target text information is added to the target template to generate the electronic record. Then, the similarity between the multiple electronic records is obtained. If the similarity is greater than or equal to the similarity threshold, a third prompt message is generated. If the similarity is less than the similarity threshold, a fourth prompt message is generated.

8. An electronic record generation system, characterized in that, include: A microphone, a display device, and an apparatus for assisting in the generation of electronic transcripts, the apparatus for assisting in the generation of electronic transcripts communicating with the microphone and the display device respectively, the apparatus for assisting in the generation of electronic transcripts being used to perform the method according to any one of claims 1 to 6.

Images