Method for generating response, and device, medium and program product

Abstract

Provided in the embodiments of the present disclosure are a method for generating a response, and a device, a storage medium and a computer program product. The method comprises: acquiring a target question related to health; invoking a health model to generate a user intent of the target question and a first response to the target question; in response to the user intent meeting a predetermined condition, invoking a target model to generate a second response to the target question; and on the basis of the first response and the second response, determining a target response to the target question on the basis of the user intent. In the method of the embodiments of the present disclosure, a health model and a target model are invoked to generate responses respectively, and an appropriate response can be selected from among the generated responses on the basis of a user intent to respond to a user, thereby improving the accuracy of responses.

Description

Methods, equipment, media, and procedures for generating responses. Technical Field

[0001] This disclosure generally relates to the field of deep learning, and more specifically to methods for generating responses, electronic devices, computer-readable storage media, and computer program products. Background Technology

[0002] With the rapid development and continuous breakthroughs in deep learning technology, many deep learning models have emerged. Deep learning models typically contain a large number of parameters and are able to learn and understand the deeper meaning of language, generating coherent and logical textual content that impacts people's lives.

[0003] With technological advancements and a deeper understanding of specific industry needs, people have begun to explore applying deep learning models to more specialized scenarios, such as the health sector. Health models specifically designed for the health field have undergone professional training in health-related knowledge. They not only help users understand health-related information but also support disease prediction, personalized treatment plan design, and other tasks, greatly promoting the intelligentization of health services.

[0004] Summary of the Invention

[0005] According to exemplary embodiments of this disclosure, a method for generating responses, an electronic device, a computer storage medium, and a computer program product are provided.

[0006] In a first aspect of this disclosure, a method for generating a response is provided, comprising: obtaining a target question related to health; invoking a health model to generate a user intent for the target question and a first response to the target question; in response to the user intent satisfying predetermined conditions, invoking a target model to generate a second response to the target question; and determining a target response to the target question based on the first response and the second response, according to the user intent.

[0007] In a second aspect of this disclosure, an electronic device is provided, comprising: at least one processing unit; and at least one memory coupled to the at least one processing unit and storing instructions for execution by the at least one processing unit, the instructions causing the electronic device to perform the method described in the first aspect of this disclosure when executed by the at least one processing unit.

[0008] In a third aspect of this disclosure, a computer-readable storage medium is provided having machine-executable instructions stored thereon, which, when executed by a device, cause the device to perform the method described in the first aspect of this disclosure.

[0009] In a fourth aspect of this disclosure, a computer program product is provided, including computer-executable instructions, wherein the computer-executable instructions, when executed by a processor, implement the method described in the first aspect of this disclosure.

[0010] The summary section is provided to introduce a series of concepts in a simplified form, which will be further described in the detailed description below. The summary section is not intended to identify key or essential features of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0011] Figure 1 shows a schematic diagram of an example environment in which embodiments of the present disclosure can be implemented;

[0012] Figure 2 shows a flowchart of a method for generating a response according to an embodiment of the present disclosure;

[0013] Figure 3 shows a schematic diagram of a method for generating a response according to an embodiment of the present disclosure;

[0014] Figure 4 shows a framework diagram of a response generation method according to an embodiment of the present disclosure;

[0015] Figure 5A shows a schematic diagram of a method for generating a response according to an embodiment of the present disclosure;

[0016] Figure 5B shows a schematic diagram of a method for generating a response according to an embodiment of the present disclosure;

[0017] Figure 6 illustrates a schematic diagram of stored data according to an embodiment of the present disclosure;

[0018] Figure 7 illustrates a schematic diagram of text recognition according to an embodiment of the present disclosure;

[0019] Figure 8 shows a schematic block diagram of an example apparatus according to some embodiments of the present disclosure;

[0020] Figure 9 shows a block diagram of an example device that can be used to implement embodiments of the present disclosure.

[0021] In all the accompanying figures, the same or similar reference numerals denote the same or similar elements. Detailed Implementation

[0022] The names of messages or information exchanged between multiple devices in the embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of these messages or information. It is understood that before using the technical solutions disclosed in the embodiments of this disclosure, users should be informed of the types, scope of use, and usage scenarios of the personal information involved in this disclosure in an appropriate manner in accordance with relevant laws and regulations, and user authorization should be obtained.

[0023] For example, upon receiving a user's active request, a prompt message is sent to the user to explicitly inform them that the requested operation will require the acquisition and use of the user's personal information. This allows the user to independently choose whether to provide personal information to the software or hardware, such as the electronic device, application, server, or storage medium performing the operations of this disclosed technical solution, based on the prompt message. As an optional but non-limiting implementation, the prompt message can be sent to the user in the form of a pop-up window, where the prompt message can be presented in text format. Furthermore, the pop-up window can also include a selection control for the user to choose "agree" or "disagree" to provide personal information to the electronic device.

[0024] It is understood that the above notification and user authorization process are merely illustrative and do not constitute a limitation on the implementation of this disclosure. Other methods that comply with relevant laws and regulations may also be applied to the implementation of this disclosure.

[0025] Embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of this disclosure are shown in the drawings, it should be understood that this disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this disclosure. It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure.

[0026] In the description of embodiments of this disclosure, the term "comprising" and similar terms should be understood as open-ended inclusion, i.e., "including but not limited to". The term "based on" should be understood as "at least partially based on". The term "one embodiment" or "the embodiment" should be understood as "at least one embodiment". The terms "first", "second", etc., may refer to different or the same objects unless explicitly stated. Other explicit and implicit definitions may also be included below.

[0027] Currently, in health-related question-answering applications, only health models trained using health-specific data are typically deployed to complete answering or interpreting tasks. However, in real-world applications, the questions users ask of applications that invoke health models are highly diverse and complex, and relying solely on health models cannot accurately answer user questions or resolve the information provided by users.

[0028] To address this issue, this disclosure proposes a method for generating responses. This method not only calls a health model to generate a first response but also generates a user intent. Furthermore, based on whether the user intent meets predetermined conditions, it calls a target model to generate a second response. This allows a target response to be determined based on the user intent and these two responses. According to the method of embodiments of this disclosure, by generating responses by separately calling the health model and the target model, an appropriate response can be selected to answer the user based on the user intent, thereby improving the accuracy of the responses.

[0029] The embodiments of the present disclosure will now be described in further detail with reference to the accompanying drawings, wherein FIG1 illustrates a schematic diagram of an example environment 100 in which the embodiments of the present disclosure can be implemented. The example environment 100 includes computing devices 110 and 120. Computing device 110 may be deployed with target models (e.g., general models, such as general large language models) and health models (e.g., large language models specifically for the health domain), both of which are trained models capable of generating corresponding content based on user requests. Computing device 120 is also shown in FIG1. ​​In some embodiments, computing device 120 communicates with computing device 110 via network 130. Network 130 may include a wired network, a wireless network, or a combination thereof, for providing communication between computing device 120 and computing device 110. In some embodiments, computing device 120 may be connected to computing device 110 via a data cable; the present disclosure does not limit the connection method between computing device 110 and computing device 120.

[0030] Computing devices 110 and 120 may include, but are not limited to, personal computers, server computers, handheld or laptop devices, mobile devices (such as mobile phones, personal digital assistants (PDAs), media players, etc.), multiprocessor systems, consumer electronics, wearable electronic devices, smart home devices, minicomputers, mainframe computers, edge computing devices, and distributed computing systems that include any one of the above systems or devices.

[0031] The computing device 120 may have an application (e.g., a client program) installed for invoking the target model and the health model. Taking system 100 in FIG1 as an example, the computing device 120 can communicate with the computing device 110 via network 130, sending the target question 122, "I caught a cold yesterday and I feel a little dizzy today. What problems do I have?" The computing device 120 invokes the health model on the computing device 110 and inputs the target question 122 into the health model. In some embodiments, the health model performs intent recognition on the target question 122 and generates the user intent of the target question 122 and a first response to the target question 122. In some embodiments, the user intent can be represented by encoding. If the user intent meets predetermined conditions, the computing device 120 invokes the target model on the computing device 110 and inputs the target question 122 into the target model, thereby generating a second response to the target question 122. Whether the user intent meets predetermined conditions can be determined according to the range and size of the encoding. In some embodiments, the computing device 120 determines the target response 124 for the target question based on the first response and the second response, according to the user intent. In this embodiment, based on the analysis of user intent, it is more appropriate for the target question 122 to be processed by a health model. Therefore, the determined target response 124 is derived from the first response, "You may have a mild cold; it is recommended to drink more hot water."

[0032] According to the method of embodiments of this disclosure, by generating responses by respectively invoking a health model and a target model, an appropriate response can be selected to answer the user based on the user's intent, thereby improving the accuracy of the response. As shown in FIG1, in environment 100, network 130 can be used to transmit data between computing device 110 and computing device 120. Network 130 has a theoretical bandwidth, which refers to the maximum transmission speed supported by network 130, representing the maximum amount of data that network 130 can transmit under ideal conditions, usually measured in bits per second (bps). For example, if the theoretical bandwidth of network 130 is 100Mbps, it means that under ideal conditions it can transmit 100 megabits of data per second. However, in reality, due to other factors that may exist in the network (e.g., signal interference, bandwidth sharing, transmission delay, etc.), the actual transmission speed of 100Mbps may not be achieved.

[0033] As understood by those skilled in the art, instances of computing device 110 can be independent physical servers, server clusters or distributed systems composed of multiple physical servers, or cloud servers providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. Servers can be connected directly or indirectly via wired or wireless communication, and this application does not impose any limitations on this.

[0034] The computing device 120 can be any type of mobile computing device, including mobile computers (e.g., personal digital assistants (PDAs), laptops, notebook computers, tablet computers, netbooks, etc.), mobile phones (e.g., cellular phones, smartphones, etc.), wearable computing devices (e.g., smartwatches, head-mounted devices, including smart glasses, etc.) or other types of mobile devices. In some embodiments, the computing device 120 can also be a fixed computing device, such as a desktop computer, game console, smart TV, etc. It should be understood that, if the computing device 120 has sufficient computing power, the computing device 120 can replace the computing device 110 to perform the above operations, or the computing device 110 and the computing device 120 can jointly perform the above operations.

[0035] It should be understood that the architecture and functionality in example environment 100 are described for illustrative purposes only and do not imply any limitation on the scope of this disclosure. Embodiments of this disclosure can also be applied to other environments with different structures and / or functionalities.

[0036] The process according to embodiments of the present disclosure will be described in detail below with reference to Figures 2 to 7. For ease of understanding, the specific data mentioned in the following description are exemplary and are not intended to limit the scope of protection of this disclosure. It is understood that the embodiments described below may also include additional actions not shown and / or actions shown may be omitted, and the scope of this disclosure is not limited in this respect.

[0037] Figure 2 illustrates a flowchart of a method 200 for generating a response according to certain embodiments of the present disclosure. At block 202, computing device 120 acquires a target question related to health. In some embodiments, computing device 120 may be equipped with an application (e.g., a client program) that invokes a health model to generate user intent and a first response, thereby enabling the execution of the method for generating a response according to embodiments of the present disclosure.

[0038] In some embodiments, the computing device 120 may launch an application and display a trigger control on the interface after the application launches (e.g., the application's startup screen or main screen). The user can initiate a trigger operation by clicking or touching the trigger control. For example, the trigger control can be a button control displayed on the current screen. Alternatively, the user can also initiate the trigger operation via methods such as voice. This disclosure does not limit the specific implementation of the trigger operation.

[0039] The computing device 120 can respond to a user-triggered operation and display a dialog interface for interaction. This dialog interface may include an input area and an information display area. The input area can receive user input and transmit it to the intelligent model 112 in the computing device 110. The model 112 then generates a corresponding result based on the user input and sends the result back to the computing device 120. The computing device 120 can display the received result in the information display area for user convenience.

[0040] In box 204, computing device 120 invokes a health model to generate a user intent for the target question and a first response to the target question. The user intent is the health model's judgment of the intent behind the target question. In some embodiments, the user intent can be represented by a series of codes, for example, a user intent encoded as 1000 indicates that the target question involves a question unrelated to health, a user intent encoded as 2000 indicates that the target question involves a question about pediatrics, and so on. The first response is the answer to the target question generated by the health model. After generating the first response, it does not need to be presented immediately; further judgment is needed based on the user intent to select a more appropriate response. The health model can be a deep learning model trained using health-related training samples. In some embodiments, the health model can be a large language model used to return to health-related questions.

[0041] In box 206, in response to the user intent meeting predetermined conditions, computing device 120 invokes the target model to generate a second response to the target question. In embodiments where the user intent is encoded, the predetermined conditions can also be conditions established based on these encodings, such as numerical ranges, numerical magnitudes, etc. Meeting the predetermined conditions means that the target model's response may be more appropriate for the target request. Therefore, it is necessary to invoke the target model to generate the second response. This target model can be, for example, a general Large Language Model (LLM). A Large Language Model is a deep learning model trained using a large amount of text data that can generate natural language text or understand the meaning of language text.

[0042] In box 208, computing device 120 determines a target response to the target question based on the first response and the second response, according to the user's intent. This target response is the final response presented to the user. As described above, target questions have diverse characteristics, and user intent reflects the deeper meaning of the target question 122. Therefore, determining whether to use the first response generated by the health model or the second response generated by the target model as the target response based on the user intent helps improve the accuracy of the response. According to the method of embodiments of this disclosure, by calling the health model and the target model respectively to generate responses, it is possible to select an appropriate response to answer the user based on the user's intent, thereby improving the accuracy of the response.

[0043] Figure 3 illustrates a schematic diagram of the response method according to an embodiment of the present disclosure. As shown in Figure 3, at 302, the user's question (i.e., the target question) is obtained. In some embodiments, the question may be text data input by the user. In other embodiments, the question may be image information uploaded by the user through a preset control. At 304, the question may be directly passed to the health model. The health model performs intent analysis based on the question, obtains the user intent at 306, and obtains a first response at 312. If the user intent meets predetermined conditions, such as the user intent indicating that the question is unrelated to health (which may be due to a bias in the user's understanding of the question), the process can proceed to 308 to invoke the target model. At 310, the target model generates a second response. At 314, based on the indication of the user intent, such as the user intent indicating that the question is a regular, typical health-related question, the first response can be determined as the target response. However, if the user intent indicates that the response from the target model is more appropriate for the question, the second response can be determined as the target response.

[0044] In some embodiments, a user can input a question on a health-related interface of the computing device 120. In some embodiments, this interface includes an image input component and a text input component. If the user triggers the image input component, image information is received as the target question. If the user triggers the text input component, text information is received as the target question. Because the interface itself is health-related, the image input component can be set to "data interpretation," so the user knows that the action is triggered in the health field. Therefore, the user does not need to input any text; the uploaded image can be directly transmitted to the health model for interpretation. For example, the user can upload an image of a test report, and the health model interprets the image as the first response. In this embodiment, the health model can be accessed through different interfaces, allowing various interfaces to reuse the same link, thus improving data processing efficiency.

[0045] Figure 4 shows a framework diagram of a method for generating a response according to an embodiment of the present disclosure. As shown in Figure 4, user 402 can trigger input operation 412 in an application client 410 installed on computing device 120. As described above, user 402 can input a question via an image input component or a text input component. In the case of inputting a question via an image input component, user 402 does not even need to input text content. In this embodiment, health agent 430 can be integrated into application client 410 as a software entity performing part of the operation of this method, for example, it can serve as a functional interface on application client 410 (e.g., a chat window for health-related questions). In some embodiments, health agent 430 can be a separate hardware entity. Requests are sent to health agent 430 via input operation 412.

[0046] If the input is image information, such as when user 402 selects the "Data Interpretation" component and uploads their own test report, the health agent 430 can perform pre-processing 432 on the input question. This pre-processing 432 detects the objects (i.e., target objects) in the image information. These objects include, but are not limited to, text in the image and the resolution of the text in the image. If the object does not meet predetermined conditions (e.g., there is no text in the image, or the image is blank), the health agent 430 can issue a prompt, such as "Image error." If these objects meet the predetermined conditions, the health agent 430 performs text recognition on the image information to obtain the text in the image information. In this embodiment, the OCR (Optical Character Recognition) 442 deployed in the health model 440 can be called to perform text recognition. Then, the health agent 430 determines the text as a question. Because the "Data Interpretation" component does not require the user to upload text, information such as "Please interpret the above content" can be added to the determined text so that the health model 440 can generate a more accurate answer. This embodiment provides multiple input interfaces, which is beneficial for users to input their questions simply, quickly, and accurately.

[0047] Upon receiving a question, the health agent 430 can act as a bridge between the application client 410 and the health model 440, forwarding the question to the health model 440 via content forwarding 434. The health model 440 performs intent recognition 444 on the question to obtain the user intent. Based on, for example, a specific encoding of the user intent, a final response (i.e., the target response) is determined from the responses generated by the health model 440 and the target model 450. In some embodiments, the response generated by the target model 440 can be used as the final response based on the user intent. In some embodiments, the health agent 430 can detect whether the user intent indicates that the target question is related to health. If the user intent indicates that the target question is unrelated to health, the health agent 430 sends a request to the target model 450 via content forwarding 434, thereby invoking the target model 450 to generate a second response to the target question. Since the question is unrelated to health, and the target model (such as a general model) has relatively stronger accuracy in answering non-health questions, the second response can be used as the final response and sent to the health agent 430.

[0048] In some embodiments, the response generated by the health model 440 can be used as the final response based on the user's intent. The health model 440 can respond to questions within a target health domain. The target health domain 452 refers to the domain that the health model 440 can respond to. For example, during pre-training and training, the health model 440 uses a sample set from the target domain for training, thereby acquiring health knowledge in the target domain and being able to respond to questions within that domain. Other domains are outside the target health domain 452 and cannot be addressed by the health model 440 in those domains. The health model 440 may possess diagnostic and treatment capabilities 446 and science popularization capabilities 448. Diagnostic and treatment capabilities 446 refer to the health model 440's ability to provide diagnostic and treatment suggestions for questions related to diagnosis and treatment, such as offering personalized treatment plans. Science popularization capabilities 448 refer to the health model 440's ability to share science popularization knowledge for questions related to science popularization, such as explaining the mechanism of action of a cold. If the question falls within the target health domain 452, regardless of whether the question involves diagnostic and treatment capabilities 446 or science popularization capabilities 448, the health model 440 can be used to provide a response. Of course, it must not conflict with other conditions of this disclosure.

[0049] The health agent 430 can process responses 436 received from the target model 450 or the health model 440, including presenting the responses on an interface and storing them in the application server 420. The application server 420 is a backend supporting the application client 410, and can, for example, provide storage services for interactive messages to the application client 410. The response can be saved in the message interaction / storage 422 of the application server 420. The response can be in text or image format, and therefore can be stored as text 424 or image 426. The health agent 430 can present the historical messages stored in the application server 420 on the interface of the application client 410 through the response processing 436.

[0050] More implementation details will be described below. Figure 5A shows a schematic diagram of a response method according to an embodiment of this disclosure. The content shown in Figure 5A is part of the method for generating a response, which is connected to the content shown in Figure 5B. As shown in Figure 5A, at 512, the user can click the image input component "Report Interpretation" on the interface of the application client 510 to upload image information (e.g., the user's detection report). At 522, the health agent 520 pops up a page accordingly. At 524, the user can choose to take a photo or upload a local image on this page. At 592, the image uploaded by the user is saved in the dedicated storage 590. Then, at 552, the health agent 520 accesses the OCR interface from the health model 550 and uses its text recognition function to perform text recognition on the image information at 526 to obtain the text. At 528, the health agent 520 can verify the extracted text to determine whether the text constitutes a qualified question. If the verification fails, at 513, a prompt is given to the application client 510, such as the prompt message "The image is incorrect, please re-upload". If the validation passes, at point 530, the current execution logic is redirected to the logic calling health model 550 via a hook function. At point 532, the cached text after text recognition technology is read, i.e., the obtained text. At point 554, this text is sent as a question to health model 550. If the user selects text output at point 514, then health agent 520 can forward the text as a question to health model 550 via a forwarding request at point 534.

[0051] Regardless of the interface through which the request is input into the health model 550, the health model 550 can identify the intent based on the query at 556 and determine the response and user intent. This user intent can be represented using a specific encoding, allowing it to be added to a blank field in the communication protocol for transmission. The health agent 520 receives the response (i.e., the first response) at 536. See Figure 5B for further details.

[0052] Figure 5B illustrates a schematic diagram of a method for generating a response according to an embodiment of the present disclosure. As described above, after the health agent 520 receives a response, it parses the received data at 538 to obtain the user intent. At 540, it determines whether the user intent meets predetermined conditions. As described with respect to Figure 4, if the user intent indicates that the target question is unrelated to health, the health agent 520 sends a request to the target model 570, thereby invoking the target model 570 to generate a second response to the target question at 572. Since the question is unrelated to health, and the target model has relatively higher accuracy in answering non-health questions, the second response can be used as the final response and sent to the health agent 520.

[0053] In some embodiments, the health model 550 is used to answer questions in the target health domain within the health category. If the user intent indicates that the question is related to health, the health agent 520 continues to detect at 544 whether the user intent relates to the target health domain. If the user intent does not relate to the target health domain, the health agent 520 detects at 516 whether the user intent relates to a knowledge answer. At 5164, if the user intent relates to a knowledge answer, the second response generated by the target model is used as the target response.

[0054] At point 5162, if the detection result indicates that the user's intent does not involve knowledge-based answers, the preset response will be determined as the target response to the target question. For example, the preset response could be, "I'm very sorry, the question you raised is beyond my current area of ​​expertise, and I may not be able to answer it. I am working hard to learn, and I will provide you with more comprehensive health consultation services in the future."

[0055] At point 546, if the question falls within the target health domain, it can be further analyzed to determine if the question pertains to the target population. The target population refers to specific groups, such as infants and pregnant women. For example, a question like "What should a pregnant woman do if she has abdominal discomfort?" would fall within the target population. At point 517, if the question pertains to the target population, a pre-set response is provided. For instance, a response might be, "Thank you very much for your question. Given that your situation falls under the category of a specific population, we recommend that you seek professional help at a reputable offline medical institution as soon as possible, or consult a specialist doctor online."

[0056] At point 548, if the target population is not involved, the system continues to check whether the question involves the target vital sign. This target vital sign indicates that the patient's illness is an urgent, life-threatening condition that requires prompt medical attention; otherwise, it may cause severe harm or death. For example, the target vital sign could be a critical illness. At point 518, if the detection does not involve the target feature, a response that identifies health model 550 as the target response is provided. At point 519, if the detection involves the target feature, a preset response is provided. For example, this preset response could be, "Thank you very much for your question. Given your description, there may be a risk of a critical illness. It is recommended that you seek professional help from a qualified medical institution as soon as possible for timely diagnosis and treatment. If you have any other health-related questions, I am happy to continue to assist you." In this embodiment, Figures 5A and 5B are used to provide response strategies for different question scenarios, offering comprehensive guidance on determining the target response based on the first and second responses, which helps improve the accuracy of the response.

[0057] Figure 6 illustrates a schematic diagram of stored data according to an embodiment of the present disclosure. As shown in Figure 6, client 602 and server 604 interact with each other. Client 602 may have client programs installed that invoke the target model and health model, while server 604 may include other devices besides the client, such as a server, a device deploying the health model, a device deploying the target model, etc. This separation architecture between client 602 and server 604 reduces the performance requirements on client 602 and improves the efficiency of generating responses.

[0058] Figure 7 illustrates a schematic diagram of text recognition according to an embodiment of the present disclosure. In this embodiment, text recognition begins at 702. As described above, after a pop-up page appears at 522, the user can choose to upload an image by taking a photo at 704, or select a local image at 706. At 708, the uploaded image undergoes preliminary detection. At 710, it is checked whether the image is blank, for example, whether it contains text. If it is blank, a prompt is given at 712, such as "Please re-upload". If it is not blank, text recognition is performed at 714. At 716, the recognized text is sent to the health model. The text recognition process is thus completed.

[0059] The pre-process shown in Figure 7 avoids sending blank or erroneous images to the health model, thus preventing the waste of computational resources and avoiding providing incorrect answers to users. Figure 8 shows a schematic block diagram of an example device 800 according to some embodiments of this disclosure. Device 800 can be implemented by software, hardware, or a combination of both. As shown in Figure 8, device 800 includes a question acquisition module 810, a first invocation module 820, a second invocation module 830, and an answer determination module 840.

[0060] In some embodiments, the question acquisition module 810 can acquire a target question related to health. The first invocation module 820 invokes a health model to generate a user intent for the target question and a first response to the target question. The second invocation module 830 can, in response to the user intent meeting predetermined conditions, invoke the target model to generate a second response to the target question. The response determination module 840 can, based on the first and second responses, determine a target response to the target question according to the user intent.

[0061] The device 800 in Figure 8 can be used to implement the process described above in conjunction with Figures 1 to 7, which will not be repeated here for the sake of brevity.

[0062] The division of modules or units in the embodiments of this disclosure is illustrative and only represents one logical functional division. In actual implementation, there may be other division methods. Furthermore, the functional units in the disclosed embodiments may be integrated into one unit, exist as separate physical entities, or two or more units may be integrated into one unit. The integrated unit described above can be implemented in hardware or as a software functional unit.

[0063] Figure 9 shows a block diagram of an example device 900 that can be used to implement embodiments of the present disclosure. It should be understood that the device 900 shown in Figure 9 is merely exemplary and should not be construed as limiting the functionality and scope of the implementations described herein. For example, device 900 can be used to correspond to computing device 120 described herein in conjunction with Figure 1 and can be used to perform the processes of Figures 1 through 7 described above.

[0064] As shown in Figure 9, device 900 is in the form of a general-purpose computing device. Components of computing device 900 may include, but are not limited to, one or more processors or processing units 910, memory 920, storage devices 930, one or more communication units 940, one or more input devices 950, and one or more output devices 960. Processing unit 910 may be a physical or virtual processor and is capable of performing various processes according to programs stored in memory 920. In a multiprocessor system, multiple processing units execute computer-executable instructions in parallel to improve the parallel processing capability of computing device 900.

[0065] Computing device 900 typically includes multiple computer storage media. Such media can be any available media accessible to computing device 900, including but not limited to volatile and non-volatile media, removable and non-removable media. Memory 920 can be volatile memory (e.g., registers, cache, random access memory (RAM)), non-volatile memory (e.g., read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory), or some combination thereof). Storage device 930 can be removable or non-removable media and may include machine-readable media, such as flash drives, disks, or any other media capable of storing information and / or data (e.g., training data for training) and accessible within computing device 900.

[0066] The computing device 900 may further include additional removable / non-removable, volatile / non-volatile storage media. Although not shown in FIG. 9, disk drives for reading from or writing to removable, non-volatile disks (e.g., "floppy disks") and optical disk drives for reading from or writing to removable, non-volatile optical disks may be provided. In these cases, each drive may be connected to a bus (not shown) via one or more data media interfaces. The memory 920 may include a computer program product 925 having one or more program modules configured to perform various methods or actions of various implementations of the present disclosure.

[0067] The communication unit 940 enables communication with other computing devices via a communication medium. Additionally, the components of the computing device 900 can function as a single computing cluster or multiple computing machines capable of communicating via communication connections. Therefore, the computing device 900 can operate in a networked environment using logical connections to one or more other servers, network personal computers (PCs), or another network node.

[0068] Input device 950 can be one or more input devices, such as a mouse, keyboard, trackball, etc. Output device 960 can be one or more output devices, such as a monitor, speaker, printer, etc. Computing device 900 can also communicate with one or more external devices (not shown) via communication unit 940 as needed. These external devices include storage devices, display devices, etc., and can communicate with one or more devices that enable user interaction with computing device 900, or with any device that enables computing device 900 to communicate with one or more other computing devices (e.g., network card, modem, etc.). Such communication can be performed via an input / output (I / O) interface (not shown).

[0069] According to an exemplary implementation of this disclosure, a computer-readable storage medium is provided that stores computer-executable instructions thereon, wherein the computer-executable instructions are executed by a processor to implement the methods described above. According to an exemplary implementation of this disclosure, a computer program product is also provided, which is tangibly stored on a non-transitory computer-readable medium and includes computer-executable instructions, which are executed by a processor to implement the methods described above. According to an exemplary implementation of this disclosure, a computer program product is provided that stores a computer program thereon, which, when executed by a processor, implements the methods described above.

[0070] Various aspects of this disclosure are described herein with reference to flowchart illustrations and / or block diagrams of methods, apparatuses, devices, and computer program products implemented according to this disclosure. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer-readable program instructions.

[0071] These computer-readable program instructions can be provided to a processing unit of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that, when executed by the processing unit of the computer or other programmable data processing apparatus, they create means for implementing the functions / actions specified in one or more blocks of the flowchart and / or block diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium that causes a computer, programmable data processing apparatus, and / or other device to operate in a particular manner. Thus, the computer-readable medium storing the instructions comprises an article of manufacture that includes instructions for implementing aspects of the functions / actions specified in one or more blocks of the flowchart and / or block diagram.

[0072] Computer-readable program instructions can be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions that execute on the computer, other programmable data processing apparatus, or other device to perform the functions / actions specified in one or more boxes of a flowchart and / or block diagram.

[0073] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of an instruction, which contains one or more executable instructions for implementing the specified logical function. In some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0074] Various implementations of this disclosure have been described above. The foregoing description is exemplary and not exhaustive, nor is it limited to the disclosed implementations. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described implementations. The terminology used herein is chosen to best explain the principles, practical applications, or improvements to technology in the market, or to enable others skilled in the art to understand the various implementations disclosed herein.

Claims

1. A method for generating a response, comprising: Ask health-related questions; The health model is invoked to generate the user intent for the target question and the first response to the target question; In response to the user intent meeting predetermined conditions, the target model is invoked to generate a second response to the target question; as well as Based on the first response and the second response, a target response is determined for the question asked about the target based on the user's intent.

2. The method of claim 1, wherein in response to the user intent satisfying a predetermined condition, invoking the target model to generate a second response to the target question comprises: Detect whether the user intent indicates that the target question is related to health; as well as In response to the user intent indicating that the target question is unrelated to health, the target model is invoked to generate a second response to the target question; Furthermore, based on the first response and the second response, the target response for the question regarding the target is determined according to the user intent, including: The second response shall be taken as the target response.

3. The method according to claim 2, wherein the health model is used to answer questions in a target health domain within a health category, and in response to the user intent satisfying predetermined conditions, invoking the target model to generate a second response to the target question further includes: In response to the user intent indicating that the target question is related to health, detect whether the user intent relates to the target health area; In response to the user intent not relating to the target health domain, detect whether the user intent relates to knowledge answering; as well as In response to the user intent relating to the knowledge answer, the target model is invoked to generate a second response to the target question.

4. The method according to claim 3, further comprising: In response to the user's intent not involving the knowledge answer, the first preset response is determined as the target response to the target question.

5. The method of claim 3, wherein determining the target response to the target question based on the first response and the second response, according to the user intent, comprises: In response to the user intent relating to the target health area, detect whether the user intent relates to the target population; as well as Since the user intent does not relate to the target population, the first response is determined to be the target response. complex.

6. The method according to claim 5, further comprising: In response to the user intent involving the target group, the second preset response is determined as the target response.

7. The method according to claim 5, further comprising: In response to the user intent relating to the target health domain, detect whether the user intent relates to the target vital signs; as well as In response to the user intent relating to the target vital signs, a third preset response is determined as the target response.

8. The method according to claim 1, wherein the target question is obtained from an interface, the interface being provided with an image input component and a text input component, and obtaining the target question includes: In response to the user triggering the image input component, image information is received as the target question; as well as In response to the user triggering the text input component, text information is received as the target question.

9. The method according to claim 8, further comprising: Detect the target object of the image information; In response to the target object satisfying a second predetermined condition, text recognition is performed on the image information to obtain the target text; as well as The target text is identified as the target question.

10. The method of claim 9, further comprising: In response to the target object not meeting the second predetermined condition, a prompt message is displayed.

11. An electronic device, comprising: At least one processing unit; At least one memory coupled to the at least one processing unit and storing instructions for execution by the at least one processing unit, the instructions causing the electronic device to perform the method according to any one of claims 1 to 10 when executed by the at least one processing unit.

12. A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method according to any one of claims 1 to 10.

13. A computer program product having a computer program stored thereon, which, when executed by a processor, implements the method according to any one of claims 1 to 10.

Images