Role interaction method, electronic equipment and storage medium

Abstract

The invention discloses a role interaction method, electronic equipment and a storage medium. The method comprises the steps that a historical record to be used in a preset interaction environment is acquired, and the historical record is used for recording records generated in the historical interaction process in the preset interaction environment; creating a plurality of proxy roles based on the historical record, the plurality of proxy roles being used for respectively simulating a plurality of real roles participating in interaction in a historical interaction process; and controlling the plurality of agent roles to perform target interaction in a preset interaction environment. According to the method and the device, the problem that the correlation degree and the accuracy of the generated reply information are relatively low when the information is generated in an artificial intelligence mode in the related technology is solved.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular, to a role interaction method, an electronic device, and a storage medium. Background Art

[0002] With the increasingly wide application of artificial intelligence technology, human life has also changed accordingly. Nowadays, artificial intelligence technology (such as machine learning models) can be used for network interaction (such as artificial intelligence medical consultation), which greatly facilitates people's lives. However, currently, network interaction based on artificial intelligence technology usually focuses on the task mode of "user question - machine learning model reply", which has certain limitations. For example, in the man - machine dialogue scenario for specific tasks, users use a simple "question - answer" inquiry method. From the content feedback by the artificial intelligence method, both the relevance and accuracy of the reply content need to be improved in various aspects, that is, the relevance and accuracy of the generated reply information are relatively low.

[0003] In response to the above problems, no effective solution has been proposed yet. Summary of the Invention

[0004] Embodiments of the present invention provide a role interaction method, an electronic device, and a storage medium, which at least solve the technical problem that the relevance and accuracy of the generated reply information are relatively low when generating information through artificial intelligence in the related art.

[0005] According to one aspect of the embodiments of the present invention, a role interaction method is provided, including: obtaining historical records to be used in a preset interaction environment, where the historical records are used to record the records generated during the historical interaction process in the preset interaction environment; creating a plurality of proxy roles based on the historical records, where the plurality of proxy roles are used to respectively simulate a plurality of real roles participating in the interaction during the historical interaction process; controlling the plurality of proxy roles to perform target interaction in the preset interaction environment.

[0006] According to another aspect of the embodiments of the present invention, a role interaction method is further provided, including: obtaining historical medical records to be used in an interaction diagnosis environment, where the historical medical records are used to record the medical records generated during the historical interaction diagnosis process in the interaction diagnosis environment; creating a plurality of proxy medical roles based on the historical medical records, where the plurality of proxy medical roles are used to respectively simulate a plurality of real medical roles participating in the interaction diagnosis during the historical interaction diagnosis process; controlling the plurality of proxy medical roles to perform auxiliary diagnosis in the interaction diagnosis environment.

[0007] According to another aspect of the embodiments of the present invention, there is also provided a role interaction method, including: obtaining a clinical diagnosis request through a first application programming interface, wherein the request data carried in the clinical diagnosis request includes: condition information; returning a clinical diagnosis response through a second application programming interface, wherein the response data carried in the clinical diagnosis response includes: an adjuvant treatment plan, the adjuvant treatment plan is determined by a proxy medical role, the proxy medical role is created based on historical medical records, the historical medical records are used to record the medical records generated during the historical diagnosis process in an interactive diagnosis environment, and the proxy medical role is used to simulate the real medical roles participating in the interactive diagnosis during the historical interactive diagnosis process.

[0008] According to another aspect of the embodiments of the present invention, there is also provided a role interaction method, including: obtaining a currently input clinical diagnosis dialogue request, wherein the information carried in the clinical diagnosis dialogue request includes: condition information; in response to the clinical diagnosis dialogue request, returning a clinical diagnosis dialogue reply, wherein the information carried in the clinical diagnosis dialogue reply includes: an adjuvant treatment plan, the adjuvant treatment plan is determined by a proxy medical role, the proxy medical role is created based on historical medical records, the historical medical records are used to record the medical records generated during the historical diagnosis process in an interactive diagnosis environment, and the proxy medical role is used to simulate the real medical roles participating in the interactive diagnosis during the historical interactive diagnosis process; displaying the treatment plan within a graphical user interface.

[0009] According to another aspect of the embodiments of the present invention, there is also provided an electronic device, including: a memory storing an executable program; a processor for running the program, wherein when the program runs, it is the role interaction method of any one of the above.

[0010] According to another aspect of the embodiments of the present invention, there is also provided a computer-readable storage medium, the computer-readable storage medium includes a stored executable program, wherein when the executable program runs, it controls the device where the computer-readable storage medium is located to execute the role interaction method of any one of the above.

[0011] According to another aspect of the embodiments of the present invention, there is also provided a computer program product, including a computer program, and when the computer program is executed by a processor, it implements the role interaction method of any one of the above.

[0012] In an embodiment of the present invention, by obtaining historical records to be used in a preset interaction environment, where the historical records are used to record the records generated during the historical interaction process in the preset interaction environment; creating a plurality of agent roles based on the historical records, where the plurality of agent roles are used to respectively simulate a plurality of real roles participating in the interaction during the historical interaction process; controlling the plurality of agent roles to perform a target interaction in the preset interaction environment. It is easy to notice that by obtaining the historical records to be used in the preset interaction environment and creating a plurality of agent roles according to the historical records, the agent roles can be controlled to simulate the target interaction in the preset interaction environment, that is, the entire target interaction process is simulated by a plurality of agent roles, avoiding the task model of the user asking questions and the artificial intelligence replying during the process of using artificial intelligence technology for interaction in the traditional situation. Thus, the interaction ability of the artificial intelligence itself can be greatly explored, and the relevance and accuracy of the generated reply information are improved. Furthermore, the technical problem that the relevance and accuracy of the generated reply information are relatively low when generating information by artificial intelligence in the related art is solved. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] The drawings described herein are used to provide a further understanding of the present invention, form a part of this application, and the schematic embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation to the present invention. In the drawings:

[0014] Figure 1 A hardware structure block diagram of a computer terminal (or mobile device) for implementing a role interaction method is shown;

[0015] Figure 2 It is a flowchart of the role interaction method according to Embodiment 1 of the present invention;

[0016] Figure 3 It is a flowchart of a medical consultation according to an embodiment of the present application;

[0017] Figure 4 It is a schematic diagram of a target prompt word of an AI intern doctor according to an embodiment of the present application;

[0018] Figure 5 It is a schematic diagram of a target prompt word of an AI medical director according to an embodiment of the present application;

[0019] Figure 6 It is a schematic diagram of an opinion interaction according to an embodiment of the present application;

[0020] Figure 7 It is a flowchart of the role interaction method according to Embodiment 2 of the present invention;

[0021] Figure 8 It is a flowchart of the role interaction method according to Embodiment 3 of the present invention;

[0022] Figure 9 is a flowchart of the role interaction method according to Embodiment 4 of the present invention;

[0023] Figure 10 is a schematic diagram of the role interaction device according to Embodiment 5 of the present application;

[0024] Figure 11 is a schematic diagram of the role interaction device according to Embodiment 6 of the present application;

[0025] Figure 12 is a schematic diagram of the role interaction device according to Embodiment 7 of the present application;

[0026] Figure 13 is a schematic diagram of the role interaction device according to Embodiment 8 of the present application;

[0027] Figure 14 is a structural block diagram of a computer terminal according to an embodiment of the present invention. Detailed implementation manners

[0028] In order to enable those skilled in the art to better understand the solution of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0029] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and do not necessarily need to describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances so that the embodiments of the present invention described herein can be implemented in an order different from those illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0030] First, some nouns or terms that appear in the process of describing the embodiments of the present application are applicable to the following explanations:

[0031] Intern doctors: Responsible for clinical diagnosis, actively interacting with patients, using medical knowledge and skills to infer diseases, and cooperating with inspectors to conduct medical examinations to confirm the diagnosis. They are also responsible for formulating treatment plans, considering the effectiveness of different treatment options.

[0032] Inspectors: Responsible for storing and providing the results of "professional medical examinations" to ensure the consistency and confidentiality of medical data. The inspection is divided into two steps: accurately interpreting the inspection request and determining the required tests, and communicating the results to the doctor according to the determined inspection items.

[0033] Patients: Played by a machine learning model, not only describing symptoms, but also including lifestyle, events, personality traits, etc., to enhance the authenticity of role-playing.

[0034] Medical directors: Evaluate the diagnostic performance of intern doctors, ask questions and evaluate the diagnosis, treatment plans, etc. of intern doctors.

[0035] Clinical interview: It is the communication between doctors and patients during the diagnosis and treatment process, collecting medical history and symptom information, which is used to evaluate the patient's health status and determine further examination and treatment plans.

[0036] Interactive evaluation: Interactive evaluation refers to a two-way communication between a system or machine and a person in a certain environment, and performance evaluation is carried out during this process. In the medical field, this may involve evaluating how medical AI interacts with patients or doctors for diagnosis.

[0037] Agent cooperation: Also known as agent collaboration, it refers to a multi-agent system, for example, an artificial intelligence hospital (also known as AIHospital), where different agents perform their respective roles and work together to complete common tasks or goals. In the field of medical AI, this may involve simulating the agent cooperation of roles such as intern doctors, patients, and attending doctors for medical visit analysis and treatment plan formulation.

[0038] Example 1

[0039] According to the embodiments of the present invention, an embodiment of a method for role interaction is also provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order than here.

[0040] The method embodiment provided by the first embodiment of this application can be executed on a mobile terminal, a computer terminal, or a similar computing device. Figure 1 The hardware structure block diagram of a computer terminal (or mobile device) for implementing the role interaction method is shown. As Figure 1As shown, the computer terminal 10 (or mobile device 10) may include one or more processors (shown as 102a, 102b, ……, 102n in the figure) (the processor may include, but is not limited to, processing devices such as a microprocessor MCU or a programmable logic device FPGA), a memory 104 for storing data, and a transmission module 106 for communication functions. In addition, it may further include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of the BUS bus), a network interface, a power supply, and / or a camera. Those of ordinary skill in the art can understand that Figure 1 the structure shown is only schematic and does not limit the structure of the above-mentioned electronic device. For example, the computer terminal 10 may further include more or fewer components than Figure 1 shown therein, or have a different configuration from Figure 1 that shown.

[0041] It should be noted that the above one or more processors and / or other data processing circuits are generally referred to as "data processing circuits" herein. The data processing circuit may be embodied in software, hardware, firmware, or any combination thereof, in whole or in part. In addition, the data processing circuit may be a single independent processing module, or be incorporated in whole or in part into any one of the other elements in the computer terminal 10 (or mobile device). As involved in the embodiments of the present application, the data processing circuit is a kind of processor control (such as the selection of a variable resistor terminal path connected to an interface).

[0042] The memory 104 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the role interaction method in the embodiments of the present invention. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory 104, that is, implements the above-mentioned role interaction method. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memories, or other non-volatile solid-state memories. In some instances, the memory 104 may further include a memory remotely disposed relative to the processor, and these remote memories may be connected to the computer terminal 10 through a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an enterprise intranet, a local area network, a mobile communication network, and combinations thereof.

[0043] The transmission device 106 is used to receive or send data via a network. Specific examples of the above-mentioned network may include a wireless network provided by the communication provider of the computer terminal 10. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 can be a Radio Frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0044] The display can be, for example, a touch-screen liquid crystal display (LCD), which enables the user to interact with the user interface of the computer terminal 10 (or mobile device).

[0045] It should be noted here that, in some embodiments, the above-mentioned Figure 1 shown computer device (or mobile device) has a touch display (also referred to as a "touch screen" or "touch display screen"). In some embodiments, the above-mentioned Figure 1 shown computer device (or mobile device) has a graphical user interface (GUI), and the user can perform human-computer interaction with the GUI through finger contacts and / or gestures on the touch-sensitive surface. The human-computer interaction function here optionally includes the following interactions: creating web pages, drawing, word processing, creating electronic documents, games, video conferencing, instant messaging, sending and receiving emails, call interfaces, playing digital videos, playing digital music, and / or web browsing, etc. The executable instructions for performing the above human-computer interaction functions are configured / stored in a computer program product or readable storage medium executable by one or more processors.

[0046] Under the above operating environment, the present application provides a Figure 2 shown role interaction method, Figure 2 which is a flowchart of the role interaction method according to Embodiment 1 of the present invention. As Figure 2 shown, the method includes the following steps:

[0047] Step S202: Obtain the historical record to be used in the preset interaction environment, where the historical record is used to record the records generated during the historical interaction process in the preset interaction environment.

[0048] The above-mentioned preset interactive environment can be an environment for interacting using a machine learning model with learning ability, such as: an environment for real-time interactive diagnosis. Optionally, the above-mentioned machine learning model with learning ability can be a model such as a Generative Pretrained Transformer. Optionally, in this application, no specific limitation is imposed on the above-mentioned machine learning model with learning ability. In this application, a neural network model with learning ability is taken as an example of a large language model for illustration.

[0049] The above-mentioned historical record can be a record of real interactive processes that have occurred in the past. For example, medical diagnosis records in a medical scenario, risk assessment records in a financial scenario, case discussion records in a judicial scenario, etc. Optionally, no specific limitation is imposed on the specific content of the historical record in this application. In this application, medical diagnosis is taken as an example of the historical record for illustration.

[0050] In an alternative embodiment, a person skilled in the art can obtain the historical record to be used in the preset interactive environment through a formal method, or a user with relevant historical records can provide the historical record for use in the preset interactive environment by themselves. Optionally, no specific limitation is imposed on the method of obtaining the historical record to be used in the preset interactive environment in this application.

[0051] Step S204: Create multiple proxy roles based on the historical record, where the multiple proxy roles are used to respectively simulate multiple real roles participating in the interaction during the historical interaction process.

[0052] The above-mentioned multiple proxy roles can be multiple roles that interact during the historical interaction process. Among them, the above-mentioned multiple proxy roles can be an AI teacher, an AI doctor, an AI actor, etc., that is, teachers, doctors, actors, etc. simulated by artificial intelligence means. Optionally, no specific limitation is imposed on the type of the multiple proxy roles in this application. In this application, the multiple proxy roles are taken as examples of an AI patient, an AI intern doctor, an AI inspector, and an AI medical director for illustration.

[0053] In an alternative embodiment, after obtaining the historical record, proxy settings can be performed based on the historical record, that is, according to the types of multiple real roles involved in the interaction process in the historical record, multiple proxy roles are created.

[0054] Optionally, assuming the historical record is a risk assessment record in a financial scenario, then when creating multiple proxy roles based on the historical record, the multiple proxy roles created can be an AI investor, an AI product manager, an AI financial expert, etc.

[0055] Optionally, assuming that the historical record is a record of case discussions in a judicial scenario, when creating multiple agent roles based on the historical record, the multiple agent roles created can be AI intern lawyers, AI judges, AI senior lawyers, etc.

[0056] Optionally, when creating multiple agent roles, the following method can be used:

[0057] Collect historical interaction records: First, historical interaction records need to be collected, including information such as conversations, behaviors, and decisions. These records can come from actual conversation records, observation records, literature materials, etc.

[0058] Analyze interaction characteristics: Analyze the collected historical interaction records to understand the interaction characteristics therein. Specifically, it can include the behavior patterns, language styles, decision-making methods, etc. of the participants.

[0059] Determine agent roles: Based on the analysis results, determine the number and types of agent roles to be created. Among them, according to the different characteristics of the participants, they can be divided into multiple agent roles, such as organizers, followers, decision-makers, etc.

[0060] Design agent role characteristics: For each agent role, design its characteristics, including behavior patterns, language styles, decision-making methods, etc. to design agent role characteristics. Specifically, reasonable settings can be made according to the actual situation in the historical interaction records.

[0061] Train agent roles: Use the historical interaction records as training samples to train each agent role. Optionally, machine learning, deep learning and other technologies can be used to let the agent roles learn the characteristics in the historical interaction and conduct simulated interactions.

[0062] Evaluate and optimize: Evaluate the trained agent roles to see if they can accurately simulate the participants in the historical interaction records. If there are deficiencies, they can be optimized and adjusted until the expected effect is achieved.

[0063] Step S206: Control multiple agent roles to conduct target interactions in a preset interaction environment.

[0064] The above target interaction can be to simulate real-time interaction diagnosis using the multiple agent roles created.

[0065] In an alternative embodiment, after creating multiple agent roles, the defined personas of the multiple agent roles can be combined to control the interaction of the multiple agent roles. Optionally, taking the above-mentioned preset interaction session as a medical consultation environment as an example for illustration, assuming the preset interaction session is a medical consultation environment, the multiple agent roles can be an AI patient, an AI intern doctor, an AI inspector, and an AI medical director respectively. The AI patient can first describe the condition to multiple AI intern doctors, each AI intern doctor among the multiple AI intern doctors can give corresponding diagnostic results respectively, and then the AI inspector records the symptoms of the AI patient and the diagnostic results given by each AI intern doctor, and submits them to the AI medical director for evaluation and giving the final conclusion. Optionally, when the diagnostic results given by the multiple AI intern doctors are inconsistent, the AI chief doctor can organize everyone to discuss.

[0066] Figure 3 is a medical consultation flowchart according to an embodiment of the present application, as Figure 3 shown, the intern doctor is responsible for constructing a medical role using the medical visit information, the patient is responsible for explaining basic information to the intern doctor, such as at least one of the chief complaint, current medical history and past medical history, personal history and family history. The inspector is responsible for performing professional medical examinations, such as physical examinations, auxiliary examinations, etc. The medical director is responsible for diagnosis and treatment, such as giving diagnostic results, giving diagnostic bases, performing treatment, etc. Optionally, the entire interaction process can be that the patient tells the intern doctor: Hello doctor, I've been having some pain in my heart recently. The intern doctor gives advice: You should first do a complete blood count test. Then the patient goes to tell the inspector: The doctor asked me to do a complete blood count test. After receiving the request, the inspector records the medical examination item: 1. Blood routine. Further, after the patient gets the test results, the patient can go to the intern doctor and ask the intern doctor: My test result is Nax, is this serious? The intern doctor can give advice: The heart is missing, you need to be sent back to the factory for repair. The patient replies: Thank you, doctor. Optionally, at this time, the diagnosis of the intern doctor is 1: Symptom A, 2: Result B, 3. Reason C, 4. Treatment D. The medical director conducts a final evaluation, and the evaluation result is 1: Result B, 2. Reason C, 3. Treatment D. It should be noted that the above-mentioned patient, intern doctor, inspector, and medical director are all AI patient, AI intern doctor, AI inspector, and AI medical director, rather than real patient, intern doctor, inspector, and medical director.

[0067] Optionally, assuming that the above-mentioned preset interaction environment is a risk assessment environment in the financial scenario, opinions on a certain asset can be expressed respectively based on the created AI investors, AI product managers, and AI financial experts, and one of them can be selected as the host, responsible for recording the opinions expressed by each person. When opinions are different, the points of disagreement can be put forward and everyone can be organized to discuss until the opinions of all are unified.

[0068] Optionally, assuming that the above-mentioned preset interaction environment is a case discussion environment in the judicial scenario, multiple created AI intern lawyers can discuss the case, and one of the AI intern lawyers is responsible for recording the opinions of all. The AI senior lawyer summarizes the discussion results and leads multiple AI intern lawyers to discuss the points of disagreement among everyone until the opinions of all are unified.

[0069] Optionally, assuming that the above-mentioned preset interaction session is a machine failure diagnosis environment, the above-mentioned multiple agent roles can be different ordinary repairmen, machine owners, inspectors, and authoritative repairmen. Optionally, the machine owner can describe the current situation of the machine, and each ordinary repairman gives a judgment result. The inspector is responsible for recording the judgment results of each ordinary repairman and the current situation of the machine, and the authoritative repairman summarizes the opinions of different ordinary repairmen to obtain a final conclusion. Optionally, in the case of different opinions, the authoritative repairman can organize everyone to discuss. It should be noted that the above-mentioned ordinary repairmen, machine owners, inspectors, and authoritative repairmen are all AI ordinary repairmen, AI machine owners, AI inspectors, and AI authoritative repairmen.

[0070] In the embodiment of the present invention, by obtaining the historical record to be used in the preset interaction environment, where the historical record is used to record the records generated during the historical interaction process in the preset interaction environment; creating multiple agent roles based on the historical record, where the multiple agent roles are used to respectively simulate multiple real roles participating in the interaction during the historical interaction process; controlling the multiple agent roles to perform target interaction in the preset interaction environment. It is easy to notice that by obtaining the historical record to be used in the preset interaction environment and creating multiple agent roles according to the historical record, the agent roles can be controlled to simulate the target interaction in the preset interaction environment, that is, the entire target interaction process is simulated through multiple agent roles, avoiding the task model of the user asking questions and the artificial intelligence replying in the traditional process of using artificial intelligence technology for interaction. Thus, the interaction ability of the artificial intelligence itself can be greatly explored, and the relevance and accuracy of the generated reply information are improved, thereby solving the technical problem that the relevance and accuracy of the generated reply information are relatively low in the related technology when generating information through the artificial intelligence method.

[0071] In the above embodiments of the present application, creating multiple agent roles based on historical records includes: classifying the recorded content of the historical records to obtain a classification result, where the classification result is used to determine various category association information to be used for creating multiple agent roles; constructing a target prompt word based on the classification result, where the target prompt word is used to determine the division of labor and responsibilities of the agent roles; and using the target prompt word to drive a machine learning model to create multiple agent roles.

[0072] The above classification result can be the result obtained by classifying the recorded content of the historical records, that is, the types of the recorded content of the historical records. Optionally, assuming it is in a medical diagnosis scenario, the above classification result can include personal information, medical history, and diagnosis records. Among them, personal information, that is, personal identity information such as the patient's name, age, gender, contact information, etc.; medical history, that is, historical information related to medicine such as the patient's medical history, surgical records, drug allergy conditions, etc.; and diagnosis records, that is, records related to the specific diagnosis and treatment process such as the patient's visit time, diagnosis result, treatment plan, doctor's advice, etc.

[0073] The above various category association information can be the basic information in the classification result. Optionally, assuming it is in a medical diagnosis scenario, the above various category association information includes the patient's basic information, professional medical examinations, and diagnosis and treatment.

[0074] The above target prompt word can be used to prompt multiple agent roles during their interaction after creating multiple agent roles using artificial intelligence technology, so as to facilitate the understanding of the background knowledge of different roles, that is, to guide the roles to interact.

[0075] In an alternative embodiment, when creating multiple agent roles based on historical records, the recorded content of the historical records can be classified first to obtain a classification result. Further, a target prompt word is constructed based on the classification result, so that the target prompt word can be used to drive a machine learning model to create multiple agent roles.

[0076] Figure 4 It is a schematic diagram of the target prompt word of an AI intern doctor according to an embodiment of the present application. As Figure 4 shown, the system information of the target prompt word can include an introduction to the identity of the AI intern doctor and a guide to the interaction process of the AI intern doctor. Specifically, it can include:

[0077] You are a professional doctor A.

[0078] You are making a diagnosis for a patient. The symptoms and examination results of the patient are as follows:

[0079] [Symptoms and results].

[0080] Regarding the patient's condition, a preliminary diagnosis report has been given: [Diagnosis report of Doctor A].

[0081] (1) Next, you will receive diagnostic opinions from other doctors, which include diagnosis results, diagnostic bases, and treatment plans. You need to critically sort out the reasons and analyze the diagnostic opinions of other doctors.

[0082] (2) During this process, please pay attention to the controversial points given by the medical director.

[0083] (3) If you find that the diagnostic opinions of other doctors are more reasonable, please improve your diagnostic opinion.

[0084] (4) If you think your diagnostic opinion is more reasonable, please keep it unchanged.

[0085] Please output according to the following format:

[0086] (1) Diagnosis result ***.

[0087] (2) Diagnostic basis ***.

[0088] (3) Treatment plan ***.

[0089] Diagnostic opinion of Doctor B.

[0090] Diagnostic opinion of Doctor C.

[0091] Diagnostic opinion of the medical director.

[0092] Optionally, guided by the above target prompt words, intern doctors can be enabled to interact, increasing the authenticity of the interaction.

[0093] Figure 5 It is a schematic diagram of the target prompt words of an AI medical director according to an embodiment of the present application. As Figure 5 shown, the system information may include an introduction to the persona of the AI medical director and an introduction to the process of what the AI medical director has to do. Specifically, the system information may include:

[0094] You are a senior medical director. You are presiding over a consultation for intern doctors on a patient's condition. The participating intern doctors are Intern Doctor A, Intern Doctor B, and Intern Doctor C.

[0095] The patient's condition is as follows:

[0096] [Symptoms and examination results].

[0097] (1) You need to listen to the diagnostic opinions of each intern doctor.

[0098] (2) Please list up to three controversial points to be discussed according to their importance.

[0099] (3) Please output in the following format:

[0100] [1] ***.

[0101] [2] ***.

[0102] Intern A [Diagnostic Opinion].

[0103] Intern B [Diagnostic Opinion].

[0104] Intern C [Diagnostic Opinion].

[0105] In the above-mentioned embodiments of the present application, the recorded content of the historical records is classified to obtain a classification result, including: based on the true role division in the historical interaction process, basic information, auxiliary information, and reply information are divided from the recorded content of the historical records to obtain a classification result. Among them, the basic information is used to describe the historical basic situation of the true role of the consulting party itself, the auxiliary information is used to describe the historical reference basis for the true role of the assisting party to provide a consulting reply to the true role of the replying party, and the reply information is used to describe the historical reply content fed back by the true role of the replying party to the true role of the consulting party based on the auxiliary information.

[0106] The above-mentioned true role division can be the roles in the historical interaction process. For example, real workers, real teachers, etc. Optionally, in the present application, the true role division is taken as an example of real patients, real inspectors, real interns, and real medical directors for illustration.

[0107] The above-mentioned basic information can be the basic information content mentioned in the historical interaction process. For example, the symptoms of the patient, etc.

[0108] The above-mentioned auxiliary information can be used to represent the relevant information of the consulting party. The above-mentioned reply information can be the reply given by the replying party to the question consulted by the consulting party. The above-mentioned true role of the assisting party can be the true role that makes auxiliary information for the real consulting party. The above-mentioned true role of the replying party can be the party that makes a reply for the consulting party in the real interaction process. Among them, the consulting party can be the party that has a question to ask and needs to interact.

[0109] In an alternative embodiment, assuming it is in a medical diagnosis scenario, the above-mentioned auxiliary information can be professional medical examinations prescribed for the patient, the above-mentioned response information can be the diagnosis and treatment performed on the patient, the true role of the consulting party can be the real patient, the true role of the assisting party can be the real examiner, and the true role of the responding party can be the real intern doctor. Optionally, based on the division of labor of the true roles in the historical interaction process, the basic information in the historical interaction process, as well as the professional medical examinations prescribed for the patient and the diagnosis and treatment performed on the patient, can be divided from the content of the historical record, so as to obtain a classification result, that is, to obtain personal information, medical history, and diagnosis and treatment records, that is, to obtain a classification result.

[0110] In the above embodiment of the present application, constructing the target prompt word based on the classification result includes: constructing the target prompt word based on at least part of the information included in the classification result and the interaction session information in the preset interaction environment, where the interaction session information is used to determine the division of labor and responsibilities to be fulfilled by the target agent role among multiple agent roles in different interaction sessions.

[0111] The above-mentioned interaction session information can be the evaluation and summary session. Optionally, assuming it is in the scenario of a medical consultation, the interaction session information can be the medical director evaluation session, or the medical director summary session, etc.

[0112] In an alternative embodiment, the target prompt word can be constructed based on at least part of the information included in the classification result and the interaction session information in the preset interaction environment. For example, the target prompt word is constructed according to the diagnosis and treatment records in the classification result and the medical director evaluation session, (1) Diagnosis result (2) Treatment plan.

[0113] In the above embodiment of the present application, the multiple agent roles include: the consulting party agent role, the assisting party agent role, and the responding party agent role. Controlling the multiple agent roles to perform the target interaction in the preset interaction environment includes: using the target prompt word to drive the machine learning model to respectively simulate the communication between the consulting party agent role and the responding party agent role for the target interaction matters in the preset interaction environment, and determining the auxiliary matters to be executed, where the auxiliary matters are used to describe the reference basis for the responding party agent role to provide a consultation response to the consulting party agent role; using the target prompt word to drive the machine learning model to respectively simulate the communication between the consulting party agent role and the assisting party agent role for the auxiliary matters, and determining the processing result corresponding to the auxiliary matters; using the target prompt word to drive the machine learning model to simulate the responding party agent role to feedback the target response content of the target interaction matters to the consulting party agent role based on the processing result corresponding to the auxiliary matters.

[0114] The above-mentioned consulting party agent role, assisting party agent role, and responding party agent role can be an AI patient, an AI inspector, and an AI intern doctor respectively, that is, a patient, an inspector, and an intern doctor simulated by a machine learning model.

[0115] The above-mentioned target interaction matter can be that the intern doctor discusses the patient's condition with the patient.

[0116] The above-mentioned assisting matters can be performing medical examinations, for example, doing an electrocardiogram, doing a urine test, etc.

[0117] The above-mentioned target response content can be the clinical judgment made by the intern doctor regarding the patient's condition.

[0118] In an optional embodiment, a target prompt can be used to drive a machine learning model to simulate the communication and discussion between a patient and an intern doctor regarding the patient's condition, and determine whether to perform a medical examination. For example, simulate the patient stating their symptoms, simulate the intern doctor analyzing the patient's symptoms, and simulate the intern doctor prescribing an examination order for the patient. Further, simulate the patient showing their examination order to the intern doctor, and then continue to simulate the intern doctor making a corresponding diagnosis result based on the examination order. Optionally, the interaction process can be displayed in the form of a dialog box on the interface of the machine learning model.

[0119] In the above embodiment of the present application, the multiple agent roles include: an evaluating party agent role. The method further includes: using a target prompt to drive a machine learning model to simulate the evaluating party agent role to evaluate the target response content from multiple preset dimensions, and obtaining an evaluation result. Among them, there is a master-slave role relationship between the evaluating party agent role and the responding party agent role. The multiple preset dimensions at least include: the analysis conclusion of the responding party agent role regarding the target interaction matter, the recommended plan of the responding party agent role regarding the target interaction matter. The evaluation result is used to evaluate the performance of the responding party agent role in handling the target interaction matter.

[0120] The above-mentioned evaluating party agent role can be an AI medical director, that is, a medical director simulated by a machine learning model.

[0121] The above-mentioned multiple preset dimensions can be different diagnostic aspects, for example, evaluating the diagnosis, treatment plan, etc. of the intern doctor.

[0122] The above-mentioned master-slave role relationship can be that the medical director is the host agent, that is, the medical director evaluates the opinions given by the intern doctor and obtains a final conclusion.

[0123] In an alternative embodiment, a target prompt can also be used to drive a machine learning model simulation to simulate a medical director to evaluate the diagnoses, treatment plans, etc. given by different interns, so as to obtain an evaluation result. That is, based on the analysis conclusion of the responder agent role for the target interaction matter and the recommended plan of the responder agent role for the target interaction matter, the performance of the responder agent role in handling the target interaction matter is evaluated.

[0124] In the above embodiments of the present application, controlling multiple agent roles to perform target interactions in a preset interaction environment includes: using a target prompt to drive a machine learning model to respectively simulate the responder agent role to conduct opinion interactions on the target response content under the chairmanship of the evaluator agent role, so as to update the target response content through the cooperation mode of the responder agent role.

[0125] The above-mentioned opinion interaction can be used to represent that under the chairmanship of the medical director, interns conduct multiple rounds of discussions.

[0126] The above cooperation mode can be the cooperation mode of the machine learning model.

[0127] The above update of the target response content can be used to represent improving one's own diagnostic opinion.

[0128] In an alternative embodiment, a target prompt can be used to drive a machine learning model to respectively simulate a medical director and an intern, and simulate the medical director organizing the interns to communicate on the diagnostic opinion and treatment plan, and update and improve the diagnostic opinion and treatment plan of the intern according to the process and results of the communication.

[0129] In the above embodiments of the present application, using a target prompt to drive a machine learning model to respectively simulate the responder agent role to conduct opinion interactions on the target response content under the chairmanship of the evaluator agent role includes: using a target prompt to drive a machine learning model to simulate the evaluator agent role to integrate the target response content to obtain an integration result; using a target prompt to drive a machine learning model to simulate the evaluator agent role to determine the disagreement content corresponding to the target response content based on the integration result; using a target prompt to drive a machine learning model to simulate the evaluator agent role to host the responder agent role to conduct multiple rounds of opinion interactions based on the disagreement content until the responder agent role reaches an agreement on the disagreement content.

[0130] The above integration result can be the medical director integrating the understanding of different interns on the patient's symptoms and medical examinations.

[0131] In an alternative embodiment, a target prompt can be used to drive a machine learning model to simulate a medical director, thereby integrating the understanding of different interns regarding patient symptoms and medical examinations performed on the patient, so as to obtain integrated content. Further, a target prompt can be used to drive a machine learning model to simulate a medical director, thereby summarizing the points of disagreement in the understanding of different interns regarding patient symptoms and medical examinations performed on the patient based on the integrated content. Further, a target prompt can be used to drive a machine learning model to simulate a medical director, thereby organizing multiple interns to conduct multiple rounds of discussions until all interns reach an agreement on the disagreement content.

[0132] Figure 6 is a schematic diagram of opinion interaction according to an embodiment of the present application, as Figure 6 shown. Assume that the interns are Intern A and Intern B. The diagnoses made by Intern A are symptom AA, medical examination BB, diagnosis result F1, diagnostic basis H1, and treatment plan H. The diagnoses made by Intern B are symptom AA, medical examination EE, diagnosis result F2, diagnostic basis H2, and treatment plan L1. Then, the integrated result integrated by the medical director here is symptom AA, medical examination BB. Further, the disagreement content determined by the medical director is the controversial points, (1) FF and FJ, (2) HH and HK, (3) H and LL, and the medical director organizes Intern A and Intern B to discuss the disagreement content. After some discussion, the conclusion reached by Intern A here is diagnosis result: FF, diagnostic basis: HH, treatment plan: H, while the conclusion reached by Intern B here is diagnosis result: FJ, diagnostic basis: HK, treatment plan: LL. That is, no agreement has been reached. Therefore, further discussion is needed until an agreement is reached. The discussion ends. Finally, when an agreement is reached, the content recognized by both Intern A and Intern B is diagnosis result: FJ, diagnostic basis: HK, treatment plan: LM. That is, the final conclusion is diagnosis result: FJ, diagnostic basis: HK, treatment plan: LM.

[0133] It should be noted that for the foregoing method embodiments, for the sake of simple description, they are all expressed as a series of action combinations. However, those skilled in the art should know that the present invention is not limited by the described action sequence, because according to the present invention, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to the present invention.

[0134] Through the description of the above embodiments, those skilled in the art can clearly understand that the role interaction method according to the above embodiments can be implemented by means of software plus a necessary general hardware platform. Of course, it can also be implemented by hardware, but in many cases, the former is a better implementation method. Based on such an understanding, the technical solution of the present invention, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions for causing a terminal device (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods described in various embodiments of the present invention.

[0135] Embodiment 2

[0136] According to an embodiment of the present invention, there is also provided another role interaction method. Figure 7 It is a flowchart of the role interaction method according to Embodiment 2 of the present invention, as Figure 7 shown, and the method includes the following steps:

[0137] Step S702: Obtain historical medical records to be used in the interactive diagnosis environment, where the historical medical records are used to record the medical records generated during the historical interactive diagnosis process in the interactive diagnosis environment.

[0138] Step S704: Create multiple proxy medical roles based on the historical medical records, where the multiple proxy medical roles are used to respectively simulate multiple real medical roles participating in the interactive diagnosis during the historical interactive diagnosis process.

[0139] Step S706: Control the multiple proxy medical roles to perform auxiliary diagnosis in the interactive diagnosis environment.

[0140] The above historical medical records can be previous medication records, medical diagnosis results, image examination results, and other medical records.

[0141] In an optional embodiment, when the role interaction method can be applied to the medical field, the medical records generated during the historical interactive diagnosis process in the interactive diagnosis environment can be obtained. Further, multiple proxy medical roles can be created based on the historical medical records, that is, multiple real medical roles participating in the interactive diagnosis during the historical interactive diagnosis process are created, and the multiple proxy medical roles are controlled to perform diagnosis in the interactive diagnosis environment.

[0142] In the above embodiments of the present application, creating multiple proxy medical roles based on historical medical records includes: classifying the recorded content of historical medical records to obtain a classification result, where the classification result is used to determine various category association information to be used for creating multiple proxy medical roles; constructing a target prompt word based on the classification result, where the target prompt word is used to determine the division of labor and responsibilities of the proxy medical roles; and using the target prompt word to drive a machine learning model to create multiple proxy medical roles.

[0143] In an alternative embodiment, after obtaining the historical medical records to be used in the interactive diagnosis environment, the recorded content of the historical medical records can be classified to obtain a classification result, and a target prompt word can be constructed based on the classification result, so that the target prompt word can be used to drive a machine learning model to create multiple proxy medical roles and control the multiple proxy medical roles to perform auxiliary diagnosis in the interactive diagnosis environment.

[0144] Embodiment 3

[0145] According to an embodiment of the present invention, another role interaction method is also provided. Figure 8 is a flowchart of the role interaction method according to Embodiment 3 of the present invention, as Figure 8 shown, the method includes the following steps:

[0146] Step S802: Obtain a clinical diagnosis request through a first application programming interface, where the request data carried in the clinical diagnosis request includes: condition information;

[0147] Step S804: Return a clinical diagnosis response through a second application programming interface, where the response data carried in the clinical diagnosis response includes: an auxiliary treatment plan, which is determined by a proxy medical role created based on historical medical records for recording the medical records generated during the historical diagnosis process in the interactive diagnosis environment, and the proxy medical role is used to simulate the real medical roles participating in the interactive diagnosis during the historical interactive diagnosis process.

[0148] In an alternative embodiment, a clinical diagnosis request can be obtained through a first application programming interface on the graphical user interface of the client 80. The client 80 is connected to the server 81 via a network. After the client 80 obtains the clinical diagnosis request, the server 81 processes the condition information contained in the clinical diagnosis request to obtain a corresponding assisted treatment plan. Among them, the assisted treatment plan is determined by an agent medical role, and the agent medical role is created based on historical medical records. The historical medical records are used to record the medical records generated during the historical diagnosis process in an interactive diagnosis environment. The agent medical role is used to simulate the real medical roles participating in the interactive diagnosis during the historical interactive diagnosis process. Further, the treatment plan can be returned in the form of a clinical diagnosis response through a second application programming interface on the client 80.

[0149] Embodiment 4

[0150] According to an embodiment of the present invention, another role interaction method is also provided. Figure 9 is a flowchart of the role interaction method according to Embodiment 4 of the present invention, as Figure 9 shown, the method includes the following steps:

[0151] Step S902: Obtain the currently input clinical diagnosis dialogue request. Among them, the information carried in the clinical diagnosis dialogue request includes: condition information;

[0152] Step S904: In response to the clinical diagnosis dialogue request, return a clinical diagnosis dialogue reply. Among them, the information carried in the clinical diagnosis dialogue reply includes: an assisted treatment plan, and the assisted treatment plan is determined by an agent medical role. The agent medical role is created based on historical medical records. The historical medical records are used to record the medical records generated during the historical diagnosis process in an interactive diagnosis environment. The agent medical role is used to simulate the real medical roles participating in the interactive diagnosis during the historical interactive diagnosis process;

[0153] Step S906: Display the treatment plan within the graphical user interface.

[0154] In an alternative embodiment, the user can give a clinical diagnosis dialogue request in any form on the graphical user interface of the client 90. The client 90 is connected to the server 91 via a network. The server 91 can give an assisted treatment plan based on the clinical diagnosis dialogue request on the client 90. Among them, the assisted treatment plan is determined by an agent medical role. The agent medical role is created based on historical medical records. The historical medical records are used to record the medical records generated during the historical diagnosis process in an interactive diagnosis environment. The agent medical role is used to simulate the real medical roles participating in the interactive diagnosis during the historical interactive diagnosis process, and the treatment plan is displayed on the graphical user interface in the form of a clinical diagnosis dialogue reply.

[0155] Embodiment 5

[0156] According to an embodiment of the present invention, there is also provided a device for implementing the above-mentioned role interaction method. Figure 10 It is a schematic diagram of a role interaction device according to Embodiment 5 of the present application, as Figure 10 shown, the device includes: an acquisition module 1002, a creation module 1004, and a control module 1006.

[0157] Among them, the acquisition module 1002 is used to acquire the historical records to be used in a preset interaction environment, where the historical records are used to record the records generated during the historical interaction process in the preset interaction environment; the creation module 1004 is used to create a plurality of proxy roles based on the historical records, where the plurality of proxy roles are used to respectively simulate a plurality of real roles participating in the interaction during the historical interaction process; the control module 1006 is used to control the plurality of proxy roles to perform target interactions in the preset interaction environment.

[0158] In the above embodiment of the present application, the creation module 1004 includes: a classification unit, which is used to classify the record content of the historical records to obtain a classification result, where the classification result is used to determine various category association information to be used for creating a plurality of proxy roles; a first construction unit, which is used to construct a target prompt word based on the classification result, where the target prompt word is used to determine the division of labor and responsibilities of the proxy roles; a second construction unit, which is used to drive a machine learning model with the target prompt word to create a plurality of proxy roles.

[0159] In the above embodiment of the present application, the classification unit includes: a division sub-unit, which is used to divide the record content of the historical records into basic information, auxiliary information, and reply information based on the real role division of labor during the historical interaction process to obtain a classification result, where the basic information is used to describe the historical basic situation of the real role of the consulting party itself, the auxiliary information is used to describe the historical reference basis for the real role of the assisting party to provide a consulting reply to the real role of the replying party, and the reply information is used to describe the historical reply content fed back by the real role of the replying party to the real role of the consulting party based on the auxiliary information.

[0160] In the above embodiment of the present application, the first construction unit includes: a construction sub-unit, which is used to construct a target prompt word based on at least part of the information included in the classification result and the interaction session information in the preset interaction environment, where the interaction session information is used to determine the division of labor and responsibilities to be fulfilled by the target proxy role among the plurality of proxy roles in different interaction sessions.

[0161] In the above embodiments of the present application, the control module 1006 includes: a first driving unit, configured to drive a machine learning model with a target prompt to respectively simulate the consulting party agent role and the responding party agent role to communicate regarding a target interaction matter in a preset interaction environment, and determine an auxiliary matter to be executed, where the auxiliary matter is used to describe a reference basis for the responding party agent to provide a consulting response to the consulting party agent; a second driving unit, configured to drive a machine learning model with a target prompt to respectively simulate the consulting party agent role and the assisting party agent role to communicate regarding the auxiliary matter, and determine a processing result corresponding to the auxiliary matter; and a third driving unit, configured to drive a machine learning model with a target prompt to simulate the responding party agent role to feedback a target response content of the target interaction matter to the consulting party agent based on the processing result corresponding to the auxiliary matter.

[0162] In the above embodiments of the present application, the device further includes: a simulation module, configured to drive a machine learning model with a target prompt to simulate an evaluating party agent role to evaluate the target response content from multiple preset dimensions, and obtain an evaluation result, where there is a master-slave role relationship between the evaluating party agent role and the responding party agent role, and the multiple preset dimensions at least include: an analysis conclusion of the responding party agent regarding the target interaction matter, a recommended plan of the responding party agent regarding the target interaction matter, and the evaluation result is used to evaluate the performance of the responding party agent in handling the target interaction matter.

[0163] In the above embodiments of the present application, the control module 1006 further includes: an interaction unit, configured to drive a machine learning model with a target prompt to respectively simulate the responding party agent role to conduct opinion interaction on the target response content under the chairmanship of the evaluating party agent role, so as to update the target response content through the cooperation mode of the responding party agent role.

[0164] In the above embodiments of the present application, the interaction unit includes: an integration subunit, configured to drive a machine learning model with a target prompt to simulate the evaluating party agent role to integrate the target response content, and obtain an integration result; a determination subunit, configured to drive a machine learning model with a target prompt to simulate the evaluating party agent role to determine disagreement content corresponding to the target response content based on the integration result; and a simulation subunit, configured to drive a machine learning model with a target prompt to simulate the evaluating party agent role to chair the responding party agent role to conduct multiple rounds of opinion interaction based on the disagreement content until the responding party agent role reaches an agreement on the disagreement content.

[0165] It should be noted that the above-mentioned acquisition module 1002, creation module 1004, and control module 1006 correspond to steps S202 to S206 in Embodiment 1. The examples and application scenarios implemented by the modules and the corresponding steps are the same, but are not limited to the content disclosed in the above-mentioned Embodiment 1. It should be noted that the above-mentioned modules, as part of the device, can run in the computer terminal 10 provided in Embodiment 1.

[0166] It should be noted that the preferred implementation schemes involved in the above embodiments of the present application are the same as the schemes, application scenarios, and implementation processes provided in Embodiment 1, but are not limited to the schemes provided in Embodiment 1.

[0167] Embodiment 6

[0168] According to an embodiment of the present invention, there is also provided another device for implementing the above-mentioned role interaction method. Figure 11 It is a schematic diagram of a role interaction device according to Embodiment 6 of the present application, as Figure 11 shown, the device includes: an acquisition module 1102, a creation module 1104, and a control module 1106.

[0169] Among them, the acquisition module 1102 is used to acquire historical medical records to be used in an interactive diagnosis environment, where the historical medical records are used to record the medical records generated during the historical interactive diagnosis process in the interactive diagnosis environment; the creation module 1104 is used to create multiple proxy medical roles based on the historical medical records, where the multiple proxy medical roles are used to respectively simulate multiple real medical roles participating in the interactive diagnosis during the historical interactive diagnosis process; the control module 1106 is used to control the multiple proxy medical roles to perform auxiliary diagnosis in the interactive diagnosis environment.

[0170] It should be noted that the above-mentioned acquisition module 1102, creation module 1104, and control module 1106 correspond to steps S702 to S706 in Embodiment 2. The examples and application scenarios implemented by the modules and the corresponding steps are the same, but are not limited to the content disclosed in the above-mentioned Embodiment 1. It should be noted that the above-mentioned modules, as part of the device, can run in the computer terminal 10 provided in Embodiment 1.

[0171] It should be noted that the preferred implementation schemes involved in the above embodiments of the present application are the same as the schemes, application scenarios, and implementation processes provided in Embodiment 2, but are not limited to the schemes provided in Embodiment 2.

[0172] Embodiment 7

[0173] According to an embodiment of the present invention, there is also provided another device for implementing the above-mentioned role interaction method. Figure 12 It is a schematic diagram of a role interaction device according to Embodiment 7 of the present application, asFigure 12 As shown, the device includes: an acquisition module 1202 and a return module 1204.

[0174] Among them, the acquisition module 1202 is used to obtain a clinical diagnosis request through a first application programming interface. Among them, the request data carried in the clinical diagnosis request includes: disease condition information; the return module 1204 is used to return a clinical diagnosis response through a second application programming interface. Among them, the response data carried in the clinical diagnosis response includes: an auxiliary treatment plan, and the auxiliary treatment plan is determined by an agent medical role. The agent medical role is created based on historical medical records, and the historical medical records are used to record the medical records generated during the historical diagnosis process in an interactive diagnosis environment. The agent medical role is used to simulate the real medical roles participating in the interactive diagnosis during the historical interactive diagnosis process.

[0175] It should be noted here that the above acquisition module 1202 and return module 1204 correspond to steps S802 to S804 in Embodiment 3. The examples and application scenarios implemented by the module and the corresponding steps are the same, but are not limited to the content disclosed in the above Embodiment 1. It should be noted that the above module, as a part of the device, can run in the computer terminal 10 provided in Embodiment 1.

[0176] It should be noted that the preferred implementation schemes involved in the above embodiments of the present application are the same as the schemes, application scenarios, and implementation processes provided in Embodiment 3, but are not limited to the schemes provided in Embodiment 3.

[0177] Embodiment 8

[0178] According to an embodiment of the present invention, there is also provided another device for implementing the above role interaction method. Figure 13 It is a schematic diagram of a role interaction device according to Embodiment 8 of the present application. As Figure 13 shown, the device includes: an acquisition module 1302 and a return module 1304.

[0179] Among them, the acquisition module 1302 is used to obtain the currently input clinical diagnosis dialogue request. Among them, the information carried in the clinical diagnosis dialogue request includes: disease condition information; the return module 1304 is used to return a clinical diagnosis dialogue reply in response to the clinical diagnosis dialogue request. Among them, the information carried in the clinical diagnosis dialogue reply includes: an auxiliary treatment plan, and the auxiliary treatment plan is determined by an agent medical role. The agent medical role is created based on historical medical records, and the historical medical records are used to record the medical records generated during the historical diagnosis process in an interactive diagnosis environment. The agent medical role is used to simulate the real medical roles participating in the interactive diagnosis during the historical interactive diagnosis process; display the treatment plan within the graphical user interface.

[0180] It should be noted that the above-mentioned acquisition module 1302 and return module 1304 correspond to steps S902 to S904 in Embodiment 4. The examples and application scenarios implemented by the module and the corresponding steps are the same, but are not limited to the content disclosed in the above-mentioned Embodiment 1. It should be noted that the above-mentioned module, as a part of the device, can run in the computer terminal 10 provided in Embodiment 1.

[0181] It should be noted that the preferred implementation schemes involved in the above-mentioned embodiments of the present application are the same as the schemes, application scenarios, and implementation processes provided in Embodiment 4, but are not limited to the schemes provided in Embodiment 4.

[0182] Embodiment 9

[0183] An embodiment of the present invention can provide a computer terminal, and the computer terminal can be any computer terminal device in a computer terminal group. Optionally, in this embodiment, the above-mentioned computer terminal can also be replaced with a terminal device such as a mobile terminal.

[0184] Optionally, in this embodiment, the above-mentioned computer terminal can be located in at least one of multiple network devices in a computer network.

[0185] In this embodiment, the above-mentioned computer terminal can execute program codes of the following steps in the role interaction method: obtaining a historical record to be used in a preset interaction environment, where the historical record is used to record the records generated during the historical interaction process in the preset interaction environment; creating a plurality of proxy roles based on the historical record, where the plurality of proxy roles are used to respectively simulate a plurality of real roles participating in the interaction during the historical interaction process; controlling the plurality of proxy roles to perform a target interaction in the preset interaction environment.

[0186] Optionally, Figure 14 is a structural block diagram of a computer terminal according to an embodiment of the present invention. As Figure 14 shown, the computer terminal A may include: one or more (only one is shown in the figure) processors 1402, a memory 1404, and the computer terminal A.

[0187] Among them, the memory 1404 can be used to store software programs and modules, such as the program instructions / modules corresponding to the role interaction method and device in the embodiments of the present invention. The processor 1402 executes various functional applications and data processing by running the software programs and modules stored in the memory, that is, implements the above-mentioned role interaction method. The memory may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memories, or other non-volatile solid-state memories. In some instances, the memory may further include a memory remotely provided with respect to the processor, and these remote memories can be connected to the terminal A through a network. Examples of the above network include but are not limited to the Internet, enterprise intranet, local area network, mobile communication network, and combinations thereof.

[0188] The processor can call the information and application programs stored in the memory through the transmission device to execute the following steps: obtaining the historical record to be used in the preset interaction environment, where the historical record is used to record the records generated during the historical interaction process in the preset interaction environment; creating a plurality of proxy roles based on the historical record, where the plurality of proxy roles are used to respectively simulate a plurality of real roles participating in the interaction during the historical interaction process; controlling the plurality of proxy roles to perform target interactions in the preset interaction environment.

[0189] Optionally, the above processor can also execute the program code of the following steps: classifying the record content of the historical record to obtain a classification result, where the classification result is used to determine various category association information to be used for creating a plurality of proxy roles; constructing a target prompt word based on the classification result, where the target prompt word is used to determine the proxy role division of labor and proxy role responsibilities; using the target prompt word to drive a machine learning model to create a plurality of proxy roles.

[0190] Optionally, the above processor can also execute the program code of the following steps: dividing the record content of the historical record into basic information, auxiliary information, and reply information based on the real role division of labor in the historical interaction process to obtain a classification result, where the basic information is used to describe the historical basic situation of the real role of the consulting party itself, the auxiliary information is used to describe the historical reference basis for the real role of the assisting party to provide a consultation reply to the real role of the reply party, and the reply information is used to describe the historical reply content fed back by the real role of the reply party to the real role of the consulting party based on the auxiliary information.

[0191] Optionally, the above processor can also execute the program code of the following steps: constructing a target prompt word based on at least part of the information included in the classification result and the interaction session information in the preset interaction environment, where the interaction session information is used to determine the division of labor and responsibilities to be fulfilled by the target proxy role among the plurality of proxy roles in different interaction sessions.

[0192] Optionally, the above-mentioned processor may also execute the program code of the following steps: driving a machine learning model with a target prompt to respectively simulate the consulting party agent role and the responding party agent role to communicate about the target interaction matters in a preset interaction environment, and determining the auxiliary matters to be executed, where the auxiliary matters are used to describe the reference basis for the responding party agent to provide a consulting reply to the consulting party agent; driving a machine learning model with a target prompt to respectively simulate the consulting party agent role and the assisting party agent role to communicate about the auxiliary matters, and determining the processing results corresponding to the auxiliary matters; driving a machine learning model with a target prompt to simulate the responding party agent role to feedback the target reply content of the target interaction matters to the consulting party agent based on the processing results corresponding to the auxiliary matters.

[0193] Optionally, the above-mentioned processor may also execute the program code of the following steps: driving a machine learning model with a target prompt to simulate the evaluating party agent role to evaluate the target reply content from multiple preset dimensions, and obtaining an evaluation result, where there is a master-slave role relationship between the evaluating party agent role and the responding party agent role, and the multiple preset dimensions at least include: the analysis conclusion of the responding party agent role for the target interaction matters, the recommended plan of the responding party agent role for the target interaction matters, and the evaluation result is used to evaluate the performance of the responding party agent role in handling the target interaction matters.

[0194] Optionally, the above-mentioned processor may also execute the program code of the following steps: driving a machine learning model with a target prompt to respectively simulate the responding party agent role to conduct opinion interaction on the target reply content under the chairmanship of the evaluating party agent role, so as to update the target reply content through the cooperation mode of the responding party agent role.

[0195] Optionally, the above-mentioned processor may also execute the program code of the following steps: driving a machine learning model with a target prompt to simulate the evaluating party agent role to integrate the target reply content, and obtaining an integration result; driving a machine learning model with a target prompt to simulate the evaluating party agent role to determine the disagreement content corresponding to the target reply content based on the integration result; driving a machine learning model with a target prompt to simulate the evaluating party agent role to host multiple rounds of opinion interaction for the responding party agent role based on the disagreement content until the responding party agent role reaches an agreement on the disagreement content.

[0196] In an embodiment of the present invention, by obtaining historical records to be used in a preset interaction environment, where the historical records are used to record the records generated during the historical interaction process in the preset interaction environment; creating a plurality of agent roles based on the historical records, where the plurality of agent roles are used to respectively simulate a plurality of real roles participating in the interaction during the historical interaction process; controlling the plurality of agent roles to perform a target interaction in the preset interaction environment. It is easy to notice that by obtaining the historical records to be used in the preset interaction environment and creating a plurality of agent roles according to the historical records, the agent roles can be controlled to simulate the target interaction in the preset interaction environment, that is, the entire target interaction process is simulated by a plurality of agent roles, avoiding the task model of the user asking questions and the artificial intelligence replying during the process of using artificial intelligence technology for interaction in the traditional situation. Thus, the interaction ability of the artificial intelligence itself can be mined to a great extent, and the relevance and accuracy of the generated reply information are improved, thereby solving the technical problem that the relevance and accuracy of the generated reply information are relatively low when generating information by artificial intelligence in the related art.

[0197] Those of ordinary skill in the art can understand that Figure 14 the structure shown is only for illustration, and the computer terminal can also be a smart phone (such as an Android phone, an iOS phone, etc.), a tablet computer, a palm computer, and terminal devices such as Mobile Internet Devices (MID), PAD, etc. Figure 14 It does not limit the structure of the above electronic device. For example, the computer terminal 10 may further include more or fewer components (such as a network interface, a display device, etc.) than those shown Figure 14 in the figure, or have a different configuration from that shown Figure 14 in the figure.

[0198] Those of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing the relevant hardware of the terminal device through a program, and the program can be stored in a computer-readable storage medium. The storage medium may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disc, etc.

[0199] Embodiment 10

[0200] The embodiment of the present invention further provides a storage medium. Optionally, in this embodiment, the above storage medium can be used to save the program code executed by the role interaction method provided in the first embodiment above.

[0201] Optionally, in this embodiment, the above storage medium may be located in any one of the computer terminals in the computer terminal group in the computer network, or in any one of the mobile terminals in the mobile terminal group.

[0202] Optionally, in this embodiment, the storage medium is configured to store program code for performing the following steps: obtaining historical records to be used in a preset interaction environment, where the historical records are used to record the records generated during the historical interaction process in the preset interaction environment; creating a plurality of agent roles based on the historical records, where the plurality of agent roles are used to respectively simulate a plurality of real roles participating in the interaction during the historical interaction process; controlling the plurality of agent roles to perform target interactions in the preset interaction environment.

[0203] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0204] In the above embodiments of the present invention, the descriptions of the various embodiments have their own emphases. For the parts not detailed in a certain embodiment, reference may be made to the relevant descriptions of other embodiments.

[0205] In several embodiments provided in the present application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only illustrative. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed mutual coupling or direct coupling or communication connection may be through some interfaces, and the indirect coupling or communication connection of the units or modules may be in an electrical or other form.

[0206] The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0207] In addition, the functional units in the various embodiments of the present invention may be integrated in one processing unit, or each unit may exist physically alone, or two or more units may be integrated in one unit. The above integrated units can be implemented in the form of hardware or in the form of software functional units.

[0208] When the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it 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. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a 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 such as USB flash drives, read-only memories (ROMs), random access memories (RAMs), mobile hard disks, magnetic disks, or optical discs that can store program codes.

[0209] The above are only the preferred embodiments of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of the present invention.

Claims

1. A role interaction method, characterized in that, Including: Obtain historical records to be used in a preset interaction environment, where the historical records are used to record the records generated during the historical interaction process in the preset interaction environment; Create multiple agent roles based on the historical records, where the multiple agent roles are used to respectively simulate multiple real roles participating in the interaction during the historical interaction process; Control the multiple agent roles to perform target interactions in the preset interaction environment.

2. The method according to claim 1, wherein Creating the multiple agent roles based on the historical records includes: Classify the recorded content of the historical records to obtain a classification result, where the classification result is used to determine various category association information to be used for creating the multiple agent roles; Construct a target prompt based on the classification result, where the target prompt is used to determine the division of labor and responsibilities of the agent roles; Use the target prompt to drive a machine learning model to create the multiple agent roles.

3. The method according to claim 2, wherein Classifying the recorded content of the historical records to obtain the classification result includes: Based on the division of labor of the real roles during the historical interaction process, divide the recorded content of the historical records into basic information, auxiliary information, and reply information to obtain the classification result, where the basic information is used to describe the historical basic situation of the real role of the consulting party itself, the auxiliary information is used to describe the historical reference basis for the real role of the auxiliary party to provide consulting replies to the real role of the reply party, and the reply information is used to describe the historical reply content that the real role of the reply party feedbacks to the real role of the consulting party based on the auxiliary information.

4. The method according to claim 3, characterized in that, Constructing the target prompt based on the classification result includes: Construct the target prompt based on at least part of the information included in the classification result and the interaction session information in the preset interaction environment, where the interaction session information is used to determine the division of labor and responsibilities to be fulfilled by the target agent role among the multiple agent roles in different interaction sessions.

5. The method according to claim 2, wherein The multiple agent roles include: a consulting party agent role, an auxiliary party agent role, and a reply party agent role. Controlling the multiple agent roles to perform target interactions in the preset interaction environment includes: Use the target prompt to drive a machine learning model to respectively simulate the communication between the consulting party agent role and the reply party agent role for the target interaction matters in the preset interaction environment, and determine the auxiliary matters to be executed, where the auxiliary matters are used to describe the reference basis for the reply party agent role to provide consulting replies to the consulting party agent role; Use the target prompt to drive a machine learning model to respectively simulate the communication between the consulting party agent role and the auxiliary party agent role for the auxiliary matters, and determine the processing results corresponding to the auxiliary matters; Use the target prompt to drive a machine learning model to simulate the reply party agent role to feedback the target reply content of the target interaction matters to the consulting party agent role based on the processing results corresponding to the auxiliary matters.

6. The method according to claim 5, wherein The multiple agent roles include: an evaluation party agent role, and the method further includes: Drive a machine learning model using the target prompt to simulate the evaluation party agent role to evaluate the target response content from multiple preset dimensions, and obtain an evaluation result, where there is a master-slave role relationship between the evaluation party agent role and the responder agent role, and the multiple preset dimensions at least include: the analysis conclusion of the responder agent role for the target interaction matter, the recommended plan of the responder agent role for the target interaction matter, and the evaluation result is used to evaluate the performance of the responder agent role in handling the target interaction matter.

7. The method according to claim 6, characterized in that Controlling the multiple agent roles to perform target interactions in the preset interaction environment includes: Drive a machine learning model using the target prompt to respectively simulate the responder agent role to conduct opinion interactions on the target response content under the chairmanship of the evaluation party agent role, so as to update the target response content through the cooperation mode of the responder agent role.

8. The method according to claim 7, characterized in that, Driving a machine learning model using the target prompt to respectively simulate the responder agent role to conduct opinion interactions on the target response content under the chairmanship of the evaluation party agent role includes: Drive a machine learning model using the target prompt to simulate the evaluation party agent role to integrate the target response content and obtain an integration result; Drive a machine learning model using the target prompt to simulate the evaluation party agent role to determine the disagreement content corresponding to the target response content based on the integration result; Drive a machine learning model using the target prompt to simulate the evaluation party agent role to host multiple rounds of opinion interactions of the responder agent role based on the disagreement content until the responder agent role reaches an agreement on the disagreement content.

9. A role interaction method, characterized in that, Includes: Obtain the historical medical records to be used in the interactive diagnosis environment, where the historical medical records are used to record the medical records generated during the historical interactive diagnosis process in the interactive diagnosis environment; Create multiple proxy medical roles based on the historical medical records, where the multiple proxy medical roles are used to respectively simulate multiple real medical roles participating in the interactive diagnosis during the historical interactive diagnosis process; Control the multiple proxy medical roles to conduct auxiliary diagnosis in the interactive diagnosis environment.

10. The method according to claim 9, wherein Creating multiple proxy medical roles based on the historical medical records includes: Classify the recorded content of the historical medical records to obtain a classification result, where the classification result is used to determine various category association information to be used for creating the multiple proxy medical roles; Construct a target prompt based on the classification result, where the target prompt is used to determine the division of labor and responsibilities of the proxy medical roles; Drive a machine learning model using the target prompt to create the multiple proxy medical roles.

11. A role interaction method, characterized in that, Includes: Obtain a clinical diagnosis request through a first application programming interface, where the request data carried in the clinical diagnosis request includes: condition information; Return a clinical diagnosis response through a second application programming interface, where the response data carried in the clinical diagnosis response includes: an adjuvant treatment plan, which is determined by a proxy medical role created based on historical medical records for recording the medical records generated during the historical diagnosis process in an interactive diagnosis environment, and the proxy medical role is used to simulate the real medical roles participating in the interactive diagnosis during the historical interactive diagnosis process.

12. A role interaction method, characterized in that, including: Obtain a currently input clinical diagnosis dialogue request, where the information carried in the clinical diagnosis dialogue request includes: condition information; In response to the clinical diagnosis dialogue request, return a clinical diagnosis dialogue reply, where the information carried in the clinical diagnosis dialogue reply includes: an adjuvant treatment plan, which is determined by a proxy medical role created based on historical medical records for recording the medical records generated during the historical diagnosis process in an interactive diagnosis environment, and the proxy medical role is used to simulate the real medical roles participating in the interactive diagnosis during the historical interactive diagnosis process; Display the treatment plan within the graphical user interface.

13. An electronic device, characterized in that, including: A memory storing an executable program; A processor for running the program, where when the program runs, it executes the role interaction method according to any one of claims 1 to 12.

14. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored executable program, where when the executable program runs, it controls the device where the computer-readable storage medium is located to execute the role interaction method according to any one of claims 1 to 12.

15. A computer program product, characterized in that, including a computer program that, when executed by a processor, implements the role interaction method according to any one of claims 1 to 12.

Images