Replay message generation method and device, electronic equipment and storage medium

Abstract

The invention discloses a reply information generation method and device, electronic equipment and a storage medium, and relates to the technical field of human-computer interaction. The specific implementation scheme is as follows: acquiring user interaction data; according to the user interaction data, reply information is generated, the reply information at least comprises first reply content, and the information amount of the first reply content corresponds to the user interaction data. The embodiment of the invention can flexibly increase and decrease the information amount of the reply information, and is convenient for the user to understand the reply information.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of human-computer interaction, and in particular to a reply information generation method and device, an electronic device and a storage medium. BACKGROUND

[0002] With the development of vehicle technology, in the field of vehicle technology, users on the vehicle interact with the vehicle in a voice interaction manner. By recognizing the user's voice question, understanding the user's intention, determining the reply content according to the user's intention, and outputting the reply content by the vehicle.

[0003] At present, the reply content determined according to the user's intention usually contains a large amount of information, and the vehicle directly outputs the reply content, which can cause the user to be difficult to understand. SUMMARY

[0004] The present application provides a reply information generation method, device, electronic device and storage medium, which can flexibly increase or decrease the information amount of the reply information, and facilitate the user to understand the reply information.

[0005] In a first aspect, the present application provides a reply information generation method, comprising:

[0006] obtaining user interaction data;

[0007] generating reply information according to the user interaction data, the reply information at least including first reply content, the information amount of the first reply content corresponding to the user interaction data.

[0008] In a second aspect, the present application further provides a reply information generation device, comprising:

[0009] an interaction data acquisition module configured to obtain user interaction data;

[0010] a reply information generation module configured to generate reply information according to the user interaction data, the reply information at least including first reply content, the information amount of the first reply content corresponding to the user interaction data.

[0011] In a third aspect, the present application further provides an electronic device, comprising:

[0012] at least one processor; and

[0013] a memory in communication with the at least one processor; wherein

[0014] the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the reply information generation method provided by any embodiment of the present application.

[0015] In a fourth aspect, the embodiments of the present application further provide a computer readable storage medium, which stores computer instructions. The computer instructions are used to make a processor execute the reply information generation method of any of the embodiments of the present application.

[0016] The embodiments of the present application can flexibly adjust the information amount of the reply information, specifically increase or decrease the information amount of the reply information, so as to feed back the reply information which is convenient for the user to understand.

[0017] It should be understood that the content described in this part is not intended to identify key or important features of the embodiments of the present application, nor is it used to limit the scope of the present application. Other features of the present application will become apparent from the following description. BRIEF DESCRIPTION OF DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments description. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without any creative effort on the basis of these drawings.

[0019] Figure 1 is a flow chart of a reply information generation method according to an embodiment of the present application;

[0020] Figure 2 is a flow chart of another reply information generation method according to an embodiment of the present application;

[0021] Figure 3 is a flow chart of another reply information generation method according to an embodiment of the present application;

[0022] Figure 4 is an application scenario diagram of a reply information generation method according to an embodiment of the present application;

[0023] Figure 5 is a schematic diagram of a reply information generation device according to an embodiment of the present application;

[0024] Figure 6 is a structural schematic diagram of an electronic device for implementing the reply information generation method according to an embodiment of the present application. DETAILED DESCRIPTION

[0025] In the following, the technical solutions in the embodiments of the present application will be described clearly and completely with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of embodiments of the present application, rather than all embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work should belong to the scope of protection of the present application.

[0026] It should be noted that the terms "first" and "second" and the like in the specification and claims of the present application and the above-described drawings are used to distinguish similar objects, and do not necessarily have to describe a specific order or a chronological order. It should be understood that the data used in this way can be interchanged under appropriate circumstances, so that the embodiments of the application described herein can be implemented in an order other than 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 including a series of steps or units does not have to be limited to those steps or units clearly listed, but can include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0027] In the technical solutions of the embodiments of the present application, the acquisition, storage and application of user information and the like conform to the relevant legal regulations and do not violate public order and good customs.

[0028] Figure 1 A flowchart of a reply information generation method provided by the embodiments of the present application, the present embodiment can be applied to the case of replying to the user's question in the human-computer interaction process between the vehicle and the user. The method can be executed by a reply information generation device, which can be realized in the form of hardware and / or software and specifically configured in an electronic device. The electronic device can be a server or a terminal device. The terminal device can include a mobile phone, a computer, a vehicle-mounted terminal or a wearable device, etc.

[0029] Referring to Figure 1 The reply information generation method includes:

[0030] S101, acquiring user interaction data.

[0031] In an optional embodiment, the user interaction data can be data related to the interaction in the process in which the human-computer interaction device interacts with the user. Exemplarily, the human-computer interaction device can include a mobile phone, a computer, a vehicle terminal, a wearable device, or the like. In an example, the human-computer interaction device is a vehicle terminal. The vehicle interacts with the user, and the vehicle terminal obtains the question content input by the user, which can be used as the user interaction data. Meanwhile, the vehicle terminal can also obtain the user-related data, which can also be used as the user interaction data.

[0032] It should be noted that the reply information generation method in the embodiments of the present application can be executed by a server or a human-computer interaction device. When the server executes the reply information generation method in the embodiments of the present application, the user interaction data obtained by the server can be data obtained by processing the question data input by the user directly collected, for example, the human-computer interaction device processes the question data to obtain the question content and the user-related data, and determines the user interaction data according to the question content and the user-related data. When the human-computer interaction device executes the reply information generation method in the embodiments of the present application, the human-computer interaction device directly collects the question data input by the user to obtain data, directly uses the collected data as the user interaction data, or processes the collected data to use the processing result as the user interaction data. The human-computer interaction device can obtain the question data input by the user by at least one of the following manners: obtaining text information input by the user, collecting voice of the user, capturing an image of the user, or recording a video of the user.

[0033] In an optional embodiment, the human-computer interaction device that interacts with the user is a vehicle. In the process in which the vehicle interacts with the user, the vehicle terminal generates interaction data related to the human-computer interaction operation, and determines the user interaction data according to the generated interaction data.

[0034] S102, generating reply information according to the user interaction data, the reply information at least including first reply content, and an information amount of the first reply content corresponding to the user interaction data.

[0035] In an optional embodiment, the reply information can be content for replying to the user interaction data. The information amount can be a numerical value obtained by quantifying the reply information. Exemplarily, the data amount included in the reply information can be used to represent, for example, the data amount can be the number of bytes, the number of characters, or the number of words (characters). Alternatively, the data amount of the effective data in the reply information can also be used to represent the information amount of the reply information, for example, the number of entities included in the reply information is used as the information amount in the reply information, and for example, the data amount of the keywords included in the reply information is used as the information amount in the reply information.

[0036] In an optional embodiment, user interaction data is used to determine the substantive content of the response information; for example, user interaction data is used to determine the answer to the user interaction data. User interaction data is also used to determine the information content of the first response content. The first response content is the substantive content of the response information; the first response content can be understood as the answer to the user interaction data, and the information content of the first response content is determined by the user interaction data.

[0037] In one optional embodiment, the generated response information is typically text data, which can be sent to a human-computer interaction device (HCI device) to display the response information to the user in text form; alternatively, the generated response information can be an image displayed to the user; the generated response information can be corresponding audio played to the user; or the generated response information can be corresponding video played to the user, etc. Alternatively, based on the response information, at least one of the following data can be generated: a corresponding image, corresponding audio, or corresponding video, and at least one of these can be sent to the HCI device.

[0038] In one example, the user's question was: Why did the dinosaurs disappear?

[0039] For example, the reply message only includes the first reply, which contains the most information. The first reply is: Wow, that's a really good question! The dinosaurs disappeared because a huge event happened on Earth a long, long time ago. Imagine a giant rock flying from the sky and crashing into the Earth with a bang—that's a meteorite. This giant rock created a lot of dust and smoke that blocked the sky, preventing sunlight from reaching the ground. The Earth became very cold, and many plants died. The dinosaurs had no food to eat, so they slowly disappeared.

[0040] For example, the response content includes the first response, which contains little information. The first response reads: This is an excellent question. The extinction of dinosaurs was due to an asteroid impact on Earth, which caused drastic environmental changes that made survival impossible. However, this process was also complex and interesting.

[0041] For example, generating response information based on the user interaction data can be achieved by inputting the user interaction data into a pre-trained deep learning model to obtain the response information output by the deep learning model. In one example, the user interaction data can be filled into a prompt template to obtain input data, which is then input into a large language model to obtain the response information output by the large language model.

[0042] For example, generating response information based on the user interaction data can be done by: generating response instruction information based on the user interaction data, and generating response information based on the response instruction information and the user interaction data.

[0043] In an optional embodiment, the response indication information is used to determine the amount of information in the response information, and the response indication information and user interaction data are used together to generate response information that conforms to the amount of information corresponding to the response indication information.

[0044] For example, the response instruction information can be a type or a tag. Specifically, the response instruction information can include types such as little information and no follow-up questions, little information and follow-up questions, a lot of information and no follow-up questions, and a lot of information and follow-up questions.

[0045] For example, response instructions can be the model's input data. Specifically, a response instruction could be: "Please provide an answer that the child can understand for a child's question: XX, and add guiding content that matches the child's interests." Another example is: "Please provide a complete answer for an adult's question: XX." Furthermore, response instructions can be described in other ways, without specific limitations.

[0046] Specifically, generating response instruction information based on the user interaction data, and generating response information corresponding to the user interaction data based on the response instruction information, may include: inputting the user interaction data into a first language model to obtain the response instruction information output by the first language model, wherein the response instruction information may be a type or a tag; querying the prompt template corresponding to the response instruction information, filling the corresponding prompt template with the user interaction data to obtain input data, and inputting the input data into a second language model to obtain the response information output by the second language model. The first language model and the second language model may be the same or different. The first language model may also be replaced by a classification model.

[0047] For example, user interaction data is input into the first language model to obtain the response instruction information output by the first language model; this response instruction information is the prompt template. User interaction data is then filled into the response instruction information to obtain the input data. This input data is then input into the second language model to obtain the response information output by the second language model. The first and second language models can be the same or different.

[0048] For example, user interaction data is input into a first language model to obtain response indication information output by the first language model. This response indication information can be a type or a tag. The response indication information and user interaction data are then input into a second language model to obtain the response information output by the second language model. The first and second language models can be the same or different. The first language model can also be replaced by a classification model.

[0049] By processing user interaction data to obtain response instruction information, and generating response information based on the response instruction information and user interaction data, effective data that influences the amount of information in the user interaction data can be further extracted and response instruction information generated. Thus, generating response information based on response instruction information and user interaction data can improve the accuracy and flexibility of response information, while reducing the end-to-end load pressure caused by response information generated solely based on user interaction data.

[0050] This invention generates a first response content with an information content adapted to user interaction data. The response information is determined based on the first response content, and the information content of the response information can be flexibly adjusted. Specifically, the information content of the first response content can be increased or decreased, thereby making it easier for users to understand the response information.

[0051] Figure 2 A flowchart of another method for generating response information provided in an embodiment of the present invention.

[0052] In an optional embodiment, "generating reply information based on the user interaction data" is further refined as follows: determining the target information quantity corresponding to the user interaction data based on the user interaction data; generating reply information based on the user interaction data and the target information quantity, wherein the information quantity of the first reply content corresponds to the target information quantity.

[0053] It should be noted that for any parts not described in detail in the embodiments of the present invention, please refer to the description in the foregoing embodiments.

[0054] See Figure 2 The method for generating response information shown includes:

[0055] S201. Obtain user interaction data.

[0056] S202. Determine the target information amount corresponding to the user interaction data based on the user interaction data.

[0057] In an optional embodiment, the target information content is used to indicate the information content of the generated first response. The target information content can be a specific numerical value, such as the number of words. Alternatively, the target information content can be a level or type. For example, the target information content can be either "more information content" or "less information content".

[0058] In one optional embodiment, the target information content can be determined as follows: For example, a mapping relationship between user interaction data and target information content can be preset, and the target information content corresponding to the user interaction data can be queried according to the mapping relationship. Alternatively, the user interaction data can be input into a pre-trained deep learning model to obtain the target information content. Specifically, feature extraction can be performed on the user interaction data to obtain feature vectors, and the feature vectors can be decoded and classified to obtain the target information content. Another example is quantizing the user interaction data to obtain the target information content.

[0059] S203. Generate reply information based on the user interaction data and the target information amount, wherein the information amount of the first reply content corresponds to the target information amount.

[0060] In an optional embodiment, the target information content corresponding to the information content of the first response content can mean that the target information content and the information content of the first response content are the same, or that the type of the target information content and the type of the information content of the first response content are the same. Specifically, the target information content and the first response content being the same type can mean, for example, that both the target information content and the first response content are of the type with more information; or that both the target information content and the first response content are of the type with less information.

[0061] In an optional embodiment, generating response information based on the user interaction data and the target information content can be achieved by: inputting the user interaction data and the target information content into a pre-trained deep learning model to obtain the response information output by the deep learning model. In one example, the user interaction data and the target information content can be filled into a prompt template corresponding to the target information content to obtain input data, which is then input into a large language model to obtain the response information output by the large language model.

[0062] For example, generating response information based on the user interaction data and the target information quantity can be achieved by: generating response instruction information based on the user interaction data and the target information quantity, and generating response information based on the response instruction information and the user interaction data. Specifically, the user interaction data is input into the first language model to obtain the target information quantity output by the first language model, and the response instruction information is queried or generated based on the target information quantity. Input data is determined based on the user interaction data and the response instruction information. The input data is then input into the second language model to obtain the response information output by the second language model. The first and second language models can be the same or different. The first language model can also be replaced by a classification model.

[0063] This invention obtains the target information amount by processing user interaction data, and generates reply information including the first reply content corresponding to the target information amount based on the target information amount and user interaction data. The target information amount can be directly extracted from user interaction data, and reply information can be generated based on the target information amount and user interaction data, which can achieve precise control of the information amount of the first reply content.

[0064] Optionally, user interaction data includes user attribute information.

[0065] In an optional embodiment, user attribute information can refer to stable and unchanging user information. User attribute information may include user identifier (identity), gender, age, height, and cognitive level, etc. For example, a user's voiceprint can be identified through voice recognition, and the voiceprint can uniquely identify the user; similarly, a user's facial features can be identified through an image, and the facial features can uniquely identify the user. Typically, users interact with the vehicle via voice, and the user identifier can be represented by the voiceprint features identified through voice recognition. Furthermore, a user's voice, image, and video can all be used to identify the user's gender, age, and cognitive level. Users can also directly input their user identifier, gender, age, height, and cognitive level via text.

[0066] In an optional embodiment, user attribute information can be collected in multiple ways, such as by receiving text information directly input by the user, by collecting the user's voice, by collecting the user's image, and by collecting the user's video. Accordingly, the media type of the user information obtained by collecting the user's voice is audio; the media type of the user information obtained by receiving text data input by the user is text; the media type of the user information obtained by collecting the user's image is image; and the media type of the user information obtained by collecting the user's video is video.

[0067] In an optional embodiment, the user's comprehension ability and / or preferred content can be determined based on user attribute information, thereby determining the target information content based on the user's comprehension ability and / or preferred content.

[0068] For example, user attribute information includes the occupation of athlete, and the user's question is about the symbolic meaning of portrait XX. It is determined that the user has a low level of interest in the artwork, and the target information content is determined to be low.

[0069] For example, user attribute information includes the profession of e-sports personnel, and the user's question is an introduction to game XX. This indicates that the user has a high level of interest in the game, and the target information volume is determined to be large.

[0070] For example, user attribute information includes the user's age being a child's age. This indicates that the user has weak comprehension abilities and that the target information amount is relatively small.

[0071] Specifically, user attribute information may include the user's age. For example, if the user's age is less than or equal to a preset age threshold, the target information content is determined to be low, and a first reply with low information content is generated; if the user's age is greater than the preset age threshold, the target information content is determined to be high, and a first reply with high information content is generated.

[0072] Specifically, user attribute information may include user identifiers. For example, when the user identifier is outside the preset identifier range, the target information amount is determined to be low, and a first reply with low information amount is generated; when the user identifier is within the preset identifier range, the target information amount is determined to be high, and a first reply with high information amount is generated.

[0073] Specifically, user attribute information may include gender. For example, when the user's gender is primary, the target information content is determined to be low, and a first reply with low information content is generated; when the user's gender is secondary, the target information content is determined to be high, and a first reply with high information content is generated.

[0074] Specifically, user attribute information may include height. For example, when a user's height is less than or equal to a preset height threshold, the target information content is determined to be low, and a first reply with low information content is generated; when a user's height is greater than the preset height threshold, the target information content is determined to be high, and a first reply with high information content is generated.

[0075] Specifically, user attribute information may include cognitive level. For example, when a user's cognitive level is lower than a preset cognitive level, the target information amount is determined to be low, and a first response with low information amount is generated; when a user's cognitive level is higher than the preset cognitive level, the target information amount is determined to be high, and a first response with high information amount is generated.

[0076] Optionally, user attribute information may also include other content such as occupation, without specific limitations.

[0077] Furthermore, when user attribute information includes multiple items such as user ID, gender, age, height, and cognitive level, the system can select whether each item meets the corresponding condition to generate a first response with the appropriate amount of information. Alternatively, a weighted score can be calculated for each item, and a first response with the appropriate amount of information can be generated based on the weighted score.

[0078] In one example, user attribute information includes the user's age. The default age threshold is 12 years old.

[0079] When the user is under 12 years old, the target information content is determined to be low, and the first reply content with low information content is generated as follows: This is a very good question. The extinction of dinosaurs was due to an asteroid impact on Earth, which caused drastic changes in the environment, making it impossible to survive. However, this process is also very complex and interesting.

[0080] When the user is 12 years of age or older, the target information content is determined to be high. The first response with high information content is: Wow, that's a very good question! Dinosaurs disappeared because a huge event happened on Earth a long, long time ago. Imagine a giant rock flying from the sky and crashing into the Earth with a bang—that's a giant meteorite. This giant rock caused a lot of dust and smoke to block out the sky, preventing sunlight from reaching the ground. The Earth became very cold, and many plants died. Dinosaurs had no food to eat, so they slowly disappeared.

[0081] Optionally, the user interaction data includes: user attribute information; determining the target information quantity corresponding to the user interaction data based on the user interaction data includes: when the user attribute information is first user attribute information, determining a first information quantity corresponding to the first user attribute information based on the first user attribute information; when the user attribute information is second user attribute information, determining a second information quantity corresponding to the second user attribute information based on the second user attribute information; the first user attribute information and the second user attribute information are different, and the first information quantity and the second information quantity are different.

[0082] In an optional embodiment, the target information quantity differs for users with different user attribute information. Specifically, the user attribute information of the first user is designated as first user attribute information, and the target information quantity is determined as the first information quantity based on the first user attribute information; the user attribute information of the second user is designated as second user attribute information, and the target information quantity is determined as the second information quantity based on the second user attribute information. When the first user attribute information and the second user attribute information are different, the first information quantity and the second information quantity are different. For example, a mapping relationship between user attribute information and target information quantity can be preset, and the corresponding target information quantity can be determined according to the level of difficulty. Alternatively, user attribute information can be input into a deep learning model to obtain the target information quantity.

[0083] By specifically limiting user interaction data to user attribute information, we can take into account the user's comprehension ability and preferred content, determine the amount of information that matches the user's comprehension ability and preferred content, generate first response content that is easy for the user to understand, and generate first response content that the user is interested in.

[0084] In an optional embodiment, the first user attribute information includes a first age; the second user attribute information includes a second age; the first age and the second age are different, and the first information content corresponding to the first age and the second information content corresponding to the second age are different.

[0085] In one example, younger users have weaker comprehension abilities, while older users have stronger comprehension abilities. Therefore, younger users can be given content with less information to reduce the difficulty of understanding the response. Since the first age group is younger than the second age group, the first group should contain less information than the second group.

[0086] In one example, the first user is younger than the second user, and correspondingly, the first user receives more information than the second. Younger users have weaker comprehension abilities, while older users have stronger comprehension abilities. Therefore, providing younger users with more informative and engaging content can help them understand the responses.

[0087] By specifically limiting user attribute information to user age, the user's comprehension ability can be determined based on age, thereby determining the amount of information appropriate to the user's age-appropriate comprehension ability. This allows for the generation of response information corresponding to the target amount of information, providing response information that is easy for the user's age to understand.

[0088] Optional, user interaction data includes: user questions.

[0089] In an optional embodiment, the user's question content may refer to the content that the user inputs to ask the human-computer interaction device during the interaction process.

[0090] Optionally, the user interaction data includes: user question content; determining the target information content corresponding to the user interaction data based on the user interaction data includes: when the difficulty level corresponding to the user question content is a first difficulty level, determining a third information content corresponding to the user interaction data based on the user question content; when the difficulty level corresponding to the user question content is a second difficulty level, determining a fourth information content corresponding to the user interaction data based on the user question content; when the first difficulty level and the second difficulty level are different, the third information content and the fourth information content are different.

[0091] In one optional embodiment, the difficulty level of the user's question is used to determine the complexity of the answer, thereby determining the target information content. The user's question can be categorized by difficulty level, such as hardness or ease. For example, a mapping relationship between difficulty level and target information content can be preset, and the corresponding target information content can be determined based on the difficulty level. Alternatively, the user's question can be input into a deep learning model to obtain the target information content.

[0092] In an optional embodiment, the target information content varies depending on the difficulty level of the user's question. This can involve the same user inputting questions of varying difficulty. For difficult user questions, the target information content can be increased, allowing for a more informative first response to facilitate user comprehension. Conversely, for simple user questions, the target information content can be reduced, allowing for a less informative first response to facilitate quick understanding.

[0093] By specifically limiting user interaction data to user questions, the amount of information can be determined based on the difficulty of the questions, and the amount of information can be adapted to generate a first response that is easy for users to understand.

[0094] Optional, user interaction data includes: user's historical conversations.

[0095] In an optional embodiment, user history dialogue can refer to the user's historical dialogue information related to user interaction data. In practice, embodiments of the present invention can be applied to multi-turn dialogue scenarios, storing the user's historical dialogue rounds as user history dialogue. Furthermore, all dialogues sent by the user by the human-computer interaction device before the current moment can also be recorded as historical dialogue.

[0096] In an optional embodiment, the electronic device implementing the response information generation method can record user data and can locally query user-related data. For example, multiple sessions of the same user can be recorded, and a session can include at least one round of dialogue. User identifiers can be stored in correspondence with user sessions, and a corresponding relationship can be established. User historical sessions can be queried based on the user identifier in the user attribute information, thereby obtaining the user's historical dialogues.

[0097] Optionally, the user interaction data includes: user historical dialogues; determining the target information quantity corresponding to the user interaction data based on the user interaction data includes: when the known information corresponding to the user historical dialogue is first known information, determining a fifth information quantity corresponding to the user interaction data based on the user historical dialogue; when the known information corresponding to the user historical dialogue is second known information, determining a sixth information quantity corresponding to the user interaction data based on the user historical dialogue; when the first known information and the second known information are different, the fifth information quantity and the sixth information quantity are different.

[0098] In an optional embodiment, user history dialogues are used to determine known information, thereby determining the target information content based on the known information. Key information can be extracted from user history dialogues as known information. The target information content differs for different user history dialogues. In fact, the first response does not need to repeat content already provided in the user history dialogue. Therefore, the known information corresponding to the first response content can be determined based on the user history dialogues. The target information content is then determined based on the known information. For example, a mapping relationship between known information and target information content can be preset, such as an inverse relationship between the information content of known information and the target information content. The corresponding target information content is then determined based on the known information. Alternatively, the known information can be input into a deep learning model to obtain the target information content.

[0099] The amount of target information varies depending on the amount of known information. The same user's question can correspond to different user history conversations with different known information, and different users can correspond to different user history conversations with different known information.

[0100] In an optional embodiment, the user's preferences and comprehension level can be determined based on the user's historical dialogue, and the amount of information can be determined based on the user's preferences and comprehension level.

[0101] By specifically limiting user interaction data to users' historical conversations, known information can be considered, thereby reducing the duplication of known content, providing effective new responses, and reducing redundant interactions.

[0102] Optional, user interaction data includes: user behavior characteristics.

[0103] In an optional embodiment, user behavior features are used to determine a user's preferred content, etc. User behavior features can be determined based on user history conversations and user attribute information, etc. For example, user behavior features indicate the user's historical preferred content. User behavior features are retrieved based on the user identifier in the user attribute information. User behavior features can be user type or tags.

[0104] Optionally, the user interaction data includes: user behavior features; determining the target information content corresponding to the user interaction data based on the user interaction data includes: when the preference content corresponding to the user behavior features is a first preference content, determining a seventh information content corresponding to the user interaction data based on the user behavior features; when the preference content corresponding to the user behavior features is a second preference content, determining an eighth information content corresponding to the user interaction data based on the user behavior features; when the first preference content and the second preference content are different, the seventh information content and the eighth information content are different.

[0105] In one optional embodiment, user behavior characteristics are used to determine preferred content. The target information content is then determined based on this preferred content. The target information content varies depending on the specific preferred content. In reality, users tend to want more content that interests them. Therefore, the target information content can be determined based on their preferred content.

[0106] For example, a mapping relationship between the similarity between preferred content and user questions and the target information content can be pre-defined. The similarity between preferred content and user questions can then be detected, and the target information content can be queried based on the similarity and correspondence. Alternatively, preferred content and user questions can be input into a deep learning model to obtain the target information content.

[0107] The amount of target information varies depending on different user preferences. The same user's question can correspond to different user preferences, and different users can have different user preferences.

[0108] In one optional embodiment, the more similar the preferred content and the user's question, the more target information is generated; conversely, the more the preferred content and the user's question deviate from each other, the less target information is generated. For example, user behavioral characteristics such as strong curiosity or multiple multi-turn conversations in the user's historical dialogues indicate a high amount of information. Alternatively, if the user's question is related to games, and the user's behavioral characteristics include a preference for astronomy or multiple simple one-turn conversations in the user's historical dialogues, the information is low and does not include second responses.

[0109] In an optional embodiment, the user's historical dialogues and / or user behavior characteristics can be queried based on the user identifier in the user attribute information to obtain query results; user interaction data can be generated based on the query results, user attribute information, and user question content.

[0110] By limiting user interaction data to user behavior characteristics, and determining the target information content based on the preferences identified by these user behavior characteristics, more response content can be provided for questions that users are interested in.

[0111] Optionally, the user interaction data may also include vehicle data.

[0112] In an optional embodiment, the human-machine interface device may be a vehicle, and the vehicle data may be vehicle-related data. Vehicle data may include vehicle status, vehicle infotainment system status, and external environment status. Vehicle status includes, but is not limited to, at least one of vehicle speed, tire pressure, and interior temperature; vehicle infotainment system status may include, but is not limited to, at least one of foreground application usage status and background application status; external environment status includes, but is not limited to, at least one of weather, GPS positioning data, and road conditions.

[0113] In an optional embodiment, vehicle data is used as the basis for responding to user questions related to the vehicle. For example, the user question might be: "How long until the vehicle arrives?" The system can obtain vehicle speed, route, current location, real-time traffic conditions along the route, and current time, and use this data as vehicle data. Based on this data, the system calculates the vehicle's arrival time and duration. The vehicle data, combined with the user question, is used to determine the substantive content of the first response, which may be the answer.

[0114] In an optional embodiment, vehicle information may include vehicle attribute information obtained locally and real-time vehicle status data, such as vehicle model, brand, size, or structure. Vehicle status can be further subdivided into vehicle status, vehicle infotainment system status, and external environment status. More specifically, vehicle status may include, but is not limited to, at least one of vehicle speed, tire pressure, and interior temperature; vehicle infotainment system status may include, but is not limited to, at least one of foreground application usage status and background application status; and external environment status may include, but is not limited to, at least one of weather, GPS (Global Positioning System) positioning data, and road conditions.

[0115] It should be noted that the acquisition, storage, and application of user information involved in the embodiments of the present invention all comply with the provisions of relevant laws and regulations, do not violate public order and good morals, and the vehicle will only acquire user information with the user's authorization. Furthermore, the user information transmitted by the vehicle is processed data, and the vehicle will not transmit the source data of the user information to the electronic device implementing the response information generation method of the embodiments of the present invention.

[0116] In one example, a brief initial response could be: "The vehicle will arrive in XX minutes, estimated arrival time is 10:32."

[0117] For example, a first reply with a lot of information could be: The vehicle has now reached location XX, and is XX meters away from the destination. The road ahead is clear, and the vehicle will arrive in XX minutes. It is expected to arrive at 10:32.

[0118] By limiting user interaction data to vehicle data and interaction scenarios to human-vehicle interaction scenarios, the input data can be enriched based on the user's questions, making the input data more diverse. Based on the richer and more complete input data, response information can be generated, which can add or remove response information in human-vehicle interaction in vehicle scenarios, helping users to understand the real-time status of the vehicle more quickly.

[0119] Optionally, the user interaction data includes: vehicle data; determining the target information quantity corresponding to the user interaction data based on the user interaction data includes: when the operation complexity corresponding to the vehicle data is a first operation complexity, determining a ninth information quantity corresponding to the user interaction data based on the vehicle data; when the operation complexity corresponding to the vehicle data is a second operation complexity, determining a tenth information quantity corresponding to the user interaction data based on the vehicle data; when the first operation complexity and the second operation complexity are different, the ninth information quantity and the tenth information quantity are different.

[0120] Operational complexity describes the complexity of processing the information to obtain the first response. It involves analyzing user queries and vehicle data, identifying vehicle data relevant to the user query, categorizing this relevant vehicle data, and determining the operational complexity.

[0121] The target information content varies depending on the complexity of the operation. Generally, the more vehicle data required for a user's question, the more complex the process to obtain the first response, and consequently, the more first response content is provided, resulting in a higher target information content. Conversely, the less vehicle data required for a user's question, the simpler the process to obtain the first response, and consequently, the less first response content is provided, resulting in a lower target information content. Different user questions can correspond to different levels of operational complexity. Generally, higher operational complexity results in a higher target information content, while lower operational complexity results in a lower target information content. For example, a mapping relationship between operational complexity and target information content can be pre-defined, and the corresponding target information content can be determined based on the operational complexity. Alternatively, the operational complexity can be input into a deep learning model to obtain the target information content.

[0122] By specifically limiting user interaction data to vehicle data, the amount of information can be determined by considering the complexity of vehicle data processing in human-vehicle interaction scenarios. This allows for the generation of first-response content that is easy for users to understand, tailored to the complexity of vehicle data processing.

[0123] In an optional embodiment, user interaction data may include user-asked questions and user attribute information.

[0124] In one example, the user's question was: Why did the dinosaurs disappear? The user's attributes included their age, which was 10 years old. Accordingly, the first response, generating minimal information, was: This is an excellent question. The dinosaurs disappeared because an asteroid impact caused drastic environmental changes, making survival impossible. However, this process was also complex and interesting.

[0125] In one example, a user asked: "Why did the dinosaurs disappear?" The user's attributes included their age, which was 20 years old. Accordingly, the first response, generating the most information, stated: "The dinosaurs disappeared because a long time ago, a huge rock flew from the sky and crashed into the earth, creating a lot of dust and smoke that blocked out the sky. Sunlight couldn't reach the ground, the earth became very cold, and many plants died. The dinosaurs had no food to eat, so they gradually disappeared."

[0126] In an optional embodiment, user interaction data may include user questions, user attribute information, and user dialogue history.

[0127] In one example, the user's question is: Why is A's speed greater than B's speed? The user's dialogue history shows the user's first question: A travels S1 kilometers at T1, and B travels S2 kilometers at T2. Who is faster, A or B? The first response was: A's speed is greater than B's speed. User attributes include age, which is 12 years old. Accordingly, the first response, which contains less information, is: S1 / T1 = V1, therefore, V1 > V2.

[0128] In one example, the user's question is: Why is A's speed greater than B's speed? The user's dialogue history shows the user's first question: A travels S1 kilometers at T1, and B travels S2 kilometers at T2. Who is faster, A or B? The first response was: A's speed is greater than B's speed. User attributes include age, which is 12 years old. Accordingly, the first response with the most information is: S1 / T1 = V1, S2 / T2 = V2, V1 > V2.

[0129] In an optional embodiment, user interaction data may include user questions, user attribute information, and user behavior characteristics.

[0130] In one example, the user's question was: Why did dinosaurs disappear? User attributes included age (10 years old) and a preference for cute and adorable things. Accordingly, the first response, with minimal information, was: This is an excellent question. The dinosaurs disappeared because an asteroid impact caused drastic environmental changes, making survival impossible. However, this process was also complex and interesting.

[0131] In one example, a user asked: "Why did the dinosaurs disappear?" User attributes included age (10 years old) and a liking for dinosaurs. The resulting first response, rich in information, stated: "The dinosaurs disappeared because a long time ago, a huge rock flew from the sky and crashed into the earth, creating dust and smoke that blocked out the sun. The earth became very cold, and many plants died. The dinosaurs had no food, so they gradually disappeared."

[0132] In an optional embodiment, user interaction data may include user questions, user attribute information, user history conversations, and user behavior characteristics.

[0133] In one example, the user's question is: Why did the dinosaurs disappear? User attributes include age (10 years old), and past conversations include questions about dolls, indicating a preference for cute and adorable items. Accordingly, the first, less informative response is generated: This is a very good question. The dinosaurs disappeared because an asteroid impact caused drastic environmental changes, making survival impossible. However, this process was also complex and interesting.

[0134] In one example, a user's question was: "Why did the dinosaurs disappear?" User attributes included age (10 years old), a history of conversations with questions related to Tyrannosaurus Rex, and a dinosaur-related behavioral characteristic. Accordingly, the first, most informative reply was generated: "The dinosaurs disappeared because a long time ago, a huge rock flew from the sky and crashed into the earth, creating a lot of dust and smoke that blocked out the sun. The earth became very cold, and many plants died. The dinosaurs had no food, so they gradually disappeared."

[0135] In an optional embodiment, user interaction data may include user questions, user attribute information, and vehicle data.

[0136] In one example, the user's question is: "How far is the service area?" User attributes include age, which is 10 years old. Based on the vehicle's location in the vehicle data, the nearest service area is determined, the distance between the vehicle and the service area is obtained, and the travel time is calculated based on the vehicle's current speed and the distance to the service area. Accordingly, the first response, with minimal information, is generated: "Arrival in 20 minutes."

[0137] In one example, the user's question is: "How far is the service area?" User attributes include age, which is 20 years old. Based on the vehicle's location in the vehicle data, the nearest service area is determined, the distance between the vehicle and the service area is obtained, and the travel time is calculated based on the vehicle's current speed and the distance to the service area. Accordingly, the first, more informative response is generated: "The nearest service area is 20km from the current location. At the current constant speed, it is estimated to arrive in 30 minutes."

[0138] In an optional embodiment, user interaction data may include user questions, user attribute information, user history conversations, and vehicle data.

[0139] In one example, the user's question is: "How far is the restaurant?" User attributes include age, which is 10 years old. The user's history shows their first question: "Are there any noodle shops nearby?" and their first reply: "Yes, there are noodle shops." Based on the vehicle's location in the vehicle data, the nearest noodle shop is determined, the distance between the vehicle and the noodle shop is obtained, and the travel time is calculated based on the vehicle's current speed and the distance to the noodle shop. Accordingly, the first reply, with less information, is generated: "Arriving in 20 minutes."

[0140] In one example, the user's question is: "How far is the restaurant?" User attributes include age, which is 20 years old. The user's history shows their first question: "Are there any noodle shops nearby?" and their first reply: "Yes, there are noodle shops." Based on the vehicle's location in the vehicle data, the nearest noodle shop is determined, the distance between the vehicle and the shop is obtained, and the travel time is calculated based on the vehicle's current speed and the distance to the shop. Accordingly, the first reply, which contains more information, is generated as: "The nearest noodle shop is 20km from the current location. At the current constant speed, it is estimated to arrive in 30 minutes."

[0141] In an optional embodiment, user interaction data may include user questions, user attribute information, user behavior characteristics, and vehicle data.

[0142] In one example, the user's question is: "What are some fun things to do nearby?" User attributes include age (10 years old) and a preference for swimming. Based on the vehicle's location in the vehicle data, the nearest swimming pool is determined, the distance between the vehicle and the pool is calculated, and the travel time is calculated using the vehicle's current speed and the distance to the pool. Accordingly, the first, less informative response is: "There's a swimming pool about 1km away."

[0143] In one example, the user's question is: "What are some fun things to do nearby?" User attributes include age (20 years old) and a preference for swimming. Based on the vehicle's location in the vehicle data, the nearest swimming pool is determined, the distance between the vehicle and the pool is calculated, and the travel time is calculated using the vehicle's current speed and the distance to the pool. The resulting first, more informative response is: "There's a swimming pool about 1km away; it should take about 10 minutes to get there."

[0144] In an optional embodiment, user interaction data may include user questions, user attribute information, user history conversations, user behavior characteristics, and vehicle data.

[0145] In one example, the user's question is: "How long will it take to get to a fun place?" User attributes include age, which is 10 years old. The user's history shows their first question: "What fun places are nearby?" and their first response: "There are swimming pools, basketball courts, and badminton courts." The user's behavioral characteristic is: "I like swimming." Based on the vehicle's location in the vehicle data, the nearest swimming pool is determined. The distance between the vehicle and the swimming pool is obtained, and the travel time is calculated based on the vehicle's current speed and the distance to the swimming pool. Accordingly, the first response, which contains less information, is generated: "There is a swimming pool more than 1 km away."

[0146] In one example, the user's question is: "What are some fun things to do nearby?" User attributes include age, which is 20 years old. The user's history shows their first question: "What are some fun places to do nearby?" The first response was: "There are swimming pools, basketball courts, and badminton courts." The user's behavioral characteristic is: "I like playing basketball." Based on the vehicle's location in the vehicle data, the nearest swimming pool is determined. The distance between the vehicle and the swimming pool is obtained, and the travel time is calculated based on the vehicle's current speed and the distance to the swimming pool. Accordingly, the first response, which generates more information, is: "There is a basketball court 2km away, estimated to take 20 minutes to reach. There is a swimming pool 1km away, estimated to take 10 minutes to reach. There are no badminton courts nearby."

[0147] In an optional embodiment, user interaction data may include user attribute information and user history conversations.

[0148] In an optional embodiment, user interaction data may include user attribute information and user behavior characteristics.

[0149] In an optional embodiment, user interaction data may include: user attribute information, user history conversations, and user behavior characteristics.

[0150] In an optional embodiment, user interaction data may include user attribute information and vehicle data.

[0151] In an optional embodiment, user interaction data may include: user attribute information, user history conversations, and vehicle data.

[0152] In an optional embodiment, user interaction data may include: user attribute information, user behavior characteristics, and vehicle data.

[0153] In an optional embodiment, user interaction data may include: user attribute information, user history conversations, user behavior characteristics, and vehicle data.

[0154] In an optional embodiment, user interaction data may include: user questions and user conversation history.

[0155] In an optional embodiment, user interaction data may include: user-asked questions and user behavior characteristics.

[0156] In an optional embodiment, user interaction data may include: user questions, user dialogue history, and user behavior characteristics.

[0157] In an optional embodiment, user interaction data may include: user-generated questions and vehicle data.

[0158] In an optional embodiment, user interaction data may include: user questions, user history conversations, and vehicle data.

[0159] In an optional embodiment, user interaction data may include: user questions, user behavior characteristics, and vehicle data.

[0160] In an optional embodiment, user interaction data may include: user questions, user dialogue history, user behavior characteristics, and vehicle data.

[0161] In an optional embodiment, user interaction data may include: user history conversations and user behavior characteristics.

[0162] In an optional embodiment, user interaction data may include: user history conversations and vehicle data.

[0163] In an optional embodiment, user interaction data may include: user history conversations, user behavior characteristics, and vehicle data.

[0164] In an optional embodiment, user interaction data may include user behavior characteristics and vehicle data.

[0165] When user interaction data includes at least two items, the target information content corresponding to the user interaction data can be determined by: determining the information content corresponding to each item; weighting the determined information content of each item to obtain a weighted result; and using the weighted result as the target information content corresponding to the user interaction data.

[0166] In one example, the user interaction data of the first user includes at least one of the following: first user attribute information, user questions of a first level of difficulty, user history dialogues of a first level of known information, user behavior characteristics of a first level of preference content, and vehicle data of a first level of operational complexity. The target information content corresponding to the user interaction data can be obtained by weighting at least one of the first, third, fifth, seventh, and ninth information contents.

[0167] Furthermore, for the same user question, different combinations of user interaction data result in different first responses.

[0168] By flexibly combining and configuring user interaction data, the flexibility of response content is increased to meet diverse interaction needs and adapt to various interaction scenarios.

[0169] Figure 3 The following is a flowchart of another method for generating response information provided in an embodiment of the present invention.

[0170] In an optional embodiment, the response information may further include a second response.

[0171] It should be noted that for any parts not described in detail in the embodiments of the present invention, please refer to the description in the foregoing embodiments.

[0172] See Figure 3 The method for generating response information shown includes:

[0173] S301. Obtain user interaction data.

[0174] S302. Generate reply information based on the user interaction data. The reply information includes at least a first reply content, the amount of information in the first reply content corresponding to the user interaction data. The reply information also includes a second reply content, the second reply content being guiding follow-up questions corresponding to the user interaction data.

[0175] In an optional embodiment, the second response content is optional. The response information may or may not include the second response content. Whether the response information includes the second response content is determined by user interaction data. As in the previous example, the decision to include the second response content may also be made by response instruction information determined based on user interaction data.

[0176] In practice, when the initial response only contains a limited amount of information, and the user is showing strong curiosity or interest, it's helpful to prompt them to ask further questions beyond the initial reply. This can be achieved by adding guiding questions to the response, encouraging the user to inquire about more than just the initial response. These guiding questions, building upon the initial reply, are designed to encourage further inquiry. They can include suggestions, counter-questions, and recommendations.

[0177] In one example, as in the previous one, the reply information excluding the second reply content is: The main reason for the extinction of dinosaurs was the asteroid impact on Earth, which led to drastic environmental changes (first reply content).

[0178] For example, the reply message including the second reply is: The main reason for the extinction of dinosaurs was an asteroid impact on Earth, which caused drastic environmental changes (first reply). Do you want to know the story of this large meteorite impact on Earth (second reply)?

[0179] Determining whether a second response is included in the reply by analyzing user interaction data allows users to choose whether to receive the full reply. Furthermore, by using the second response to guide follow-up questions, combined with the first response that can be flexibly adjusted in terms of information content, the system can prompt and guide users to continue asking questions to obtain the full reply while maintaining the conciseness of the response. This allows users to ask questions step by step, reflecting a realistic dialogue scenario, thereby improving the authenticity of the reply and enhancing the user experience.

[0180] Optionally, generating response information based on the user interaction data includes: determining the guidance type corresponding to the user interaction data based on the user interaction data; generating response information based on the user interaction data and the guidance type, wherein the response information includes whether the second response content corresponds to the guidance type.

[0181] The guidance type is used to determine whether to generate a second response. For example, guidance types include guided and unguided types. Whether the response information includes a second response corresponds to the guidance type as follows: if the guidance type is guided, the response information includes a second response; if the guidance type is unguided, the response information does not include a second response.

[0182] In an optional embodiment, generating response information based on the user interaction data and the guidance type can be achieved by: inputting the user interaction data and the guidance type into a pre-trained deep learning model to obtain the response information output by the deep learning model. In one example, the user interaction data and the guidance type can be filled into a prompt template corresponding to the guidance type to obtain input data, which is then input into a large language model to obtain the response information output by the large language model.

[0183] For example, generating response information based on the user interaction data and the guidance type can be achieved by: generating response instruction information based on the user interaction data and the guidance type, and generating response information based on the response instruction information and the user interaction data. Specifically, the user interaction data is input into the first language model to obtain the guidance type output by the first language model, and response instruction information is queried or generated based on the guidance type. Input data is determined based on the user interaction data and the response instruction information. The input data is then input into the second language model to obtain the response information output by the second language model. The first and second language models can be the same or different. The first language model can also be replaced by a classification model.

[0184] In an optional embodiment, generating response information based on the user interaction data includes: determining the target information volume corresponding to the user interaction data; determining the guidance type corresponding to the user interaction data; and generating response information based on the user interaction data, the target information volume, and the guidance type.

[0185] In an optional embodiment, generating response information based on the user interaction data, target information content, and guidance type can be achieved by inputting the user interaction data, target information content, and guidance type into a pre-trained deep learning model to obtain the response information output by the deep learning model. In one example, the user interaction data and guidance type can be filled into prompt templates corresponding to the target information content and guidance type to obtain input data, which is then input into a large language model to obtain the response information output by the large language model.

[0186] For example, generating response information based on the user interaction data, target information content, and guidance type can be achieved by: generating response instruction information based on the user interaction data, target information content, and guidance type; and generating response information based on the response instruction information and user interaction data. Specifically, the user interaction data is input into a first language model to obtain the target information content and guidance type output by the first language model, which are then used as the response instruction information. Input data is determined based on the user interaction data and the response instruction information. The input data is then input into a second language model to obtain the response information output by the second language model. The first and second language models can be the same or different. The first language model can also be replaced by a classification model.

[0187] This invention provides a method to obtain a guidance type by processing user interaction data, and to generate response information, including or excluding a second response, based on the guidance type and user interaction data. The method can directly extract the guidance type from the user interaction data and generate response information based on the guidance type and user interaction data, thereby enabling precise control over whether to generate a second response.

[0188] In an optional embodiment, the user interaction data includes: user attribute information; determining the guidance type corresponding to the user interaction data based on the user interaction data includes: when the user attribute information is first user attribute information, determining a first guidance type corresponding to the first user attribute information based on the first user attribute information; when the user attribute information is second user attribute information, determining a second guidance type corresponding to the second user attribute information based on the second user attribute information; the first user attribute information is different from the second user attribute information, and the first guidance type is different from the second guidance type.

[0189] In an optional embodiment, user attribute information is further used to indicate whether to generate second response content. The user's comprehension level can be determined based on the user attribute information, and based on the user's comprehension level, it can be determined whether to generate second response content adapted to the user's comprehension level. Furthermore, the user's content interests can be determined based on the user attribute information, and based on the user's content interests, it can be determined whether to generate second response content adapted to the user's content interests. Different guidance types are used for users with different user attribute information. For example, a mapping relationship between user attribute information and guidance types can be preset, and the corresponding guidance type can be determined based on the user attribute information.

[0190] For example, user attribute information includes the occupation of athlete, and the user's question is about the symbolic meaning of portrait XX. If it is determined that the user has a low level of interest in the artwork, a second response may not be generated.

[0191] For example, if a user's attribute information includes their profession as an e-sports player and their question is about an introduction to a certain game, and it's determined that the user has a high level of interest in the game, a second response can be generated.

[0192] For example, if user attribute information includes the user's age as a child, it indicates that the user has weak comprehension skills, and a second response can be generated.

[0193] By determining whether to generate a second response based on user attribute information in user interaction data, the user's comprehension ability and preferred content can be taken into account. This allows for a better understanding of the user's comprehension ability and preferred content, helping them to access more interesting and understandable content.

[0194] Optionally, the user interaction data includes: user question content; determining the guidance type corresponding to the user interaction data based on the user interaction data includes: when the difficulty level corresponding to the user question content is a first difficulty level, determining a third guidance type corresponding to the user interaction data based on the user question content; when the difficulty level corresponding to the user question content is a second difficulty level, determining a fourth guidance type corresponding to the user interaction data based on the user question content; when the first difficulty level and the second difficulty level are different, the third guidance type and the fourth guidance type are different.

[0195] In an optional embodiment, questions from difficult users can have a second response added, allowing for more rounds of feedback with the first response. This facilitates the user's understanding of the detailed and complete response, thus determining the guidance type as guided. Questions from easy users can be answered without a second response, allowing for fewer rounds of feedback with the first response. This facilitates the user's quick understanding of the first response, thus determining the guidance type as unguided. The guidance type varies depending on the difficulty level of the user's question. For example, a mapping relationship between difficulty level and guidance type can be preset, and the corresponding guidance type can be determined based on the difficulty level.

[0196] By specifically limiting user interaction data to user questions, we can determine whether to add a second response based on the difficulty level of the questions. We can also add appropriate guidance content to match the difficulty level of the questions, helping users obtain more detailed and complete responses.

[0197] Optionally, the user interaction data includes: user history dialogue; determining the guidance type corresponding to the user interaction data based on the user interaction data includes: when the known information corresponding to the user history dialogue is first known information, determining a fifth guidance type corresponding to the user interaction data based on the user history dialogue; when the known information corresponding to the user history dialogue is second known information, determining a sixth guidance type corresponding to the user interaction data based on the user history dialogue; when the first known information and the second known information are different, the fifth guidance type and the sixth guidance type are different.

[0198] In one optional embodiment, the user can choose whether to repeatedly obtain known information. Correspondingly, a second response can be added to the known information, thus determining the guidance type as "guided," guiding the user to obtain previously replied content that they still want. Alternatively, the known information can be used to determine whether the user has explored content in the same domain as their question, thereby judging their interest in that domain. If the user is determined to be interested, a second response can be added to that domain, thus determining the guidance type as "guided." The guidance type differs for different known information and historical dialogues. For example, a mapping relationship between the similarity between known information and the user's question and the guidance type can be preset, the similarity between the known information and the user's question can be calculated, and the corresponding guidance type can be determined based on the similarity.

[0199] By specifically limiting user interaction data to users' historical conversations, known information can be considered to determine whether to add a second response, which can help users retrieve previously accessed content or content of interest.

[0200] Optionally, the user interaction data includes: user behavior characteristics; determining the guidance type corresponding to the user interaction data based on the user interaction data includes: when the preference content corresponding to the user behavior characteristics is a first preference content, determining a seventh guidance type corresponding to the user interaction data based on the user behavior characteristics; when the preference content corresponding to the user behavior characteristics is a second preference content, determining an eighth guidance type corresponding to the user interaction data based on the user behavior characteristics; when the first preference content and the second preference content are different, the seventh guidance type and the eighth guidance type are different.

[0201] In an optional embodiment, the user desires more content of interest. The type of guidance can be determined based on the user's preferred content, deciding whether to add a second response. For example, when the preferred content is similar to the user's question, the guidance type is "guided"; when the preferred content is dissimilar, the guidance type is "no guidance." Different guidance types apply to different preferred content. For instance, a mapping relationship between the similarity between preferred content and the user's question and the guidance type can be preset, the similarity between the preferred content and the user's question can be calculated, and the corresponding guidance type can be determined based on the similarity.

[0202] By limiting user interaction data to user behavior characteristics, and determining the type of guidance based on the preferred content identified by these characteristics, more response content can be provided for questions that users are interested in.

[0203] Optionally, the user interaction data includes: vehicle data; determining the guidance type corresponding to the user interaction data based on the user interaction data includes: when the operation complexity corresponding to the vehicle data is a first operation complexity, determining a ninth guidance type corresponding to the user interaction data based on the vehicle data; when the operation complexity corresponding to the vehicle data is a second operation complexity, determining a tenth guidance type corresponding to the user interaction data based on the vehicle data; when the first operation complexity and the second operation complexity are different, the ninth guidance type and the tenth guidance type are different.

[0204] Generally, the more vehicle data a user's question requires, the more complex the process of obtaining the first response becomes, and the more detailed the response content provided. This necessitates providing complete content through multiple rounds of dialogue, i.e., adding a second response, resulting in a guided response type. Conversely, the less vehicle data a user's question requires, the simpler the process of obtaining the first response becomes, eliminating the need for multiple rounds of dialogue to provide more content. This necessitates not adding a second response, resulting in an unguided response type. Generally, higher operational complexity corresponds to a guided response type, while lower operational complexity corresponds to an unguided response type.

[0205] Different guidance types are used for different vehicle data. For example, a mapping relationship between operation complexity and guidance type can be preset, and the corresponding guidance type can be determined based on the operation complexity.

[0206] By specifically limiting user interaction data to vehicle data, we can consider whether to add a second response based on the complexity of vehicle data processing in human-vehicle interaction scenarios. This allows for guidance types that adapt to the complexity of vehicle data processing, helping users to gain a more comprehensive understanding of the vehicle's status.

[0207] When user interaction data includes at least two items, determining the guidance type corresponding to the user interaction data can be done by: determining the guidance type for each item; weighting the guidance types determined for each item to obtain a weighted result; and using the weighted result as the guidance type corresponding to the user interaction data.

[0208] In one example, the user interaction data of the first user includes at least one of the following: first user attribute information, user questions of a first level of difficulty, user history dialogues of a first level of known information, user behavior characteristics of a first level of preferred content, and vehicle data of a first level of operational complexity. The guidance type corresponding to the user interaction data can be obtained by weighting at least one of the first guidance type, third guidance type, fifth guidance type, seventh guidance type, and ninth guidance type.

[0209] Furthermore, for the same user's question, different combinations of user interaction data result in different second responses.

[0210] In an optional embodiment, generating response information based on the user interaction data includes: determining the target information volume and guidance type based on the user interaction data, and determining the response type; querying the response indication information corresponding to the response type from a set of preset response indication information; and generating response information corresponding to the user interaction data based on the response indication information and the user interaction data.

[0211] In an optional embodiment, the response type is used to query response instruction information. The response instruction information may include types such as low information content with no follow-up questions, low information content with follow-up questions, high information content with no follow-up questions, and high information content with follow-up questions. The amount of information can be determined based on the target information content, and the presence or absence of guidance can be determined based on the guidance type, thereby combining the target information content and the guidance type to obtain the response type.

[0212] Multiple response indication messages can be pre-configured, and a correspondence between response indication messages and response types can be established. Based on the response type, the corresponding response indication message can be queried from the correspondence.

[0213] For example, the response instruction for a response type with little information and no guiding follow-up questions could be: Please provide a simple answer to the question: XX.

[0214] For example, the response instructions for a response type with limited information and guiding follow-up questions could be: Please provide an answer that the child can understand for a child's question: XX, and add guiding content that matches the child's interests.

[0215] For example, the response instruction for a response type that contains a lot of information but does not provide any follow-up questions could be: Please provide a complete answer to an adult's question: XX.

[0216] For example, the reply instruction information for a reply type that contains a lot of information and guides follow-up questions could be: Please provide a detailed answer to the question: XX, and add more guiding content for the answer.

[0217] By configuring response types and corresponding response instructions for each type, different response needs can be preset. Based on these needs, corresponding response instructions can be provided, and the most suitable response content can be generated. This adapts to response requirements, provides accurate response content, and increases the flexibility of response content to meet diverse interaction needs and adapt to various interaction scenarios. Furthermore, by presetting response instructions and establishing a correspondence between response types and response instructions, the response instructions can be quickly determined, thereby improving interaction efficiency.

[0218] Optionally, the response types include: little information and no follow-up questions, little information and some follow-up questions, a lot of information and no follow-up questions, and a lot of information and some follow-up questions.

[0219] In one optional embodiment, the response information corresponding to the type with little information and no guided follow-up questions only includes the first response content and does not include the second response content, and the first response content has relatively little information; the response information corresponding to the type with little information and guided follow-up questions includes both the first response content and the second response content, and the first response content has relatively little information; the response information corresponding to the type with a lot of information and no guided follow-up questions only includes the first response content and does not include the second response content, and the first response content has relatively much information; the response information corresponding to the type with a lot of information and guided follow-up questions includes both the first response content and the second response content, and the first response content has relatively much information.

[0220] As in the previous example, responses with limited information and no follow-up questions include: The main reason for the extinction of dinosaurs was an asteroid impact that caused drastic environmental changes (first response content).

[0221] For example, responses with limited information but prompting follow-up questions might include: The main reason for the extinction of dinosaurs was an asteroid impact that caused drastic environmental changes (first response). Would you like to know the story of this massive meteorite impact (second response)?

[0222] For example, responses to questions that are information-rich but lack follow-up questions might include: "Wow, that's a really good question! Dinosaurs disappeared because a huge event happened on Earth a long, long time ago. Imagine a giant rock flying down from the sky and crashing into the Earth with a bang—that's a meteorite. This giant rock created a lot of dust and smoke that blocked out the sky, preventing sunlight from reaching the ground. The Earth became very cold, and many plants died. Dinosaurs had no food to eat, so they slowly disappeared (first response content)."

[0223] For example, responses to questions that are informative and encourage follow-up questions might include: "This is an excellent question. The extinction of dinosaurs was caused by an asteroid impact, which drastically altered the environment, making survival impossible. This process is also complex and interesting (first response). Would you like to hear about the impact process?" (second response)

[0224] By configuring multiple response types and corresponding response instructions for each type, the diversity of response types is increased, thereby enriching the response instructions. This allows for adaptation to various interaction scenarios and flexible adjustment of the information content in the response.

[0225] Optionally, generating reply instruction information based on the user interaction data includes: inputting the user interaction data into a classification model to obtain the reply type output by the classification model; and querying the reply instruction information corresponding to the reply type from a preset reply prompt template based on the reply type.

[0226] In an optional embodiment, a classification model is used to classify user interaction data to obtain response types. For example, the classification model can be a large language model. The specific processing steps of the classification model are: encoding the user interaction data to obtain a corresponding vector representation, and decoding the vector to obtain the response type.

[0227] By inputting user interaction data into a classification model to obtain response types, complex user interaction data can be processed, and response information can be generated quickly.

[0228] Optionally, the reply instruction information includes a reply prompt template; generating reply information corresponding to the user interaction data based on the reply instruction information and the user interaction data includes: filling the user interaction data into the reply prompt template to obtain target input content; and inputting the target input content into the generation model to obtain reply information corresponding to the user interaction data.

[0229] In an optional embodiment, the response prompt template includes slots for user interaction data. Filling the response prompt template with user interaction data involves adding the data to the corresponding slots. The filled response prompt template serves as the target input content. This target input content is used as input data for the generative model. The generative model processes the target input content to generate response information. For example, the specific processing steps of the generative model are: encoding the target input content to obtain a corresponding vector representation, and decoding the vector representation to obtain the response information. The generative model can be a large language model, used to input natural language content and generate response information.

[0230] By limiting the response instruction information to a response prompt template and inputting user interaction data into the response prompt template to obtain the target input content, and then inputting the target input content into the generative model to obtain the response information, complex question content can be processed, accurate and coherent response information can be generated, the realism of the interaction can be improved, and response information can be generated quickly, thus improving the efficiency of the interaction.

[0231] It should be noted that the aforementioned classification model and generative model can be the same model or two different, independent models. Optionally, user interaction data is input into the classification model to obtain the response type output by the classification model; the response indication information corresponding to the response type is queried, and the user interaction data is filled into the response indication information to obtain the target input content; the target input content is input into the generative model to obtain the response information output by the generative model. The size of the classification model is smaller than that of the generative model. Specifically, the number of levels in the classification model is less than the number of levels in the generative model, and / or the number of parameters in the classification model is less than the number of parameters in the generative model. This two-level connected model structure processes user interaction data. Simple operations, i.e., generating response types, can be performed by the small-scale, fast classification model at the front end, while complex operations, i.e., generating response information, can be performed by the large-scale, powerful inference and production model at the back end, which has better instruction compliance capabilities. This allows for reasonable task allocation, improving the speed of response information generation while ensuring accuracy, thus achieving a balance between cost and effectiveness.

[0232] Optionally, the generative model is obtained by: acquiring standard question content and corresponding standard response information; splitting the standard response information to obtain multiple sub-response information and the response order of each sub-response information; the information content of the standard response information is greater than the information content of each sub-response information; adding corresponding guiding follow-up questions to the sub-response information that responds first to obtain target response information; combining at least one target response information and the last sub-response information in the response order with the standard question content to generate multiple training samples; and fine-tuning the pre-trained initial model using each training sample to obtain the generative model.

[0233] In an optional embodiment, the initial model can be a pre-trained general-domain adapted large language model. The generating model can be a fine-tuned general-domain large language model to obtain a model adapted to the response content generation method provided in this embodiment of the invention. Standard question content and standard response information are a pair of question-response data, and standard response information is the correct response information to the standard question content. It should be noted that there can be more than one correct response information to the standard question content. One correct response information can be selected as the standard response information, or multiple correct response information can be selected and fused to obtain the standard response information. The methods for obtaining standard question content and standard response information can include at least one of the following: manual generation, obtaining high-scoring question-answer pairs and topic content and corresponding comments from question-and-answer websites, etc.

[0234] In an optional embodiment, the standard response information is split into sub-response information with a general-to-specific logical relationship, or into sub-response information with a progressive or inferential relationship. The information content of the standard response information is greater than that of the sub-response information.

[0235] A standard, example response: The dinosaurs disappeared because a huge event happened on Earth a long, long time ago. Imagine a giant rock flying down from the sky and crashing into the Earth with a bang—that's a meteorite. This rock created a lot of dust and smoke that blocked out the sky, preventing sunlight from reaching the ground. The Earth became very cold, and many plants died. The dinosaurs had no food to eat, so they slowly disappeared.

[0236] In an optional embodiment, the total sub-response information in the split sub-response information is: The extinction of dinosaurs was due to an asteroid impact on Earth, causing drastic environmental changes that made survival impossible. A molecular response is: A huge rock flew from the sky and crashed into Earth with a bang—this is a large meteorite. This large rock caused a lot of dust and smoke to obscure the sky, making the Earth very cold, leading to the extinction of the dinosaurs. A molecular response is: The Earth became very cold, and many plants died. The dinosaurs had no food to eat, so they gradually disappeared.

[0237] For example, the standard response information is: S1 / T1=V1, S2 / T2=V2, V1>V2.

[0238] One of the sub-response messages in the split sub-response messages is: S1 / T1 = V1; another of the sub-response messages in the split sub-response messages is: S2 / T2 = V2; and yet another of the sub-response messages in the split sub-response messages is: V1 > V2.

[0239] In the example above, each sub-response message only includes a portion of the content in the standard response message. Therefore, the information content of the standard response message is greater than the information content of each sub-response message.

[0240] In addition, there are other splitting methods and relationships between sub-response information, which can be set as needed, without specific limitations.

[0241] The response order indicates the output sequence of the sub-response messages derived from the breakdown of the same standard response message. The response order of each sub-response message can be determined based on the logical relationships between them.

[0242] For example, when dealing with sub-response information in a total-to-sub logical relationship, the total sub-response information is usually responded to first, followed by the sub-response information. Furthermore, the order of sub-response information can be determined based on importance, causal relationship, or other logical relationships. If the sub-response information is in a parallel relationship, the order of response can be randomly determined.

[0243] For example, the order of responses to sub-response messages with progressive or inferential relationships can be determined according to the relationship. Specifically, in inferential relationships, sub-response messages without dependencies are responded to first, while those that depend on other sub-response messages are responded to after those dependent sub-response messages.

[0244] In an optional embodiment, each sub-response message is divided into two categories: the last sub-response message in the response order and the sub-response messages that are all present except the last one, i.e., the sub-response messages that are present in the response order first. There is at least one sub-response message that is present in the response order first, and this includes all sub-response messages except the last one. Follow-up questions can be added to the first-response messages to obtain the target response. Specifically, after adding follow-up questions to the first-response messages, the user can be guided to further inquire about the subsequent sub-response messages. The follow-up questions can be determined based on standard question content, uniform wording, or a summary of the keywords or the first sentence of the next adjacent sub-response message. These can be set as needed and are not specifically limited. Since there are no sub-response messages that are present in the last-response message, it is not necessary to add follow-up questions to them.

[0245] In an optional embodiment, for the same standard question content, each target response is combined with the standard question content to obtain corresponding training samples. The last sub-response information in the response order is combined with the standard question content to obtain corresponding training samples. The number of sub-response information obtained by splitting the standard response information is the same as the number of generated training samples.

[0246] In addition, a large number of standard question contents and corresponding standard question content pairs can be set, and multiple training samples can be generated for each pair. The initial model is then fine-tuned using a large number of training samples. For example, the initial model can be fine-tuned using the SFT (Self-optimizing and Self-adjusting) method.

[0247] Besides splitting standard response information into sub-response information, standard response information can also be combined with standard question content to generate training samples. Furthermore, standard response information can be simplified using at least one method to obtain at least one sub-response information, each of which is then combined with standard question content to generate corresponding training samples. Additionally, at least one related piece of information from the standard response information can be obtained and expanded to obtain at least one response with more information. Each expanded response is then combined with standard question content to generate corresponding training samples. These generated training samples can be input into the initial model for fine-tuning. This significantly expands the coverage of the training samples, improves their representativeness, and using these training samples to fine-tune the initial model improves the generalization ability of the fine-tuned generative model, increases the accuracy of the generated model's responses, and enhances the diversity of the responses.

[0248] By fine-tuning the initial model, the generated model can be made capable of differentiating user interaction data, adjusting the amount of information in the response, and guiding content production. This allows for gradual guidance of user interaction, step by step, and improved dialogue authenticity.

[0249] In a specific application scenario, such as Figure 4 As shown, during vehicle interaction, the vehicle acquires the user's questions and multimodal data, and sends them to the server. The server generates the response information using the following method:

[0250] S401. Obtain user interaction data.

[0251] In an optional embodiment, user interaction data includes at least the content of user queries and user attribute information. User interaction data may also include, but is not limited to, user history conversations, user behavior characteristics, and vehicle data. User attribute information includes at least user identifier and age. User interaction data can be multimodal data and may include data from at least one media type.

[0252] S402. The server outputs the determination result of adult or child through the classification model.

[0253] In one optional embodiment, the target information content for adults is high, and the target information content for children is low.

[0254] S403. When determining an adult dialogue link, the server uses a classification model to output a judgment result on whether there are dialogue guidance words.

[0255] In one optional embodiment, there is no dialogue guidance script, corresponding to the guidance type being "no guidance". There is a dialogue guidance script, corresponding to the guidance type being "guided".

[0256] S404. When determining the dialogue without guidance, obtain the prompt template corresponding to adults without guidance.

[0257] S405. When it is determined that there is a dialogue guidance script, obtain the prompt template for adults with corresponding guidance.

[0258] S406. When determining the dialogue link for children, the server uses a classification model to output a judgment result on whether there are dialogue guidance words.

[0259] S407. When determining the dialogue prompt without guidance, obtain the prompt template corresponding to children and no guidance.

[0260] S408. When it is determined that there is a dialogue guidance script, obtain the prompt template for the child and the corresponding guidance.

[0261] In an optional embodiment, the target information content is determined based on whether the child is a child or an adult, the guidance type is determined based on whether there is a dialogue guidance script, the response type is determined based on the target information content and the guidance type, and the response instruction information corresponding to the response type is queried. The response instruction information may be a prompt template.

[0262] In an optional embodiment, a classification model determines whether the user is a child or an adult (based on the amount of information provided) and whether guidance is required, i.e., the response type, based on the user's age, dialogue context (i.e., historical dialogue), and the content of the user's question. A prompt template is then selected and input into the generation model. Here, "child" corresponds to less information, and "adult" corresponds to more information.

[0263] In one optional embodiment, the classification model first determines whether the dialogue link is for adults or children based on age. Adult dialogue links correspond to response types with more information, while child dialogue links correspond to response types with less information. Secondly, the classification model determines whether the response type is guided or not based on the user's question content and user information. Alternatively, the classification model can determine whether the response type is guided or not based on the user's question content, user information, and vehicle information.

[0264] S409. Integrate user questions, multimodal data, and prompt templates.

[0265] S410. Input the fusion result into the model, and the model outputs the response content and dialogue guidance script.

[0266] In an optional embodiment, the dialogue guidance script can be empty. The fusion result is input into the generative model to obtain the response information output by the generative model. The response information includes at least the response content (first response content) and optionally includes guiding follow-up questions (second response content).

[0267] The classification model is a small-scale pre-model, while the generative model is a large-scale post-model. The classification model is also a large language model, capable of performing simple tasks. Simple tasks, such as vehicle control tasks, can be directly input into the classification model. The classification model generates vehicle control commands, which are then sent to the vehicle's infotainment system. The system controls certain vehicle entities without needing further processing before being input into the generative model. For example, if the entity is a window, the control command would be the instruction to open or close the window. Furthermore, the classification model can also, based on the user's question, call a search interface to obtain relevant external knowledge. This external knowledge is then sent to the large language model, specifically used to populate the response instructions. This fusion of user question content and external knowledge provides the input data for the large language model, enabling it to better understand the user's question and improve response accuracy. This approach reduces the need for a high-performance large language model to perform all tasks, which would otherwise be costly. By using a two-tiered model structure with low and high configurations, tasks can be rationally allocated, balancing cost reduction with response accuracy.

[0268] This invention integrates multimodal data with user information to address the lack of multimodal information in previous model architectures. This allows for the perception of more and richer information, thereby improving the accuracy of the output responses. By appropriately breaking down the dialogue process and gradually guiding the user, simulating conversations between users, it can guide user questions, improve the authenticity of dialogue responses, and enhance user experience. Furthermore, it can shorten response content, making it easier for users to understand, especially improving the human-computer interaction experience for children. Finally, through reasonable multi-level model allocation, it achieves a balance between cost and effectiveness, thereby reducing costs and response latency.

[0269] Figure 5 This is a schematic diagram of a response information generation device provided in an embodiment of the present invention. This embodiment of the present invention is applicable to situations where a user's questions are answered during human-computer interaction between a vehicle and a user. The device can execute a response information generation method and can be implemented in hardware and / or software. The device can be configured in an electronic device.

[0270] See Figure 5 The response information generation device shown includes:

[0271] The interaction data acquisition module 501 is used to acquire user interaction data;

[0272] The reply information generation module 502 is used to generate reply information based on the user interaction data. The reply information includes at least a first reply content, and the amount of information in the first reply content corresponds to the user interaction data.

[0273] The embodiments of the present invention generate response information with an amount of information adapted to user interaction data. The amount of information in the response information can be flexibly adjusted, specifically increasing or decreasing the amount of information in the response information, thereby providing feedback that is easy for users to understand.

[0274] Optionally, the response information generation module 502 includes:

[0275] The target information content determination unit is used to determine the target information content corresponding to the user interaction data based on the user interaction data.

[0276] The first response content generation unit is used to generate response information based on the user interaction data and the target information amount, wherein the information amount of the first response content corresponds to the target information amount.

[0277] Optionally, the user interaction data includes: user attribute information;

[0278] The target information content determination unit is specifically used for:

[0279] When the user attribute information is the first user attribute information, the first information quantity corresponding to the first user attribute information is determined based on the first user attribute information.

[0280] When the user attribute information is the second user attribute information, the second information quantity corresponding to the second user attribute information is determined according to the second user attribute information; the first user attribute information is different from the second user attribute information, and the first information quantity is different from the second information quantity.

[0281] Optionally, the first user attribute information includes a first age; the second user attribute information includes a second age; the first age and the second age are different, and the first information content corresponding to the first age and the second information content corresponding to the second age are different.

[0282] Optionally, the user interaction data includes: user-asked questions;

[0283] The target information content determination unit is specifically used for:

[0284] When the difficulty level of the user's question is the first difficulty level, the third information content corresponding to the user interaction data is determined based on the user's question.

[0285] When the difficulty level corresponding to the user's question is the second difficulty level, a fourth information quantity corresponding to the user interaction data is determined based on the user's question; when the first difficulty level is different from the second difficulty level, the third information quantity is different from the fourth information quantity.

[0286] Optionally, the user interaction data includes: user's historical conversations;

[0287] The target information content determination unit is specifically used for:

[0288] When the known information corresponding to the user's historical dialogue is the first known information, the fifth information quantity corresponding to the user interaction data is determined based on the user's historical dialogue.

[0289] When the known information corresponding to the user's historical dialogue is the second known information, the sixth information quantity corresponding to the user interaction data is determined based on the user's historical dialogue; when the first known information is different from the second known information, the fifth information quantity is different from the sixth information quantity.

[0290] Optionally, the user interaction data includes: user behavior characteristics;

[0291] The target information content determination unit is specifically used for:

[0292] When the preference content corresponding to the user behavior feature is the first preference content, the seventh information content corresponding to the user interaction data is determined according to the user behavior feature;

[0293] When the preference content corresponding to the user behavior feature is the second preference content, the eighth information content corresponding to the user interaction data is determined according to the user behavior feature; when the first preference content is different from the second preference content, the seventh information content is different from the eighth information content.

[0294] Optionally, the user interaction data includes: vehicle data;

[0295] The target information content determination unit is specifically used for:

[0296] When the operation complexity corresponding to the vehicle data is the first operation complexity, the ninth information quantity corresponding to the user interaction data is determined based on the vehicle data.

[0297] When the operation complexity corresponding to the vehicle data is the second operation complexity, the tenth information quantity corresponding to the user interaction data is determined based on the vehicle data; when the first operation complexity is different from the second operation complexity, the ninth information quantity is different from the tenth information quantity.

[0298] Optionally, the response information may also include a second response, which is a follow-up question corresponding to the user interaction data.

[0299] Optionally, the response information generation module 502 includes:

[0300] The guidance type determination unit is used to determine the guidance type corresponding to the user interaction data based on the user interaction data.

[0301] The second response content judgment unit is used to generate response information based on the user interaction data and the guidance type, and to determine whether the response information includes the second response content corresponding to the guidance type.

[0302] Optionally, the user interaction data includes: user attribute information;

[0303] The guidance type determination unit is specifically used for:

[0304] When the user attribute information is the first user attribute information, the first guidance type corresponding to the first user attribute information is determined according to the first user attribute information;

[0305] When the user attribute information is the second user attribute information, the second boot type corresponding to the second user attribute information is determined according to the second user attribute information; the first user attribute information is different from the second user attribute information, and the first boot type is different from the second boot type.

[0306] Optionally, the user interaction data includes: user-asked questions;

[0307] The guidance type determination unit is specifically used for:

[0308] When the difficulty level of the user's question is the first difficulty level, the third guidance type corresponding to the user interaction data is determined based on the user's question.

[0309] When the difficulty level corresponding to the user's question is the second difficulty level, a fourth guidance type corresponding to the user interaction data is determined based on the user's question; when the first difficulty level is different from the second difficulty level, the third guidance type is different from the fourth guidance type.

[0310] Optionally, the user interaction data includes: user's historical conversations;

[0311] The guidance type determination unit is specifically used for:

[0312] When the known information corresponding to the user's historical dialogue is the first known information, the fifth guidance type corresponding to the user interaction data is determined based on the user's historical dialogue.

[0313] When the known information corresponding to the user's historical dialogue is the second known information, the sixth guidance type corresponding to the user interaction data is determined based on the user's historical dialogue; when the first known information is different from the second known information, the fifth guidance type is different from the sixth guidance type.

[0314] Optionally, the user interaction data includes: user behavior characteristics;

[0315] The guidance type determination unit is specifically used for:

[0316] When the preference content corresponding to the user behavior feature is the first preference content, the seventh guidance type corresponding to the user interaction data is determined according to the user behavior feature;

[0317] When the preference content corresponding to the user behavior feature is the second preference content, the eighth guidance type corresponding to the user interaction data is determined according to the user behavior feature; when the first preference content is different from the second preference content, the seventh guidance type is different from the eighth guidance type.

[0318] Optionally, the user interaction data includes: vehicle data;

[0319] The guidance type determination unit is specifically used for:

[0320] When the operation complexity corresponding to the vehicle data is the first operation complexity, the ninth guidance type corresponding to the user interaction data is determined based on the vehicle data.

[0321] When the operation complexity corresponding to the vehicle data is the second operation complexity, the tenth guidance type corresponding to the user interaction data is determined based on the vehicle data; when the first operation complexity is different from the second operation complexity, the ninth guidance type is different from the tenth guidance type.

[0322] The response information generation device provided in this embodiment of the invention can execute the response information generation method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of executing the response information generation method.

[0323] Figure 6A schematic diagram of an electronic device 600 that can be used to implement embodiments of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0324] like Figure 6 As shown, the electronic device 600 includes at least one processor 601 and a memory, such as a read-only memory (ROM) 602 or a random access memory (RAM) 603, communicatively connected to the at least one processor 601. The memory stores computer programs executable by the at least one processor. The processor 601 can perform various appropriate actions and processes based on the computer program stored in the ROM 602 or loaded into the RAM 603 from storage unit 608. The RAM 603 may also store various programs and data required for the operation of the electronic device 600. The processor 601, ROM 602, and RAM 603 are interconnected via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0325] Multiple components in electronic device 600 are connected to I / O interface 605, including: input unit 606, such as keyboard, mouse, etc.; output unit 607, such as various types of displays, speakers, etc.; storage unit 608, such as disk, optical disk, etc.; and communication unit 609, such as network card, modem, wireless transceiver, etc. Communication unit 609 allows electronic device 600 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0326] Processor 601 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 601 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 601 performs the various methods and processes described above, such as response information generation methods.

[0327] In some embodiments, the response information generation method may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 608. In some embodiments, part or all of the computer program may be loaded and / or installed on electronic device 600 via ROM 602 and / or communication unit 609. When the computer program is loaded into RAM 603 and executed by processor 601, one or more steps of the response information generation method described above may be performed. Alternatively, in other embodiments, processor 601 may be configured to perform the response information generation method by any other suitable means (e.g., by means of firmware).

[0328] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0329] Computer programs used to implement the methods of the present invention can be written in any combination of one or more programming languages. These computer programs can be provided to the processor of a general-purpose computer, a special-purpose computer, or other programmable response information generation device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The computer programs can be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0330] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0331] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0332] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or computing systems that include middleware components (e.g., application servers), or computing systems that include frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0333] A computing system can include clients and servers. Clients and servers are generally geographically separated and typically interact via communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system. It addresses the shortcomings of traditional physical hosts and VPS (Virtual Private Server) services, such as high management difficulty and weak business scalability.

[0334] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0335] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A method for generating response information, characterized in that, The method includes: Obtain user interaction data; Based on the user interaction data, a reply message is generated, which includes at least a first reply message, the amount of information in the first reply message corresponding to the user interaction data.

2. The method according to claim 1, characterized in that, The step of generating response information based on the user interaction data includes: Based on the user interaction data, determine the target information content corresponding to the user interaction data; Based on the user interaction data and the target information amount, a reply is generated, wherein the information amount of the first reply content corresponds to the target information amount.

3. The method according to claim 2, characterized in that, The user interaction data includes: user attribute information; The step of determining the target information content corresponding to the user interaction data based on the user interaction data includes: When the user attribute information is the first user attribute information, the first information quantity corresponding to the first user attribute information is determined based on the first user attribute information. When the user attribute information is the second user attribute information, the second information quantity corresponding to the second user attribute information is determined according to the second user attribute information; the first user attribute information is different from the second user attribute information, and the first information quantity is different from the second information quantity.

4. The method according to claim 3, characterized in that, The first user attribute information includes a first age; the second user attribute information includes a second age; the first age and the second age are different, and the first information content corresponding to the first age and the second information content corresponding to the second age are different.

5. The method according to claim 2, characterized in that, The user interaction data includes: user-asked questions; The step of determining the target information content corresponding to the user interaction data based on the user interaction data includes: When the difficulty level of the user's question is the first difficulty level, the third information content corresponding to the user interaction data is determined based on the user's question. When the difficulty level corresponding to the user's question is the second difficulty level, a fourth information quantity corresponding to the user interaction data is determined based on the user's question; when the first difficulty level is different from the second difficulty level, the third information quantity is different from the fourth information quantity.

6. The method according to claim 2, characterized in that, The user interaction data includes: user's historical conversations; The step of determining the target information content corresponding to the user interaction data based on the user interaction data includes: When the known information corresponding to the user's historical dialogue is the first known information, the fifth information quantity corresponding to the user interaction data is determined based on the user's historical dialogue. When the known information corresponding to the user's historical dialogue is the second known information, the sixth information quantity corresponding to the user interaction data is determined based on the user's historical dialogue; when the first known information is different from the second known information, the fifth information quantity is different from the sixth information quantity.

7. The method according to claim 2, characterized in that, The user interaction data includes: user behavior characteristics; The step of determining the target information content corresponding to the user interaction data based on the user interaction data includes: When the preference content corresponding to the user behavior feature is the first preference content, the seventh information content corresponding to the user interaction data is determined according to the user behavior feature; When the preference content corresponding to the user behavior feature is the second preference content, the eighth information content corresponding to the user interaction data is determined according to the user behavior feature; when the first preference content is different from the second preference content, the seventh information content is different from the eighth information content.

8. The method according to claim 2, characterized in that, The user interaction data includes: vehicle data; The step of determining the target information content corresponding to the user interaction data based on the user interaction data includes: When the operation complexity corresponding to the vehicle data is the first operation complexity, the ninth information quantity corresponding to the user interaction data is determined based on the vehicle data. When the operation complexity corresponding to the vehicle data is the second operation complexity, the tenth information quantity corresponding to the user interaction data is determined based on the vehicle data; when the first operation complexity is different from the second operation complexity, the ninth information quantity is different from the tenth information quantity.

9. The method according to claim 1, characterized in that, The response information also includes a second response, which is a follow-up question corresponding to the user interaction data.

10. The method according to claim 9, characterized in that, The step of generating response information based on the user interaction data includes: Based on the user interaction data, determine the guidance type corresponding to the user interaction data; Based on the user interaction data and the guidance type, a response message is generated, and the response message may include the second response content corresponding to the guidance type.

11. The method according to claim 10, characterized in that, The user interaction data includes: user attribute information; The step of determining the guidance type corresponding to the user interaction data based on the user interaction data includes: When the user attribute information is the first user attribute information, the first guidance type corresponding to the first user attribute information is determined according to the first user attribute information; When the user attribute information is the second user attribute information, the second boot type corresponding to the second user attribute information is determined according to the second user attribute information; the first user attribute information is different from the second user attribute information, and the first boot type is different from the second boot type.

12. The method according to claim 10, characterized in that, The user interaction data includes: user-asked questions; The step of determining the guidance type corresponding to the user interaction data based on the user interaction data includes: When the difficulty level of the user's question is the first difficulty level, the third guidance type corresponding to the user interaction data is determined based on the user's question. When the difficulty level corresponding to the user's question is the second difficulty level, a fourth guidance type corresponding to the user interaction data is determined based on the user's question; when the first difficulty level is different from the second difficulty level, the third guidance type is different from the fourth guidance type.

13. The method according to claim 10, characterized in that, The user interaction data includes: user's historical conversations; The step of determining the guidance type corresponding to the user interaction data based on the user interaction data includes: When the known information corresponding to the user's historical dialogue is the first known information, the fifth guidance type corresponding to the user interaction data is determined based on the user's historical dialogue. When the known information corresponding to the user's historical dialogue is the second known information, the sixth guidance type corresponding to the user interaction data is determined based on the user's historical dialogue; when the first known information is different from the second known information, the fifth guidance type is different from the sixth guidance type.

14. The method according to claim 10, characterized in that, The user interaction data includes: user behavior characteristics; The step of determining the guidance type corresponding to the user interaction data based on the user interaction data includes: When the preference content corresponding to the user behavior feature is the first preference content, the seventh guidance type corresponding to the user interaction data is determined according to the user behavior feature; When the preference content corresponding to the user behavior feature is the second preference content, the eighth guidance type corresponding to the user interaction data is determined according to the user behavior feature; when the first preference content is different from the second preference content, the seventh guidance type is different from the eighth guidance type.

15. The method according to claim 10, characterized in that, The user interaction data includes: vehicle data; The step of determining the guidance type corresponding to the user interaction data based on the user interaction data includes: When the operation complexity corresponding to the vehicle data is the first operation complexity, the ninth guidance type corresponding to the user interaction data is determined based on the vehicle data. When the operation complexity corresponding to the vehicle data is the second operation complexity, the tenth guidance type corresponding to the user interaction data is determined based on the vehicle data; when the first operation complexity is different from the second operation complexity, the ninth guidance type is different from the tenth guidance type.

16. A response information generation device, characterized in that, include: The interaction data acquisition module is used to acquire user interaction data; The reply information generation module is used to generate reply information based on the user interaction data. The reply information includes at least a first reply content, and the amount of information in the first reply content corresponds to the user interaction data.

17. An electronic device, characterized in that, The electronic device includes: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the response information generation method according to any one of claims 1-15.

18. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that, when executed by a processor, implement the response information generation method according to any one of claims 1-15.

Images