Information processing method and device, equipment, storage medium and product
By dividing and summarizing the time period of historical dialogues and building target prompt information, the problem of dialogue response quality and interaction efficiency caused by the huge amount of historical dialogue content in intelligent dialogue services is solved, and more efficient and accurate reply is achieved.
Patent Information
- Application Number
- CN202510038869.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-01-09
- Publication Date
- 2025-04-18
AI Technical Summary
As users use smart dialogue services increase their duration, the amount of historical dialogue content is huge, making it difficult to take into account the quality of dialogue responses and interaction efficiency.
By dividing historical dialogue into multiple time periods according to time periods, summarizing and processing them separately, and building target prompt information is built so that a preset dialogue model based on the large language model can generate more accurate replies.
Improves the efficiency of replying and the accuracy of replying information, and improves the user experience.
Smart Images
Figure CN119849638A_ABST
Abstract
Description
Technical Field
[0001] The embodiments of the present disclosure relate to the field of computer technology, and in particular, to information processing methods, devices, equipment, storage media, and products. Background Art
[0002] At present, Large Language Model (LLM) is widely used in intelligent dialogue services. It can understand and generate natural language, and then realize intelligent dialogue and interaction with users.
[0003] In order to communicate with users more accurately, the large language model needs to combine the historical conversation content to answer the user's input in this round of dialogue. However, as users continue to use intelligent dialogue services, the historical conversation content usually has a huge amount of data, making it difficult to balance the quality of dialogue responses and interaction efficiency. Summary of the invention
[0004] The embodiments of the present disclosure provide information processing methods, devices, equipment, storage media and products, which can optimize existing information processing solutions based on intelligent dialogue responses.
[0005] In a first aspect, an embodiment of the present disclosure provides an information processing method, including:
[0006] In response to receiving input information of the current round of dialogue, performing information matching for the input information in a historical information database, and determining first prompt information according to the information matching result, wherein the historical information database includes a plurality of historical dialogue information, and the historical dialogue information is obtained by summarizing historical dialogues within a first historical period;
[0007] Acquire second prompt information, wherein the second prompt information is obtained by summarizing historical conversations in a second historical period, and the second historical period is later than the first historical period;
[0008] Determine target prompt information according to the first prompt information, the second prompt information and the third prompt information, wherein the third prompt information is a historical conversation in a third historical period, the third historical period is later than the second historical period, and the third historical period includes at least a previous round of conversation before the current round of conversation;
[0009] The response information corresponding to the input information is determined based on the target prompt information using a preset dialogue model, wherein the preset dialogue model is constructed based on a large language model.
[0010] In a second aspect, the present disclosure also provides an information processing device, including:
[0011] a first prompt information determination module, configured to, in response to receiving input information of the current round of dialogue, perform information matching for the input information in a historical information database, and determine first prompt information according to the information matching result, wherein the historical information database includes a plurality of historical dialogue information, and the historical dialogue information is obtained by summarizing historical dialogues within a first historical period;
[0012] A second prompt information acquisition module, used to acquire second prompt information, wherein the second prompt information is obtained by summarizing historical conversations in a second historical period, and the second historical period is later than the first historical period;
[0013] a target prompt information determination module, configured to determine target prompt information according to the first prompt information, the second prompt information and the third prompt information, wherein the third prompt information is a historical conversation within a third historical period, the third historical period is later than the second historical period, and the third historical period includes at least a previous round of conversation before the current round of conversation;
[0014] A reply information determination module is used to determine the reply information corresponding to the input information based on the target prompt information using a preset dialogue model, wherein the preset dialogue model is constructed based on a large language model.
[0015] In a third aspect, an embodiment of the present disclosure further provides an electronic device, the electronic device comprising:
[0016] one or more processors;
[0017] a storage device for storing one or more programs,
[0018] When the one or more programs are executed by the one or more processors, the one or more processors implement the information processing method provided by the embodiment of the present disclosure.
[0019] In a fourth aspect, an embodiment of the present disclosure further provides a storage medium containing computer executable instructions, which, when executed by a computer processor, are used to execute the information processing method provided by the embodiment of the present disclosure.
[0020] In a fifth aspect, an embodiment of the present disclosure further provides a computer program product, including a computer program, which implements the information processing method provided by the embodiment of the present disclosure when executed by a processor.
[0021] The information processing scheme provided by the disclosed embodiment, in response to receiving input information of the current round of dialogue, performs information matching for the input information in a historical information library, and determines first prompt information according to the information matching result, wherein the historical information library includes multiple historical dialogue information, and the historical dialogue information is obtained by summarizing historical dialogues in a first historical period, obtains second prompt information, wherein the second prompt information is obtained by summarizing historical dialogues in a second historical period later than the first historical period, determines target prompt information according to the first prompt information, the second prompt information and the third prompt information, wherein the third prompt information is a historical dialogue in a third historical period later than the second historical period, and the third historical period at least includes the previous round of dialogue of the current round of dialogue, and determines reply information corresponding to the input information based on the target prompt information using a preset dialogue model, wherein the preset dialogue model is constructed based on a large language model. By adopting the above technical solution, the historical conversations between the virtual object and the user can be divided into multiple time periods according to the distance from the current round of conversation, and different time periods are processed differently. The processing results of the historical conversations corresponding to multiple time periods are used to construct target prompt information, providing richer and more accurate reference information for the preset conversation model built based on the large language model, improving the reply efficiency and the accuracy of the reply information, and thus improving the user experience. BRIEF DESCRIPTION OF THE DRAWINGS
[0022] The above and other features, advantages and aspects of the embodiments of the present disclosure will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. Throughout the accompanying drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic and the originals and elements are not necessarily drawn to scale.
[0023] Figure 1 A flowchart of an information processing method provided by an embodiment of the present disclosure;
[0024] Figure 2 A schematic diagram of historical period division provided by an embodiment of the present disclosure;
[0025] Figure 3 A flowchart of another information processing method provided by an embodiment of the present disclosure;
[0026] Figure 4 A schematic diagram of a historical conversation processing process provided by an embodiment of the present disclosure;
[0027] Figure 5 A schematic diagram of a reply information determination process provided by an embodiment of the present disclosure;
[0028] Figure 6 A schematic diagram of the structure of an information processing device provided by an embodiment of the present disclosure;
[0029] Figure 7 A schematic diagram of the structure of an electronic device provided by an embodiment of the present disclosure. DETAILED DESCRIPTION
[0030] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as being limited to the embodiments described herein, which are instead provided for a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not intended to limit the scope of protection of the present disclosure.
[0031] It should be understood that the various steps described in the method embodiments of the present disclosure may be performed in different orders and / or in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present disclosure is not limited in this respect.
[0032] The term "including" and its variations used herein are open inclusions, i.e., "including but not limited to". The term "based on" means "based at least in part on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". The relevant definitions of other terms will be given in the following description.
[0033] It should be noted that the concepts such as "first" and "second" mentioned in the present disclosure are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.
[0034] It should be noted that the modifications of "one" and "plurality" mentioned in the present disclosure are illustrative rather than restrictive, and those skilled in the art should understand that unless otherwise clearly indicated in the context, it should be understood as "one or more".
[0035] It is understandable that before using the technical solutions disclosed in the embodiments of the present disclosure, the types, scope of use, usage scenarios, etc. of the personal information involved in the present disclosure should be informed to the user and the user's authorization should be obtained in an appropriate manner in accordance with relevant laws and regulations.
[0036] For example, in response to receiving an active request from a user, a prompt message is sent to the user to clearly prompt the user that the operation requested to be performed will require obtaining and using the user's personal information. Thus, the user can autonomously choose whether to provide personal information to software or hardware such as an electronic device, application, server, or storage medium that performs the operation of the technical solution of the present disclosure according to the prompt message.
[0037] As an optional but non-limiting implementation, in response to receiving an active request from the user, the prompt information may be sent to the user in the form of a pop-up window, in which the prompt information may be presented in text form. In addition, the pop-up window may also carry a selection control for the user to choose "agree" or "disagree" to provide personal information to the electronic device.
[0038] It is understandable that the above notification and the process of obtaining user authorization are merely illustrative and do not constitute a limitation on the implementation of the present disclosure. Other methods that meet the relevant laws and regulations may also be applied to the implementation of the present disclosure.
[0039] It is understandable that the data involved in this technical solution (including but not limited to the data itself, the acquisition or use of the data) shall comply with the requirements of relevant laws, regulations and relevant provisions.
[0040] Figure 1 A schematic diagram of an information processing flow provided in an embodiment of the present disclosure is applicable to situations of information processing based on intelligent dialogue responses. The method can be executed by an information processing device, which can be implemented in the form of software and / or hardware. Optionally, it can be implemented by an electronic device, which can be a mobile terminal such as a mobile phone, a smart watch, a smart speaker, a tablet computer, and a personal digital assistant, or a personal computer (PC) or a server, or a vehicle.
[0041] like Figure 1 As shown, the method includes:
[0042] Step 101: In response to receiving input information of this round of dialogue, information matching is performed for the input information in a historical information library, and first prompt information is determined according to the information matching result, wherein the historical information library includes a plurality of historical dialogue information, and the historical dialogue information is obtained by summarizing historical dialogues within a first historical period.
[0043] In the disclosed embodiment, a virtual object may be configured in the electronic device. The virtual object may have different names in different application scenarios, such as an intelligent assistant, a voice assistant, a voice robot, or a car assistant, etc. The virtual object may provide services for intelligent dialogue and interaction with the user. During the dialogue between the user and the virtual object, an alternating dialogue may be conducted. A round of dialogue may include the user's input information and the virtual object's reply information. The user's input information may be voice information or text information, etc., and may include various dialogue contents such as chat content and instruction content. The type of reply information may match the type of input information. For example, the user asks a question by voice, and the virtual object may answer by voice.
[0044] In the disclosed embodiment, the virtual object responds to the user through a preset dialogue model, which is constructed based on a large language model. The sample data volume and model parameters of the large language model can be at a massive level, for example, it can have billions or even hundreds of billions of parameters, and it is trained on large-scale text data or other modal data. It can be used for various natural language processing tasks, has good text generation capabilities, and can be applied to language translation and question-answering systems.
[0045] By providing prompt information to the large language model, the large language model can be helped to better understand the context of the user input, thereby generating a more accurate and relevant response. In the disclosed embodiment, the target prompt information is constructed based on the historical conversations in three different historical time periods, which can improve the richness and accuracy of the prompt information. Among them, the three different historical time periods can be continuous, and the length of each historical time period can be preset or dynamically determined according to the actual number of conversation rounds.
[0046] Exemplarily, the first historical period may be the historical period farthest from the current round of dialogue, and the historical dialogues in the first historical period may be called long-term historical dialogues, which may form the long-term memory of the virtual object. Optionally, the first historical period includes a preset number of unit time periods, and the unit time period may be, for example, 1 day, and the preset number may be set according to actual needs, such as 30.
[0047] Exemplarily, the first historical period usually includes multiple rounds of dialogue. The historical dialogue in the first historical period can be segmented, and each segment of the historical dialogue is summarized to obtain multiple historical dialogue information. The multiple historical dialogue information is added to the historical information library, and each historical dialogue information can be understood as a memory block. Optionally, the historical dialogue information in the preset historical information library is determined by the following method: the historical dialogue in the first historical period is segmented into multiple historical dialogue segments based on the preset segmentation method; for each historical dialogue segment in the multiple historical dialogue segments, the current historical dialogue segment is summarized to obtain the historical dialogue information corresponding to the current historical dialogue segment. Among them, the preset segmentation method is not specifically limited. For example, it can be divided according to the above-mentioned unit time period, such as one historical dialogue segment per day; for another example, it can be divided according to the startup and shutdown of the electronic device, and the historical dialogue between each startup and shutdown is divided into the same historical dialogue segment. Taking the vehicle as an example, the historical dialogue between the vehicle power-on and the vehicle power-off is divided into the same historical dialogue segment; for another example, the historical dialogue in the first historical period is understood by using a large language model, and the historical dialogue belonging to the same topic is divided into the same historical dialogue segment. The process of summarizing historical conversations can be completed by an independent task management module. For example, the first summary task is triggered every unit time period (such as 0 o'clock every day). When the first summary task is triggered, the historical conversations in the first historical period corresponding to the current moment are obtained and summarized in the above manner to dynamically update the historical information library. The current historical conversation segment can be summarized by using a large language model to abstractly summarize a large amount of fragmented chat content into a description after thinking about it by the large language model, thereby removing a large amount of parts that have no memory value, retaining important parts, and avoiding the loss of important information.
[0048] After receiving the input information of the current round of dialogue, information matching is performed in the historical information database for the input information, such as performing semantic retrieval in the historical information database, to obtain historical dialogue information associated with the input information, such as obtaining historical dialogue information whose similarity with the input information is greater than a preset similarity threshold. If historical dialogue information matching the input information is retrieved, the first prompt information can be determined based on the retrieved historical dialogue information, such as determining the retrieved historical dialogue information as the first prompt information; if historical dialogue information matching the input information is not retrieved, the first prompt information can be determined to be empty.
[0049] Step 102: Obtain second prompt information, wherein the second prompt information is obtained by summarizing historical conversations in a second historical period, and the second historical period is later than the first historical period.
[0050] Exemplarily, the second historical period may be a historical period that is centered with respect to the first and third historical periods and the current round of dialogue. The historical dialogues within the second historical period may be referred to as mid-term historical dialogues, which may form mid-term memories of virtual objects.
[0051] Optionally, the second historical period and the third historical period mentioned below are included in the same unit time period, such as the time period before the current round of dialogue on the same day. The second historical period is later than the first historical period and earlier than the third historical period. The third historical period includes at least the previous round of dialogue before the current round of dialogue. Among them, the third historical period includes a preset number of historical dialogues (the specific value can be set according to actual needs, such as 3), and the historical time except the third historical period in the unit time period to which the second historical period and the third historical period belong is divided into the second historical period. The third historical period can be the historical period closest to the current round of dialogue. The historical dialogues in the third historical period can be called short-term historical dialogues, which can form the short-term memory of the virtual object.
[0052] For example, Figure 2 This is a schematic diagram of historical time period division provided by an embodiment of the present disclosure. In the figure, Q is the abbreviation of question, which represents the user's input information, and A is the abbreviation of answer, which represents the system's reply information. QA in the figure represents one or more rounds of dialogue. Figure 2 As shown, the -i day to the -1 day is the first historical period, and the historical conversations before the -i-1 day can be ignored. From the start time of the second historical period, it is the 0th day, that is, the current day. The input information of this round of conversation is represented by Q[0], and the reply information is represented by A[0]. The third historical period includes m rounds of historical conversations, that is, from QA[-m-1] to QA[-1], and on day 0, the historical conversations other than those included in the third historical period are included in the second historical period, such as from QA[-n] to QA[-m] in the figure.
[0053] In the disclosed embodiment, the second prompt information is obtained by summarizing the historical conversations in the second historical period. As the conversation proceeds, the number of historical conversations continues to increase. If the third historical period includes a preset number of historical conversations, the number of historical conversations contained in the second historical period continues to increase. Optionally, each additional round of historical conversations triggers a second summary task. When the second summary task is triggered, the historical conversations in the second historical period corresponding to the current moment are obtained, and the obtained historical conversations are summarized to dynamically determine the second prompt information. In this way, the second prompt information corresponding to the current round of conversations can be obtained before the input information of the current round of conversations is received, thereby improving the speed of determining the target prompt information and thereby improving the response efficiency of the virtual object.
[0054] Exemplarily, by summarizing the historical conversations within the second historical period, the virtual object can still remember the key points of the historical conversations beyond the third historical period. For example, when a user is playing idioms with a virtual object, if there is no second prompt information, after a certain number of dialogue rounds, the virtual object may forget that it is playing idioms. When the user says an idiom again, it may be mistakenly believed that the idiom needs to be explained, and the reply information is the interpretation of the idiom, which affects the user experience. By summarizing the second prompt information, the preset dialogue model virtual object can be reminded that it is playing idioms with the user, without having to record the content of each idiom in the entire process of the chain, ensuring information processing efficiency while making the reply information more accurate and ensuring user experience.
[0055] Step 103: determine target prompt information based on the first prompt information, the second prompt information and the third prompt information, wherein the third prompt information is a historical conversation within a third historical period, the third historical period is later than the second historical period, and the third historical period includes at least the previous round of conversation before the current round of conversation.
[0056] In the embodiment of the present disclosure, the historical dialogue in the third historical period is the historical dialogue closest to the current round of dialogue, and it is not necessary to summarize it, but directly serve as part of the target prompt information. The target prompt information is used as a reference for the preset dialogue model constructed based on the large language model in the next step to generate reply information corresponding to the input information. Optionally, in addition to the first prompt information, the second prompt information and the third prompt information, the target prompt information may also include other prompt information, which is not specifically limited in the embodiment of the present disclosure.
[0057] Step 104: Determine reply information corresponding to the input information based on the target prompt information using a preset dialogue model, wherein the preset dialogue model is constructed based on a large language model.
[0058] Exemplarily, the input data of the preset dialogue model is determined according to the target prompt information and the input information, the reply information corresponding to the input information is determined according to the output data of the preset dialogue model, and the reply information is fed back to the user to complete the current round of dialogue.
[0059] The information processing method provided by the embodiment of the present disclosure, in response to receiving input information of the current round of dialogue, performs information matching for the input information in a historical information library, and determines first prompt information according to the information matching result, wherein the historical information library includes multiple historical dialogue information, and the historical dialogue information is obtained by summarizing historical dialogues in a first historical period, obtains second prompt information, wherein the second prompt information is obtained by summarizing historical dialogues in a second historical period later than the first historical period, determines target prompt information according to the first prompt information, the second prompt information and the third prompt information, wherein the third prompt information is a historical dialogue in a third historical period later than the second historical period, and the third historical period at least includes the previous round of dialogue of the current round of dialogue, and uses a preset dialogue model to determine reply information corresponding to the input information based on the target prompt information, wherein the preset dialogue model is constructed based on a large language model. By adopting the above technical solution, the historical conversations between the virtual object and the user can be divided into multiple time periods according to the distance from the current round of conversation, and different time periods are processed differently. The processing results of the historical conversations corresponding to multiple time periods are used to construct target prompt information, providing richer and more accurate reference information for the preset conversation model built based on the large language model, improving the reply efficiency and the accuracy of the reply information, and thus improving the user experience.
[0060] In some embodiments, the historical conversation information includes first information determined based on the conversation content in the current historical conversation segment, and second information supplementing the conversation content in the current historical conversation segment. Thus, when summarizing the historical conversation segments in the first historical period, not only the first information determined based on the conversation content itself is included, but also the second information supplementing it is included, so that when performing information matching, it is easier and more accurate to obtain relevant information, that is, it is easier and more accurate to recall relevant memories.
[0061] Exemplarily, the second information may include project information of multiple supplementary items, and the supplementary items may include at least one of the conversation time, conversation environment, and subject type of the target subject in the conversation content. The conversation time may refer to the time when the conversation occurs, and the project information may be time information, and the specific type is not limited, such as date (such as January 1), yesterday, the day before yesterday, or a holiday (such as New Year's Day); the conversation environment may include weather, temperature, or noise level; the target subject in the conversation content may be the object discussed in the conversation, such as songs, movies, sports events, or attractions, etc. The subject type may be the category of the target subject, and the specific classification method is not limited. Taking movies as an example, the subject type may include comedy, fantasy, or documentary, etc.
[0062] For example, a user and a virtual object chat about movie A, which is a fantasy movie, but the user did not mention that it was a fantasy movie during the chat. A few days later, if the user asks the virtual object "Have we talked about fantasy movies before?", if no supplementary information is added to the previous historical conversation about movie A, the content cannot be retrieved. If supplementary information is added when summarizing the historical conversation, such as supplementary information about fantasy movies, for example, the historical conversation information obtained after the summary can be "The owner and I talked about fantasy movie A on January 1, and he thought the plot of movie A was exciting,...", then when matching information, the probability of retrieving the above historical conversation can be increased, and the accuracy of the reply information can be improved.
[0063] In some embodiments, the second information is determined by: determining the target conversation scene corresponding to the current historical conversation segment; determining the target supplementary item corresponding to the target conversation scene according to a preset mapping relationship, wherein the preset mapping relationship includes a correspondence between a preset conversation scene and a preset supplementary item, the target conversation scene is a preset conversation scene, the target supplementary item belongs to the preset supplementary item, and the preset supplementary item includes at least one of the conversation time, conversation environment, and the subject type of the target subject in the conversation content; determining the item information of the target supplementary item to obtain the second information. Thus, by presetting the preset mapping relationship, it is possible to quickly determine which items (target supplementary items) need to be supplemented for the current conversation scene, thereby improving the efficiency of determining the second information.
[0064] Exemplarily, multiple dialogue scenes (recorded as preset dialogue scenes) can be preset. For different preset dialogue scenes, supplementary items (recorded as preset supplementary items) that need to be supplemented for each preset dialogue scene can be preset according to actual needs, so as to establish a correspondence between each preset dialogue scene and the preset supplementary items. Different preset dialogue scenes can correspond to the same or different preset supplementary items, and the number of preset supplementary items corresponding to each preset dialogue scene is not limited. When it is necessary to determine the second information, for each historical dialogue segment, the dialogue scene of the current historical dialogue segment can be identified, and then the current dialogue scene can be determined, which is recorded as the target dialogue scene. The target dialogue scene is one of the preset dialogue scenes, that is, the target dialogue scene belongs to the preset dialogue scene. Therefore, the identification of the dialogue scene can be regarded as a classification task. After determining the target dialogue scene, the preset mapping relationship is queried, and the preset supplementary items in the preset mapping relationship that have a corresponding relationship with the target dialogue scene are determined as the target supplementary items corresponding to the target dialogue scene.
[0065] Figure 3This is a flow chart of another information processing method provided by the embodiment of the present disclosure. The embodiment of the present disclosure is optimized based on each optional solution in the above embodiment. The historical information library also includes object portrait information of virtual objects. The object portrait information is dynamically updated according to the historical conversations in the first historical period. Specifically, the method includes the following steps:
[0066] Step 301: In response to receiving input information of this round of conversation, determine a matching type corresponding to the input information, wherein the matching type includes an object portrait matching type and / or a historical conversation matching type.
[0067] In the disclosed embodiment, the object portrait information is the portrait information of the virtual object itself. As the dialogue with the user continues, the virtual object also needs to remember its own characteristics or preferences and other related information. As in the example above, the user and the virtual object chat about movie A. The user asks the virtual object whether it likes movie A. The virtual object's answer to such questions is not pre-set, but is generated by the preset dialogue model according to the actual situation. Assuming that the virtual object's answer is that it likes it, the virtual object needs to remember its answer and update its object portrait information to avoid the virtual object giving a contradictory answer that it does not like movie A when the user asks the virtual object again a few days later without recording it, which affects the user's dialogue interaction experience.
[0068] Figure 4 A schematic diagram of a historical conversation processing process provided by an embodiment of the present disclosure is shown in FIG. Figure 4 As shown, the user inputs on the user side, and the dialogue system main controller on the system side (which can be integrated into the electronic device of the embodiment of the present disclosure) receives the user input (input information) and generates a system reply (reply information) and returns it to the user side. The user input and reply information constitute a historical dialogue, and the dialogue system main controller stores the historical dialogue in the dialogue history management module. The dialogue system main controller can periodically trigger the creation of a first summary task through a signal. The long-term memory task management is responsible for executing the first summary task. The historical dialogue is pulled from the dialogue history management module by dialogue abstract understanding, and the memory block (historical dialogue information) is summarized and stored in the long-term memory storage. The virtual object portrait extraction pulls the historical dialogue from the dialogue history management module, and the abstract portrait result (object portrait information of the virtual object) is stored in the long-term memory storage. The long-term memory storage includes a historical information library.
[0069] Figure 5 A schematic diagram of a reply information determination process provided by an embodiment of the present disclosure is shown in FIG. Figure 5As shown, historical conversation information is stored in the memory block storage, and object portrait information of virtual objects is stored in the virtual object portrait storage. After the conversation system main controller on the system side (which can be integrated into the electronic device of the embodiment of the present disclosure) receives user input (input information), it determines the matching type corresponding to the input information, that is, whether historical conversation information matching or object portrait information matching is required, or both matching are required. The matching type can be related to the current conversation scenario, such as online retrieval scenario, chat scenario and other scenarios.
[0070] Step 302: perform information matching for the input information in the historical information library according to the matching type, and determine the first prompt information according to the information matching result, wherein the historical information library includes object portrait information of the virtual object and multiple historical conversation information, and the historical conversation information is obtained by summarizing the historical conversations in a first historical time period, and the first historical time period includes a preset number of unit time periods.
[0071] like Figure 5 As shown, after determining the matching type, a semantic search service is used to perform a corresponding semantic search to obtain portrait features in the historical conversation information and / or object portrait information that match the input information, and then generate the first prompt information based on the matching result.
[0072] Step 303: Obtain second prompt information, wherein the second prompt information is obtained by summarizing historical conversations in a second historical period, and the second historical period is later than the first historical period.
[0073] Step 304: determine the target prompt information based on the first prompt information, the second prompt information, and the third prompt information, wherein the third prompt information is a historical conversation within a third historical period, the second historical period and the third historical period are included in the same unit time period, the third historical period includes a preset number of rounds of historical conversations, and the preset number of rounds of historical conversations includes the previous round of conversations before the current round of conversations.
[0074] Step 305: Determine reply information corresponding to the input information based on the target prompt information using a preset dialogue model, wherein the preset dialogue model is constructed based on a large language model.
[0075] The information processing method provided by the disclosed embodiment includes not only historical conversation information but also object portrait information of virtual objects in the historical information library for information matching. The object portrait information is dynamically updated according to the historical conversations in the first historical period, so that the preset conversation model can refer to the dynamically updated object portrait information of the virtual object to give more accurate reply information, thereby avoiding inconsistent reply content in conversations with longer intervals and improving user experience.
[0076] In some embodiments, it further includes: obtaining multimedia information played in the second historical period and / or the third historical period; determining fourth prompt information according to the multimedia information; determining target prompt information according to the first prompt information, the second prompt information and the third prompt information, including: determining target prompt information according to the first prompt information, the second prompt information, the third prompt information and the fourth prompt information. In this way, the prompt information can be further enriched, helping the preset dialogue model to understand more dialogue backgrounds, give more accurate responses, and further improve user experience.
[0077] Exemplarily, multimedia information may include relevant information of multimedia content such as the played music, videos, audio books, radio information, and news information. When the part related to the multimedia information is mentioned in the conversation, the reference resolution can be effectively completed, further improving the accuracy of the reply information.
[0078] Figure 6 A schematic diagram of the structure of an information processing device provided by an embodiment of the present disclosure is shown in FIG. Figure 6 As shown, the device comprises:
[0079] A first prompt information determination module 601 is used to, in response to receiving input information of the current round of dialogue, perform information matching for the input information in a historical information database, and determine first prompt information according to the information matching result, wherein the historical information database includes a plurality of historical dialogue information, and the historical dialogue information is obtained by summarizing historical dialogues in a first historical period;
[0080] A second prompt information acquisition module 602, configured to acquire second prompt information, wherein the second prompt information is obtained by summarizing historical conversations in a second historical period, the second historical period being later than the first historical period;
[0081] A target prompt information determination module 603 is used to determine target prompt information according to the first prompt information, the second prompt information and the third prompt information, wherein the third prompt information is a historical conversation in a third historical period, the third historical period is later than the second historical period, and the third historical period includes at least the previous round of conversation of the current round of conversation;
[0082] The reply information determination module 604 is used to determine the reply information corresponding to the input information based on the target prompt information by using a preset dialogue model, wherein the preset dialogue model is constructed based on a large language model.
[0083] The information processing device provided by the disclosed embodiment can divide the historical conversation between the virtual object and the user into multiple time periods according to the distance from the current round of conversation, and different time periods are processed in different ways. The processing results of the historical conversations corresponding to the multiple time periods are used to construct target prompt information, so as to provide richer and more accurate reference information for the preset conversation model constructed based on the large language model, improve the reply efficiency and the accuracy of the reply information, and thus enhance the user experience.
[0084] Optionally, the historical conversation information in the preset historical information library is determined in the following manner: dividing the historical conversations in the first historical time period into multiple historical conversation segments based on a preset segmentation method; for each historical conversation segment in the multiple historical conversation segments, summarizing the current historical conversation segment to obtain historical conversation information corresponding to the current historical conversation segment, wherein the historical conversation information includes first information determined based on the conversation content in the current historical conversation segment, and second information supplementing the conversation content in the current historical conversation segment.
[0085] Optionally, the second information is determined in the following manner: determining a target dialogue scene corresponding to the current historical dialogue segment; determining a target supplementary item corresponding to the target dialogue scene based on a preset mapping relationship, wherein the preset mapping relationship includes a correspondence between a preset dialogue scene and a preset supplementary item, the target dialogue scene is one of the preset dialogue scenes, the target supplementary item belongs to the preset supplementary item, and the preset supplementary item includes at least one of the dialogue time, dialogue environment, and the subject type of the target subject in the dialogue content; determining the item information of the target supplementary item to obtain the second information.
[0086] Optionally, the historical information library also includes object portrait information of virtual objects, and the object portrait information is dynamically updated according to historical conversations in the first historical period;
[0087] Wherein, the first prompt information determination module includes:
[0088] a matching type determining unit, configured to determine, in response to receiving input information of the current round of dialogue, a matching type corresponding to the input information, wherein the matching type includes an object portrait matching type and / or a historical dialogue matching type;
[0089] The information matching unit is used to perform information matching for the input information in a historical information database according to the matching type, and determine the first prompt information according to the information matching result.
[0090] Optionally, the device further comprises:
[0091] A multimedia information acquisition module, used to acquire multimedia information played in the second historical period and / or the third historical period;
[0092] A fourth prompt information determining module, configured to determine fourth prompt information according to the multimedia information;
[0093] The target prompt information determination module is specifically used to determine the target prompt information according to the first prompt information, the second prompt information, the third prompt information and the fourth prompt information.
[0094] Optionally, the first historical time period includes a preset number of unit time periods, the second historical time period and the third historical time period are included in the same unit time period, and the third historical time period includes a preset number of rounds of historical conversations.
[0095] The information processing device provided by the embodiments of the present disclosure can execute the information processing method provided by any embodiment of the present disclosure, and has the corresponding functional modules and beneficial effects of the execution method.
[0096] It is worth noting that the various units and modules included in the above-mentioned device are only divided according to functional logic, but are not limited to the above-mentioned division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other, and are not used to limit the protection scope of the embodiments of the present disclosure.
[0097] Figure 7 This is a schematic diagram of the structure of an electronic device provided by an embodiment of the present disclosure. Figure 7 , which shows an electronic device (eg, Figure 7 The terminal device in the embodiment of the present disclosure may include but is not limited to mobile terminals such as mobile phones, notebook computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 7 The electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.
[0098] like Figure 7As shown, the electronic device 700 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 701, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 702 or a program loaded from a storage device 708 into a random access memory (RAM) 703. In the RAM 703, various programs and data required for the operation of the electronic device 700 are also stored. The processing device 701, the ROM 702, and the RAM 703 are connected to each other via a bus 704. An edit / output (I / O) interface 705 is also connected to the bus 704.
[0099] Typically, the following devices may be connected to the I / O interface 705: an input device 706 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 707 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 708 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 709. The communication device 709 may allow the electronic device 700 to communicate with other devices wirelessly or by wire to exchange data. Although Figure 7 The electronic device 700 is shown with various devices, but it should be understood that it is not required to implement or possess all the devices shown. More or fewer devices may be implemented or possessed instead.
[0100] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a non-transitory computer-readable medium, and the computer program contains program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device 709, or installed from a storage device 708, or installed from a ROM 702. When the computer program is executed by the processing device 701, the above-mentioned functions defined in the method of the embodiment of the present disclosure are executed.
[0101] The names of the messages or information exchanged between multiple devices in the embodiments of the present disclosure are only used for illustrative purposes and are not used to limit the scope of these messages or information.
[0102] The electronic device provided by the embodiment of the present disclosure and the information processing method provided by the above embodiment belong to the same inventive concept. The technical details not fully described in this embodiment can be referred to the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.
[0103] The embodiments of the present disclosure provide a computer storage medium on which a computer program is stored. When the program is executed by a processor, the information processing method provided by the above embodiments is implemented.
[0104] The embodiments of the present disclosure provide a computer program product, including a computer program, which implements the information processing method provided by the above embodiments when executed by a processor.
[0105] It should be noted that the computer-readable medium disclosed above may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, device or device. In the present disclosure, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, in which a computer-readable program code is carried. This propagated data signal may take a variety of forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. The computer readable signal medium may also be any computer readable medium other than a computer readable storage medium, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer readable medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.
[0106] The computer-readable medium may be included in the electronic device, or may exist independently without being incorporated into the electronic device.
[0107] The computer-readable medium carries one or more programs. When the one or more programs are executed by the electronic device, the electronic device: in response to receiving input information of the current round of dialogue, performs information matching for the input information in a historical information library, and determines first prompt information according to the information matching result, wherein the historical information library includes multiple historical dialogue information, and the historical dialogue information is obtained by summarizing historical dialogues in a first historical period; obtains second prompt information, wherein the second prompt information is obtained by summarizing historical dialogues in a second historical period, and the second historical period is later than the first historical period; determines target prompt information according to the first prompt information, the second prompt information and the third prompt information, wherein the third prompt information is a historical dialogue in a third historical period, the third historical period is later than the second historical period, and the third historical period includes at least the previous round of dialogue of the current round of dialogue; determines reply information corresponding to the input information based on the target prompt information using a preset dialogue model, wherein the preset dialogue model is constructed based on a large language model.
[0108] Computer program code for performing the operations of the present disclosure may be written in one or more programming languages or a combination thereof, including, but not limited to, object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., through the Internet using an Internet service provider).
[0109] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present disclosure. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some implementations as replacements, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.
[0110] The units involved in the embodiments described in the present disclosure may be implemented by software or hardware. The name of a module does not limit the module itself in some cases. For example, a reply information determination module may also be described as a "module for determining the reply information corresponding to the input information based on the target prompt information using a preset dialogue model."
[0111] The functions described above herein may be performed at least in part by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), and the like.
[0112] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
[0113] According to one or more embodiments of the present disclosure, there is provided an information processing method, including:
[0114] In response to receiving input information of the current round of dialogue, performing information matching for the input information in a historical information database, and determining first prompt information according to the information matching result, wherein the historical information database includes a plurality of historical dialogue information, and the historical dialogue information is obtained by summarizing historical dialogues within a first historical period;
[0115] Acquire second prompt information, wherein the second prompt information is obtained by summarizing historical conversations in a second historical period, and the second historical period is later than the first historical period;
[0116] Determine target prompt information according to the first prompt information, the second prompt information and the third prompt information, wherein the third prompt information is a historical conversation in a third historical period, the third historical period is later than the second historical period, and the third historical period includes at least a previous round of conversation before the current round of conversation;
[0117] The response information corresponding to the input information is determined based on the target prompt information using a preset dialogue model, wherein the preset dialogue model is constructed based on a large language model.
[0118] According to one or more embodiments of the present disclosure, the historical conversation information in the preset historical information library is determined in the following manner:
[0119] Segmenting the historical conversations in the first historical period into a plurality of historical conversation segments based on a preset segmentation method;
[0120] For each of the multiple historical conversation segments, the current historical conversation segment is summarized to obtain historical conversation information corresponding to the current historical conversation segment, wherein the historical conversation information includes first information determined based on the conversation content in the current historical conversation segment, and second information supplementing the conversation content in the current historical conversation segment.
[0121] According to one or more embodiments of the present disclosure, the second information is determined in the following manner:
[0122] Determining a target dialogue scene corresponding to the current historical dialogue segment;
[0123] Determine the target supplementary item corresponding to the target conversation scene according to a preset mapping relationship, wherein the preset mapping relationship includes a correspondence between a preset conversation scene and a preset supplementary item, the target conversation scene is one of the preset conversation scenes, the target supplementary item belongs to the preset supplementary item, and the preset supplementary item includes at least one of the conversation time, the conversation environment, and the subject type of the target subject in the conversation content;
[0124] The item information of the target supplementary item is determined to obtain the second information.
[0125] According to one or more embodiments of the present disclosure, the historical information library also includes object portrait information of the virtual object, and the object portrait information is dynamically updated according to the historical conversations in the first historical period;
[0126] The step of performing information matching for the input information in a historical information database in response to receiving the input information of the current round of dialogue, and determining the first prompt information according to the information matching result includes:
[0127] In response to receiving input information of the current round of conversation, determining a matching type corresponding to the input information, wherein the matching type includes an object portrait matching type and / or a historical conversation matching type;
[0128] Information matching is performed for the input information in a historical information database according to the matching type, and first prompt information is determined according to the information matching result.
[0129] According to one or more embodiments of the present disclosure, the present invention further includes:
[0130] Acquire multimedia information played in the second historical period and / or the third historical period;
[0131] Determine fourth prompt information according to the multimedia information;
[0132] The determining target prompt information according to the first prompt information, the second prompt information and the third prompt information includes:
[0133] Target prompt information is determined according to the first prompt information, the second prompt information, the third prompt information and the fourth prompt information.
[0134] According to one or more embodiments of the present disclosure, the first historical time period includes a preset number of unit time periods, the second historical time period and the third historical time period are included in the same unit time period, and the third historical time period includes a preset number of rounds of historical conversations.
[0135] The above description is only a preferred embodiment of the present disclosure and an explanation of the technical principles used. Those skilled in the art should understand that the scope of disclosure involved in the present disclosure is not limited to the technical solutions formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above disclosed concept. For example, the above features are replaced with the technical features with similar functions disclosed in the present disclosure (but not limited to) by each other to form a technical solution.
[0136] In addition, although each operation is described in a specific order, this should not be understood as requiring these operations to be performed in the specific order shown or in a sequential order. Under certain circumstances, multitasking and parallel processing may be advantageous. Similarly, although some specific implementation details are included in the above discussion, these should not be interpreted as limiting the scope of the present disclosure. Some features described in the context of a separate embodiment can also be implemented in a single embodiment in combination. On the contrary, the various features described in the context of a single embodiment can also be implemented in multiple embodiments individually or in any suitable sub-combination mode.
[0137] Although the subject matter has been described in language specific to structural features and / or methodological logical actions, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or actions described above. On the contrary, the specific features and actions described above are merely example forms of implementing the claims.
Claims
1. An information processing method, characterized in that: include: In response to receiving input information of the current round of dialogue, performing information matching for the input information in a historical information database, and determining first prompt information according to the information matching result, wherein the historical information database includes a plurality of historical dialogue information, and the historical dialogue information is obtained by summarizing historical dialogues within a first historical period; Acquire second prompt information, wherein the second prompt information is obtained by summarizing historical conversations in a second historical period, and the second historical period is later than the first historical period; Determine target prompt information according to the first prompt information, the second prompt information and the third prompt information, wherein the third prompt information is a historical conversation in a third historical period, the third historical period is later than the second historical period, and the third historical period includes at least a previous round of conversation before the current round of conversation; The response information corresponding to the input information is determined based on the target prompt information using a preset dialogue model, wherein the preset dialogue model is constructed based on a large language model.
2. The method according to claim 1, characterized in that The historical conversation information in the preset historical information library is determined in the following manner: Segmenting the historical conversations in the first historical period into a plurality of historical conversation segments based on a preset segmentation method; For each of the multiple historical conversation segments, the current historical conversation segment is summarized to obtain historical conversation information corresponding to the current historical conversation segment, wherein the historical conversation information includes first information determined based on the conversation content in the current historical conversation segment, and second information supplementing the conversation content in the current historical conversation segment.
3. The method according to claim 2, characterized in that The second information is determined by: Determining a target dialogue scene corresponding to the current historical dialogue segment; Determine the target supplementary item corresponding to the target conversation scene according to a preset mapping relationship, wherein the preset mapping relationship includes a correspondence between a preset conversation scene and a preset supplementary item, the target conversation scene is one of the preset conversation scenes, the target supplementary item belongs to the preset supplementary item, and the preset supplementary item includes at least one of the conversation time, the conversation environment, and the subject type of the target subject in the conversation content; The item information of the target supplementary item is determined to obtain the second information.
4. The method according to claim 1, characterized in that: The historical information library also includes object portrait information of virtual objects, and the object portrait information is dynamically updated according to the historical conversations in the first historical period; The step of performing information matching for the input information in a historical information database in response to receiving the input information of the current round of dialogue, and determining the first prompt information according to the information matching result includes: In response to receiving input information of the current round of conversation, determining a matching type corresponding to the input information, wherein the matching type includes an object portrait matching type and / or a historical conversation matching type; Information matching is performed for the input information in a historical information database according to the matching type, and first prompt information is determined according to the information matching result.
5. The method according to claim 1, characterized in that Also includes: Acquire multimedia information played in the second historical period and / or the third historical period; Determine fourth prompt information according to the multimedia information; The determining target prompt information according to the first prompt information, the second prompt information and the third prompt information includes: Target prompt information is determined according to the first prompt information, the second prompt information, the third prompt information and the fourth prompt information.
6. The method according to claim 1, characterized in that The first historical time period includes a preset number of unit time periods, the second historical time period and the third historical time period are included in the same unit time period, and the third historical time period includes a preset number of rounds of historical conversations.
7. An information processing device, characterized in that: include: a first prompt information determination module, configured to, in response to receiving input information of the current round of dialogue, perform information matching for the input information in a historical information database, and determine first prompt information according to the information matching result, wherein the historical information database includes a plurality of historical dialogue information, and the historical dialogue information is obtained by summarizing historical dialogues within a first historical period; A second prompt information acquisition module, used to acquire second prompt information, wherein the second prompt information is obtained by summarizing historical conversations in a second historical period, and the second historical period is later than the first historical period; a target prompt information determination module, configured to determine target prompt information according to the first prompt information, the second prompt information and the third prompt information, wherein the third prompt information is a historical conversation within a third historical period, the third historical period is later than the second historical period, and the third historical period includes at least a previous round of conversation before the current round of conversation; A reply information determination module is used to determine the reply information corresponding to the input information based on the target prompt information using a preset dialogue model, wherein the preset dialogue model is constructed based on a large language model.
8. An electronic device, characterized in that: The electronic device comprises: one or more processors; a storage device for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the information processing method according to any one of claims 1 to 6.
9. A storage medium containing computer executable instructions, characterized in that: When the computer executable instructions are executed by a computer processor, they are used to perform the information processing method according to any one of claims 1 to 6.
10. A computer program product, characterized in that The invention comprises a computer program, which, when executed by a processor, implements the information processing method according to any one of claims 1 to 6.