Dialog processing method, voice dialog method, computing device, computer-readable storage medium, and computer program product

Abstract

The embodiments of the present disclosure provide a dialog processing method, a voice dialog method, a computing device, a computer-readable storage medium, and a computer program product. The dialog processing method comprises: obtaining target dialog data; on the basis of the target dialog data, determining a processing policy of the target dialog data; and when it is determined that the processing policy is to process the target dialog data on the basis of reference dialog data, inputting the target dialog data and the reference dialog data into a target dialog model, to obtain a reply result of the target dialog data, the reference dialog data being obtained by expanding dialog content on the basis of dialog context information. When the target dialog data is processed on the basis of the reference dialog data, the reference dialog data is used to assist the target dialog model in dialog processing, so that the target dialog model can acquire knowledge from the reference dialog data, and actively output content by combining a semantic understanding capability of the model, so as to avoid dialog interruption and improve the fluency and naturalness of dialog processing.

Description

Dialogue processing methods, voice dialogue methods, computing devices, computer-readable storage media, and computer program products Technical Field

[0001] This disclosure relates to the field of computer technology, and in particular to dialogue processing methods, voice dialogue methods, computing devices, computer-readable storage media, and computer program products. Background Technology

[0002] With the development of artificial intelligence, human-computer dialogue technology has gradually become a widely used technology. Human-computer dialogue focuses on enabling machines to understand, generate, and use natural language to communicate effectively with humans. This involves multiple technical aspects such as speech recognition, natural language understanding, dialogue management, and natural language generation, with the aim of enabling machines to chat, provide information retrieval, and perform service booking tasks like humans.

[0003] Currently, in human-computer interaction, chatbots may experience interruptions in dialogue due to their inability to respond to user input, which does not conform to real-life human conversation scenarios and results in poor realism and fluency in human-computer interaction. Therefore, there is an urgent need for a more realistic and fluent dialogue processing solution. Summary of the Invention

[0004] In view of the above, embodiments of this disclosure provide a dialogue processing method. One or more embodiments of this disclosure also relate to a voice dialogue method, a dialogue processing apparatus, a voice dialogue device, a computing device, a computer-readable storage medium, and a computer program product, in order to solve the technical defects existing in the prior art.

[0005] According to a first aspect of the present disclosure, a dialogue processing method is provided, comprising:

[0006] Obtain target dialogue data;

[0007] Based on the target dialogue data, determine the processing strategy for the target dialogue data;

[0008] When the processing strategy is determined to be processing the target dialogue data based on the reference dialogue data, the target dialogue data and the reference dialogue data are input into the target dialogue model to obtain the response results of the target dialogue data. The reference dialogue data is obtained by expanding the dialogue content based on the dialogue context information.

[0009] According to a second aspect of the present disclosure, a voice dialogue method is provided, comprising:

[0010] Acquire voice dialogue data;

[0011] Based on the voice dialogue data, determine the processing strategy for the voice dialogue data;

[0012] Given that the processing strategy is to process voice dialogue data based on reference dialogue data, the voice dialogue data and reference dialogue data are input into the target dialogue model to obtain the response results of the voice dialogue data. The reference dialogue data is obtained by expanding the dialogue content based on the dialogue context information.

[0013] According to a third aspect of the present disclosure, a voice dialogue method is provided, comprising:

[0014] Obtain the client's voice dialogue data;

[0015] Based on the voice dialogue data, determine the processing strategy for the voice dialogue data;

[0016] When the processing strategy is determined to be processing voice dialogue data based on reference dialogue data, the voice dialogue data and reference dialogue data are input into the target dialogue model to obtain the response results of the voice dialogue data. The reference dialogue data is obtained by expanding the dialogue content based on the dialogue context information.

[0017] The system reads the response results of the voice dialogue data to the client.

[0018] According to a fourth aspect of the present disclosure, a dialogue processing apparatus is provided, comprising:

[0019] The first acquisition module is configured to acquire target dialogue data;

[0020] The first determining module is configured to determine the processing strategy for the target dialogue data based on the target dialogue data.

[0021] The first input module is configured to input the target dialogue data and the reference dialogue data into the target dialogue model when the processing strategy is determined to be processing the target dialogue data based on the reference dialogue data, and to obtain the response result of the target dialogue data. The reference dialogue data is obtained by expanding the dialogue content based on the dialogue context information.

[0022] According to a fifth aspect of the present disclosure, a voice dialogue device is provided, comprising:

[0023] The second acquisition module is configured to acquire voice dialogue data;

[0024] The second determining module is configured to determine the processing strategy for the voice dialogue data based on the voice dialogue data.

[0025] The second input module is configured to input the voice dialogue data and the reference dialogue data into the target dialogue model when the processing strategy is determined to be processing the voice dialogue data based on the reference dialogue data, and to obtain the response result of the voice dialogue data. The reference dialogue data is obtained by expanding the dialogue content based on the dialogue context information.

[0026] According to a sixth aspect of the present disclosure, a voice dialogue device is provided, comprising:

[0027] The third acquisition module is configured to acquire the client's voice dialogue data;

[0028] The third determining module is configured to determine the processing strategy for the voice dialogue data based on the voice dialogue data.

[0029] The third input module is configured to input the voice dialogue data and the reference dialogue data into the target dialogue model when the processing strategy is determined to be processing the voice dialogue data based on the reference dialogue data, and to obtain the response result of the voice dialogue data. The reference dialogue data is obtained by expanding the dialogue content based on the dialogue context information.

[0030] The broadcast module is configured to broadcast the response results of the voice dialogue data to the client.

[0031] According to a seventh aspect of the present disclosure, a computing device is provided, comprising:

[0032] Memory and processor;

[0033] The memory is used to store computer programs / instructions, and the processor is used to execute the computer programs / instructions, which, when executed by the processor, implement the steps of the methods provided in the first, second, or third aspects described above.

[0034] According to an eighth aspect of the present disclosure, a computer-readable storage medium is provided that stores a computer program / instructions that, when executed by a processor, implement the steps of the methods provided in the first, second, or third aspects described above.

[0035] According to a ninth aspect of the present disclosure, a computer program product is provided, including a computer program / instructions that, when executed by a processor, implement the steps of the methods provided in the first, second, or third aspects described above.

[0036] This disclosure provides a dialogue processing method according to one embodiment, comprising: acquiring target dialogue data; determining a processing strategy for the target dialogue data based on the target dialogue data; and, if the processing strategy is determined to be processing the target dialogue data based on reference dialogue data, inputting the target dialogue data and reference dialogue data into a target dialogue model to obtain a response result for the target dialogue data, wherein the reference dialogue data is obtained by expanding the dialogue content based on dialogue context information. First, by parsing the target dialogue data, a processing strategy for the target dialogue data is determined, making the processing strategy more closely aligned with the target dialogue data and ensuring the accuracy of dialogue processing. Second, since the reference dialogue data is obtained by expanding the dialogue content based on dialogue context information, the reference dialogue data is more consistent with real-world dialogue scenarios. Furthermore, if the processing strategy is determined to be processing the target dialogue data based on reference dialogue data, the reference dialogue data is used to assist the target dialogue model in dialogue processing, enabling the target dialogue model to acquire knowledge from the reference dialogue data and, combined with its own semantic understanding capabilities, actively output content, avoiding dialogue interruptions and improving the fluency and naturalness of dialogue processing. Attached Figure Description

[0037] Figure 1 is an architecture diagram of a dialogue processing system provided in an embodiment of this disclosure;

[0038] Figure 2 is an architecture diagram of another dialogue processing system provided in an embodiment of this disclosure;

[0039] Figure 3 is a flowchart of a dialogue processing method provided in an embodiment of this disclosure;

[0040] Figure 4 is a flowchart of a voice dialogue method provided in an embodiment of this disclosure;

[0041] Figure 5 is a flowchart of the processing procedure of a voice dialogue method provided in an embodiment of this disclosure;

[0042] Figure 6 is a flowchart of another voice dialogue method provided in an embodiment of this disclosure;

[0043] Figure 7 is a schematic diagram of the structure of a dialogue processing device provided in an embodiment of the present disclosure;

[0044] Figure 8 is a schematic diagram of the structure of a voice dialogue device provided in an embodiment of this disclosure;

[0045] Figure 9 is a schematic diagram of another dialogue processing device provided in an embodiment of this disclosure;

[0046] Figure 10 is a structural block diagram of a computing device provided in an embodiment of this disclosure. Detailed Implementation

[0047] Numerous specific details are set forth in the following description to provide a full understanding of this disclosure. However, this disclosure can be implemented in many other ways than those described herein, and those skilled in the art can make similar extensions without departing from the spirit of this disclosure. Therefore, this disclosure is not limited to the specific implementations disclosed below.

[0048] The terminology used in one or more embodiments of this disclosure is for the purpose of describing particular embodiments only and is not intended to be limiting of the one or more embodiments of this disclosure. The singular forms “a,” “the,” and “the” as used in one or more embodiments of this disclosure and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in one or more embodiments of this disclosure refers to and includes any or all possible combinations of one or more associated listed items.

[0049] It should be understood that although the terms first, second, etc., may be used to describe various information in one or more embodiments of this disclosure, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, first may also be referred to as second without departing from the scope of one or more embodiments of this disclosure, and similarly, second may also be referred to as first. Depending on the context, the word “if” as used herein may be interpreted as “when”, “in response to a determination”, or “when…”.

[0050] Furthermore, it should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in one or more embodiments of this disclosure are all information and data authorized by the user or fully authorized by all parties. Moreover, the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and corresponding operation entry points are provided for users to choose to authorize or refuse.

[0051] In one or more embodiments of this disclosure, a large model refers to a deep learning model with a large number of model parameters, typically containing hundreds of millions, tens of billions, hundreds of billions, trillions, or even tens of trillions of model parameters. A large model can also be called a foundation model. It is pre-trained using large-scale unlabeled corpora to produce a pre-trained model with hundreds of millions of parameters. Such models can adapt to a wide range of downstream tasks and have good generalization ability. Examples include Large Language Models (LLMs) and multi-modal pre-training models.

[0052] In practical applications, large models only require a small number of samples to fine-tune the pre-trained model before they can be applied to different tasks. Large models can be widely used in fields such as Natural Language Processing (NLP) and Computer Vision. Specifically, they can be applied to computer vision tasks such as Visual Question Answering (VQA), Image Captioning (IC), and Image Generation, as well as natural language processing tasks such as text-based sentiment classification, text summarization, and machine translation. The main application scenarios for large models include digital assistants, intelligent robots, search, online education, office software, e-commerce, and intelligent design.

[0053] First, the terms and concepts involved in one or more embodiments of this disclosure will be explained.

[0054] An agent is a software entity that can perceive its environment and make decisions or perform actions based on its perceptions to achieve a specific goal.

[0055] Multi-agent: refers to a system that utilizes multiple agents to cooperate in achieving a specific goal.

[0056] Voice chat is a technology that enables communication via voice, typically in real-time. Voice chat allows users to converse using a microphone, speaker, or headset. In the field of artificial intelligence, voice chat generally refers to systems where users interact with intelligent assistants via voice, such as smart speakers or virtual assistants on smartphones.

[0057] Conversation Management Unit (CMU): In voice interaction applications, the conversation management unit is the component responsible for controlling the flow of dialogue between the user and the conversation manager. The CMU typically processes user input, executes appropriate dialogue processing strategies, manages the state of the dialogue, determines when to return control to the user, and how to respond to user input. The conversation management unit is a key component in ensuring natural and effective human-computer interaction.

[0058] Function call: In programming, a function call refers to the process of referencing or executing a function within a program. A function is a block of code containing a series of programming instructions designed to accomplish a specific task. When the program reaches the code that calls the function, it jumps to the location where the function is defined, executes the code within the function, and then returns to the place where the function was called to continue executing the rest of the program.

[0059] Deep self-attention (Transformer) models are deep learning architectures based on the attention mechanism, primarily used for processing sequential data. A Transformer model consists of repeatable, stackable encoder and decoder units. This design allows the Transformer to efficiently learn long-term dependencies, making it suitable for various natural language processing tasks, including machine translation, text summarization, and question answering systems.

[0060] Large-scale language models are an artificial intelligence technique based on deep learning neural networks, particularly the Transformer architecture, trained on massive amounts of text data to generate language. These models are capable of understanding language, generating coherent text, answering questions, translating, summarizing, and performing many other language-related tasks.

[0061] Currently, most common voice dialogue frameworks are subordinate designs of assistants. Intelligent assistants respond to human requests via voice to advance the conversation. For example, a user can click the "call button" to have a voice call with an intelligent assistant, but the conversation must follow a question-and-answer paradigm: "user speaks" - "intelligent assistant speaks" - "user speaks." During the conversation, the intelligent assistant lacks the ability to proactively interrupt, propose topics, or continuously repeat information, resulting in a dialogue that doesn't resemble the experience of genuine human casual conversation. Furthermore, in real conversations, humans typically have both active and passive roles; for example, relatively introverted people are not adept at guiding the conversation, leading to interruptions and a lack of realism and fluency.

[0062] To address the aforementioned issues, this disclosure proposes a dialogue processing scheme. The scheme determines a processing strategy for the dialogue data based on its actual condition. When the processing strategy is determined to be processing the target dialogue data based on reference dialogue data, the target dialogue data and reference dialogue data are input into the target dialogue model to obtain the response result of the target dialogue data. The reference dialogue data is used to assist the target dialogue model in dialogue processing, enabling the target dialogue model to acquire knowledge from the reference dialogue data and, combined with its own semantic understanding capabilities, proactively output content, avoiding dialogue interruptions and improving the fluency and naturalness of dialogue processing.

[0063] This disclosure provides a dialogue processing method, and also relates to a voice dialogue method, a dialogue processing apparatus, a voice dialogue device, a computing device, a computer-readable storage medium, and a computer program product, which will be described in detail in the following embodiments.

[0064] Referring to Figure 1, Figure 1 shows an architecture diagram of a dialogue processing system provided in an embodiment of the present disclosure. The dialogue processing system may include a client 100 and a server 200.

[0065] Client 100 is used to send target dialogue data to server 200.

[0066] Server 200 is used to determine the processing strategy for the target dialogue data based on the target dialogue data; if the processing strategy is determined to be processing the target dialogue data based on the reference dialogue data, the server inputs the target dialogue data and the reference dialogue data into the target dialogue model to obtain the response result of the target dialogue data, wherein the reference dialogue data is obtained by expanding the dialogue content based on the dialogue context information; and sends the response result of the target dialogue data to client 100.

[0067] Client 100 is also used to receive the response results of the target dialogue data sent by server 200.

[0068] The solution applied in this embodiment firstly involves analyzing the target dialogue data to determine a processing strategy that better aligns with the target dialogue data, ensuring the accuracy of dialogue processing. Secondly, since the reference dialogue data is derived by expanding the dialogue content based on dialogue context information, it better reflects real-world dialogue scenarios. Furthermore, when the processing strategy is determined to be processing the target dialogue data based on the reference dialogue data, the reference dialogue data assists the target dialogue model in dialogue processing. This allows the target dialogue model to acquire knowledge from the reference dialogue data that matches real-world dialogue scenarios and, combined with its own semantic understanding capabilities, actively output content, avoiding dialogue interruptions and improving the fluency and naturalness of dialogue processing.

[0069] Referring to Figure 2, which illustrates an architecture diagram of another dialogue processing system provided in an embodiment of this disclosure, the dialogue processing system may include multiple clients 100 and a server 200. The clients 100 may include edge devices, and the server 200 may include cloud devices. Multiple clients 100 can establish communication connections through the server 200. In a dialogue processing scenario, the server 200 is used to provide dialogue processing services between the multiple clients 100. Each client 100 can act as a sender or receiver, communicating through the server 200.

[0070] Users can interact with server 200 through client 100 to receive data sent by other clients 100, or send data to other clients 100. In a dialogue processing scenario, a user can publish a data stream to server 200 through client 100, and server 200 can generate a response based on the data stream and push the response to other clients with established communication. Client 100 and server 200 establish a connection via a network. The network provides the medium for the communication link between client 100 and server 200. The network can include various connection types, such as wired, wireless communication links, or fiber optic cables. Data transmitted by client 100 may need to undergo encoding, transcoding, compression, or other processing before being published to server 200.

[0071] Client 100 can be a browser, application (APP), web application such as Hyper Text Markup Language 5 (H5) application, or a lightweight application (also known as a mini-program) or cloud application. Client 100 can be developed based on the software development kit (SDK) of the corresponding service provided by server 200, such as based on the Real Time Communication (RTC) SDK. Client 100 can be deployed in electronic devices and depends on the device to run or on certain APPs on the device. Electronic devices may have a display screen and support information browsing, such as personal mobile terminals such as mobile phones, tablets, and personal computers. Various other types of applications can also be configured in electronic devices, such as human-computer interaction applications, model training applications, text processing applications, web browser applications, shopping applications, search applications, instant messaging tools, email clients, and social platform software.

[0072] Server 200 may include servers providing various services, such as servers providing communication services to multiple clients, servers supporting backend training of models used on clients, and servers processing data sent by clients. It should be noted that server 200 can be implemented as a distributed server cluster composed of multiple servers, or as a single server. The server can also be a server in a distributed system, or a server integrated with blockchain. The server can also be a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms, or an intelligent cloud computing server or intelligent cloud host with artificial intelligence technology.

[0073] It is worth noting that the dialogue processing method provided in the embodiments of this disclosure is generally executed by the server. However, in other embodiments of this disclosure, the client may also have similar functions to the server, thereby executing the dialogue processing method provided in the embodiments of this disclosure. In other embodiments, the dialogue processing method provided in the embodiments of this disclosure may also be executed jointly by the client and the server.

[0074] Referring to Figure 3, Figure 3 shows a flowchart of a dialogue processing method provided in an embodiment of this disclosure, which specifically includes the following steps:

[0075] Step 302: Obtain target dialogue data.

[0076] In one or more embodiments of this disclosure, when dialogue processing begins, target dialogue data can be acquired, and further dialogue processing can be performed based on the target dialogue data to obtain the response result corresponding to the target dialogue data.

[0077] For example, target dialogue data refers to data generated by users during dialogue interaction. Target dialogue data can be data from different scenarios, such as meeting data in a conference scenario, discussion data in a podcast scenario, etc. Target dialogue data can be single-turn dialogue data, such as dialogue data in a question-and-answer scenario, or multi-turn dialogue data, such as dialogue data in scenarios like casual conversation, negotiation, guidance, and debate. The dialogue processing tasks corresponding to target dialogue data may be various, such as speech summarization tasks, speech-to-text transcription tasks, speech content retrieval tasks, etc. Target dialogue data can include at least one of target dialogue content and target dialogue state.

[0078] The target dialogue content refers to the content of actual communication between at least one user. This user can be either a human or a machine.

[0079] The target dialogue content includes questions posed, facts stated, emotions expressed, and viewpoints shared, forming the dialogue entity. The target dialogue content can be in text or audio format. Text content can be a transcript of an audio or video dialogue.

[0080] The target dialogue state reflects the current progress of the conversation. It includes a dialogue state and a silent state. A dialogue state indicates that the user has output content during the conversation, while a silent state indicates that the user has not output content. The target dialogue state helps determine if there are any interruptions in the conversation, thus ensuring the continuity of the conversation even when interruptions occur.

[0081] It should be noted that if the target dialogue data includes the target dialogue content, it means that the user has output content during the current dialogue, and the target dialogue state is a dialogue state. If the target dialogue data does not include the target dialogue content, it means that the user has not output content during the current dialogue, and the target dialogue state is a silent state.

[0082] In practical applications, there are various ways to acquire target dialogue data, and the specific method chosen depends on the actual situation. This disclosure does not impose any limitations on these methods. In one possible implementation, target dialogue data can be received from a user via a client. In another possible implementation, target dialogue data can be read from other data acquisition devices or databases.

[0083] Step 304: Determine the processing strategy for the target dialogue data based on the target dialogue data.

[0084] In one or more embodiments of this disclosure, after obtaining the target dialogue data, a processing strategy for the target dialogue data can be determined based on the target dialogue data.

[0085] For example, the target dialogue data processing strategy refers to the method adopted to understand and resolve the target dialogue data. The processing strategy includes at least one of the following: processing the target dialogue data based on reference dialogue data, processing the target dialogue data using a third-party plugin, and processing the target dialogue data using a target processing model. That is, at least one of the following: a processing strategy based on reference dialogue data, a processing strategy based on a third-party plugin, and a processing strategy based on a target processing model.

[0086] It should be noted that when the target dialogue data includes target dialogue content, the processing strategy for the target dialogue data can be determined based on the target dialogue content. When the target dialogue data does not include target dialogue content but includes target dialogue state, the processing strategy for the target dialogue data can be determined based on the target dialogue state. When the target dialogue data includes both target dialogue content and target dialogue state, the processing strategy for the target dialogue data can be determined based on either the target dialogue content or the target dialogue state.

[0087] Step 306: If the processing strategy is determined to be processing the target dialogue data based on the reference dialogue data, input the target dialogue data and the reference dialogue data into the target dialogue model to obtain the response result of the target dialogue data. The reference dialogue data is obtained by expanding the dialogue content based on the dialogue context information.

[0088] In one or more embodiments of this disclosure, target dialogue data is acquired; after determining a processing strategy for the target dialogue data based on the target dialogue data, further, if the processing strategy is determined to be processing the target dialogue data based on reference dialogue data, the target dialogue data and reference dialogue data are input into the target dialogue model to obtain the response result of the target dialogue data.

[0089] For example, reference dialogue data is used to assist the target dialogue model in processing the target dialogue data. Since the reference dialogue data is obtained by expanding the dialogue content based on dialogue context information, it can be a dialogue script. The reference dialogue data can include dialogue data from one dialogue character or multiple dialogue characters. The target dialogue model can be a pre-trained large model or a machine learning model trained based on multiple sample dialogue data, sample reference dialogue data for each sample dialogue data, and sample response results. The target dialogue model has semantic understanding capabilities, used to efficiently extract dialogue knowledge from the reference dialogue data and process the target dialogue data based on this knowledge to obtain the response results for the target dialogue data. Dialogue context information refers to the environment and conditions in which the reference dialogue data is generated. Dialogue context information includes the dialogue topic, dialogue characters, dialogue location, a brief summary of the dialogue content, the development of the dialogue plot, etc. Therefore, dialogue context information can be the outline of a dialogue script.

[0090] In practical applications, when target dialogue data and reference dialogue data are input into the target dialogue model, preset prompts can be obtained. Dialogue prompts are then constructed based on the target dialogue data and the preset prompts. The dialogue prompts and reference dialogue data are then input into the target dialogue model to obtain the response result for the target dialogue data. For example, the preset prompts could be: "The target dialogue data includes the target dialogue state [target dialogue state] and the target dialogue content [target dialogue content]. Please determine the response result for the target dialogue data based on the reference dialogue data." The specific preset prompts are selected according to the actual situation, and this embodiment does not impose any limitations on them.

[0091] It should be noted that, taking the aforementioned preset prompt information as an example, when constructing dialogue prompt information based on target dialogue data and preset prompt information, the target dialogue state and target dialogue content from the target dialogue data can be added to the preset prompt information to obtain the dialogue prompt information. If the target dialogue data includes target dialogue content, the target dialogue model can match reference dialogue content that is related to the target dialogue content from the reference dialogue data based on the target dialogue content, and determine the reference dialogue content as the response result. The relationship between the reference dialogue content and the target dialogue content can be semantically related, scene-related, sentiment-related, etc., and the specific selection depends on the actual situation; this disclosure does not impose any limitations on this. If the target dialogue data does not include target dialogue content, the dialogue prompt information will also not include target dialogue content. In this case, the target dialogue model can randomly select the response result of the target dialogue data from the reference dialogue data. Alternatively, if the target dialogue data does not include target dialogue content, the user's historical dialogue content can be obtained and added to the position of the target dialogue content in the preset prompt information, thereby enabling the target dialogue model to select the response result of the target dialogue data from the reference dialogue data based on the historical dialogue content.

[0092] The solution applied in this embodiment firstly involves analyzing the target dialogue data to determine a processing strategy that better aligns with the target dialogue data, ensuring the accuracy of dialogue processing. Secondly, since the reference dialogue data is derived by expanding the dialogue content based on dialogue context information, it better reflects real-world dialogue scenarios. Furthermore, when the processing strategy is determined to be processing the target dialogue data based on the reference dialogue data, the reference dialogue data is used to assist the target dialogue model in dialogue processing. This allows the target dialogue model to acquire knowledge from the reference dialogue data and, combined with its own semantic understanding capabilities, proactively output content, avoiding dialogue interruptions and improving the fluency and naturalness of dialogue processing.

[0093] In one optional embodiment of this disclosure, when determining the processing strategy for the target dialogue data, in addition to considering the target dialogue data itself, the user's profile can also be considered. For example, if the user profile indicates that the user has an introverted personality, the processing strategy for the target dialogue data can be determined to be processing the target dialogue data based on reference dialogue data. If the user profile indicates that the user has an extroverted personality, the processing strategy for the target dialogue data can be determined to be direct processing without the assistance of reference dialogue data; that is, the processing strategy for the target dialogue data can be a processing strategy based on a third-party plugin or a processing strategy based on the target processing model.

[0094] In one optional embodiment of this disclosure, the target dialogue data includes target dialogue content; the above-mentioned determination of the processing strategy for the target dialogue data based on the target dialogue data may include the following steps:

[0095] Intent recognition is performed on the target dialogue content to determine the target dialogue intent.

[0096] When the target dialogue intent is an instruction processing intent, the processing strategy is determined to be either a third-party plugin processing the target dialogue data or a target processing model processing the target dialogue data.

[0097] When the target dialogue intent is a non-instruction processing intent, the processing strategy is determined to be to process the target dialogue data based on the reference dialogue data.

[0098] For example, a target conversational intent reflects a user's desired outcome or task through the conversation. Target conversational intents include instruction processing intents and non-instruction processing intents. Instruction processing intents indicate that the user has a specific goal or task to achieve during the conversation. For example, if the target conversation content is a question-and-answer dialogue like "How's the weather today?" or a translation like "Translate 'hello' into English," the target conversational intent is an instruction processing intent. Non-instruction processing intents indicate that the user does not have a specific goal or task to achieve during the conversation. For example, if the target conversation content is casual conversation like "Ugh, I'm so bored," the target conversational intent is a non-instruction processing intent.

[0099] In practical applications, there are various methods for identifying the intent of a target dialogue content and determining the target dialogue intent. The specific method chosen depends on the actual situation, and this disclosure does not impose any limitations on these methods. In one possible implementation, the target dialogue content can be input into an intent recognition model to obtain the target dialogue intent. This intent recognition model is a machine learning model trained based on multiple sample dialogue contents and their corresponding sample intent labels. In another possible implementation, dialogue keywords can be extracted from the target dialogue content. These dialogue keywords are then matched with intent keywords for different dialogue intents. The dialogue intent corresponding to the intent keyword that successfully matches the dialogue keywords is determined as the target dialogue intent.

[0100] The solution applied in this disclosure identifies the target dialogue intent of the target dialogue content and determines the processing strategy for the target dialogue data based on the target dialogue intent, thereby improving the adaptability of the processing strategy to the target dialogue data and enhancing the accuracy of dialogue processing.

[0101] In one optional embodiment of this disclosure, the target dialogue data includes the target dialogue state; the above-mentioned determination of the processing strategy for the target dialogue data based on the target dialogue data may include the following steps:

[0102] When the target dialogue state is silent, the processing strategy is determined to be to process the target dialogue data based on the reference dialogue data.

[0103] It's important to note that a silent state indicates the target dialogue data does not contain the target dialogue content. During the current dialogue, the user provides no input and remains silent. For example, a user might initiate a dialogue task but be introverted and unsure how to begin, or have no problem to solve. In this case, the target dialogue state is silent. To avoid dialogue interruptions due to the user's lack of input, the server can proactively interact with the user to guide the dialogue and advance the process. When the server actively interacts with the user, the processing strategy can be determined as processing the target dialogue data based on reference dialogue data, using the reference data to assist in guiding the user.

[0104] In practical applications, when the target dialogue state is a dialogue state, the processing strategy can be determined to be at least one of the following: a third-party plugin processes the target dialogue data, or a target processing model processes the target dialogue data.

[0105] The solution of this disclosure embodiment determines the processing strategy as processing target dialogue data based on reference dialogue data when the target dialogue state is silent. This takes into account the influence of user personality and uses reference dialogue data to encourage users to actively participate in the dialogue, so that introverted users can also successfully engage in dialogue, thereby lowering the threshold for user dialogue and increasing user participation.

[0106] In one optional embodiment of this disclosure, after determining the processing strategy for the target dialogue data based on the target dialogue data, the following steps may be further included:

[0107] When the processing strategy involves a third-party plugin processing the target dialogue data, the third-party plugin is invoked to process the target dialogue data and obtain the response result of the target dialogue data.

[0108] When the processing strategy is to process the target dialogue data using the target processing model, the target dialogue data is input into the target processing model to obtain the response results of the target dialogue data.

[0109] For example, a third-party plugin is a plugin developed by a third-party configuration provider. Third-party plugins can be implemented on the server side through pluggable coding design, and they can adapt to various web application programming interface (API) call formats. Third-party plugins can be plugins with different functions, such as weather query plugins, ticket purchase plugins, drawing plugins, etc. A target processing model refers to a machine learning model that can process instruction processing tasks, such as large models, text translation models, text summarization extraction models, information extraction models, etc.

[0110] In one optional embodiment of this disclosure, when it is determined that the processing strategy is not based on reference dialogue data to process the target dialogue data, it is explained that the target dialogue data includes target dialogue content. The processing strategy for the target dialogue data can be selected from two options: processing the target dialogue data using a third-party plugin and processing the target dialogue data using a target processing model. In one possible implementation, the target dialogue data can be processed using a target processing model first. If the target processing model cannot process the target dialogue data, the processing strategy is determined to be processing the target dialogue data using a third-party plugin. In another possible implementation, the target dialogue data can be processed using a third-party plugin first. If the third-party plugin cannot process the target dialogue data, the processing strategy is determined to be processing the target dialogue data using a target processing model.

[0111] In practical applications, when calling a third-party plugin to process target dialogue data, the corresponding function of the third-party plugin can be executed, and the function execution result can be used as the response result for the target dialogue data. Furthermore, the function execution result can be input into the target dialogue model for integration and processing to obtain the response result for the target dialogue data.

[0112] By applying the solution of this disclosure embodiment, the target dialogue data is processed by calling a third-party plugin or target processing model, thereby realizing instruction compliance in the dialogue processing process and improving the user experience in the dialogue processing process.

[0113] In one optional embodiment of this disclosure, after determining that the processing strategy is to process the target dialogue data based on the reference dialogue data, and inputting the target dialogue data and the reference dialogue data into the target dialogue model to obtain the response result of the target dialogue data, the following steps may be further included:

[0114] Obtain the updated target dialogue data, which is the data obtained from the feedback of the response results.

[0115] Return to the steps that determine the processing strategy for the target dialogue data based on the target dialogue data, until the dialogue termination condition is met and the target response result is obtained.

[0116] It should be noted that, when the processing strategy is determined to be processing the target dialogue data based on the reference dialogue data, after inputting the target dialogue data and the reference dialogue data into the target dialogue model and obtaining the response result of the target dialogue data, the response result of the target dialogue data can be sent to the client. Users can then conduct multi-turn dialogue tasks based on the response result. The updated target dialogue data may include at least one of the updated target dialogue content and the updated target dialogue state.

[0117] In practical applications, there are various ways to obtain updated target dialogue data, and the specific method should be selected according to the actual situation. This disclosure does not limit the methods used. In one possible implementation, after sending the response to the client, if updated target dialogue content is received from the client within a preset time (e.g., 5 seconds, this value is only an example), the updated target dialogue state is determined to be a dialogue state, and the updated target dialogue data is determined based on the updated target dialogue content and the updated target dialogue state. If no updated target dialogue content is received from the client within the preset time, the updated target dialogue state is determined to be a silent state, and the updated target dialogue data is determined based on the updated target dialogue state. The preset time can be shortened sequentially according to the iteration of dialogue rounds. In another possible implementation, a silent state and a dialogue state can be sent to the client simultaneously. If the client selects the silent state, it indicates that the user currently has no content output, and the silent state is determined as the updated target dialogue state. The updated target dialogue data is then determined based on the updated target dialogue state. When the client selects a dialogue state, it indicates that the user currently has content to output. The dialogue state is then determined as the updated target dialogue state, and the updated target dialogue content sent by the client is received. The updated target dialogue data is then determined based on the updated target dialogue content and the updated target dialogue state.

[0118] The solution of this disclosure, by obtaining the updated target dialogue data of the client in response to the response after sending the reply result to the client, and then starting multi-round dialogue iteration processing, makes the dialogue processing more flexible and improves the user's dialogue processing experience.

[0119] In one optional embodiment of this disclosure, after determining that the processing strategy is to process the target dialogue data based on the reference dialogue data, and inputting the target dialogue data and the reference dialogue data into the target dialogue model to obtain the response result of the target dialogue data, the following steps may be further included:

[0120] Identify at least one target dialogue role corresponding to the response result.

[0121] The response result is read aloud based on the voice configuration information of at least one target dialogue character.

[0122] For example, the target dialogue character refers to the virtual persona used to communicate with the user during the dialogue process. For instance, the target dialogue character could be a girl with a soft, gentle voice, or a loud, confident, and sunny boy. Voice configuration information refers to the configuration information of the target dialogue character when communicating with the user via voice, including timbre, volume, speech rate, language type, etc.

[0123] It should be noted that there are several ways to send the response to the target dialogue data to the client. The response can be sent directly to the client in text form, or the response can be read aloud to the client.

[0124] In practical applications, there are multiple ways to determine at least one target dialogue role corresponding to the response result, and the specific method chosen depends on the actual situation. This disclosure does not impose any limitations on this approach. In one possible implementation, at least one dialogue role can be randomly selected from multiple candidate dialogue roles as the target dialogue role corresponding to the response result. In another possible implementation, since the response result is based on reference dialogue data, the reference dialogue segment to which the response result belongs can be determined, and the dialogue role of the reference dialogue segment can be identified as the dialogue role of the response result.

[0125] By applying the solution of this disclosure embodiment, at least one target dialogue role corresponding to the response result is directly determined, and the response result is broadcast by voice according to the voice configuration information of the target dialogue role. Users do not need to manually select the broadcast voice, speech rate and other information, which reduces the threshold for users to participate in dialogue processing, simplifies the user operation process and improves the user experience.

[0126] In one optional embodiment of this disclosure, the method for generating reference dialogue data is described. That is, before inputting the target dialogue data and reference dialogue data into the target dialogue model to obtain the response result of the target dialogue data, when the processing strategy is determined to be processing the target dialogue data based on the reference dialogue data, the following steps may be included:

[0127] Obtain the dialogue generation data.

[0128] Extract contextual information from the generated dialogue data to determine the contextual information of the dialogue.

[0129] Expand the dialogue content based on the contextual information to obtain reference dialogue data.

[0130] For example, dialogue generation data is used to generate reference dialogue data. Dialogue generation data can be information, facts, related reports, etc. There are various ways to obtain dialogue generation data, and the specific method is selected according to the actual situation; this disclosure does not limit this approach. In one possible implementation, dialogue generation data can be read from other data acquisition devices or databases. In another possible implementation, dialogue generation data can be collected from a large number of data sources according to data acquisition rules. These data acquisition rules may include multiple dimensions such as real-time performance, popularity, and factual accuracy, and data sources may include news websites, entertainment websites, science websites, etc. For example, the web application programming interface of the data source can be called to automatically collect data from information websites, and natural language processing technology can be used to identify and extract dialogue generation data from the collected data. Alternatively, information can be crawled from relevant resource websites under the premise of conforming to machine protocols, and natural language processing technology can be used to identify and extract dialogue generation data from the crawled data.

[0131] It should be noted that there are multiple ways to extract contextual information from dialogue-generated data and determine the dialogue contextual information. The specific method should be selected based on the actual situation, and this disclosure does not impose any limitations on this approach. In one possible implementation, the content of the dialogue-generated data can be analyzed to capture keywords, phrases, and sentence structures, understand the development path and logical relationships of the dialogue-generated data, and obtain the dialogue contextual information. In another possible implementation, a text generation model can be used to extract contextual information from the dialogue-generated data to determine the dialogue contextual information.

[0132] Furthermore, there are various ways to expand the dialogue content based on the dialogue context information to obtain reference dialogue data. The specific method chosen depends on the actual situation, and this disclosure does not impose any limitations on this approach. In one possible implementation, multiple users can be invited to participate in discussions and co-creation based on the dialogue context information through an online collaboration platform to obtain reference dialogue data. In another possible implementation, a text generation model can be used to expand the dialogue content based on the dialogue context information to obtain reference dialogue data.

[0133] The solution of this disclosure embodiment uses a phased reference dialogue data generation scheme. In the first phase, dialogue context information is generated based on dialogue generation data. In the second phase, script expansion is performed based on the dialogue context information to obtain reference dialogue data. This integrates scattered dialogue generation data to obtain reference dialogue data that can provide in-depth content analysis and knowledge. This allows the target dialogue model to acquire knowledge from the reference dialogue data and, combined with the model's own semantic understanding ability, actively output content, avoid dialogue interruptions, and improve the fluency and naturalness of dialogue processing.

[0134] In one optional embodiment of this disclosure, the above-described extraction of contextual information from the dialogue generation data to determine the dialogue contextual information may include the following steps:

[0135] The extracted prompts and dialogue generation data are input into the text generation model to obtain dialogue context information. The extracted prompts are used to guide the text generation model to extract context information from the dialogue generation data.

[0136] Expanding dialogue content based on contextual information to obtain reference dialogue data may include the following steps:

[0137] The expanded prompts and dialogue context information are input into the text generation model to obtain reference dialogue data. The expanded prompts are used to guide the text generation model to expand the dialogue content based on the dialogue context information.

[0138] For example, the text generation model can be a pre-trained large model, or a machine learning model trained based on multiple sample dialogue generation data, sample dialogue context information of each sample dialogue generation data, and sample reference dialogue data. The expansion prompts can include the number of characters, dialogue style, and story structure of the reference dialogue data, thereby guiding the text generation model to generate multi-character, logical, and engaging reference dialogue data. The expansion prompts can be text expansion prompts, such as "Please expand the dialogue content based on the dialogue context information. The expanded reference dialogue data includes three dialogue characters. Output the expanded reference dialogue data." The extraction prompts can be text extraction prompts, such as "Please extract context information from the dialogue generation data. Output the extracted dialogue context information."

[0139] It should be noted that when generating reference dialogue data based on the generated dialogue data, the data type of the generated dialogue data can be identified. Based on the data type, the generated dialogue data can be divided into multiple generated data sets, and then reference dialogue data corresponding to each generated data set can be generated. Among these, data types include entertainment data types, popular science data types, and so on.

[0140] The solution applied in this disclosure can efficiently and accurately generate dialogue context information and reference dialogue data by leveraging the semantic understanding capabilities of the text generation model and by extracting and expanding prompt information.

[0141] In one optional embodiment of this disclosure, after obtaining the reference dialogue data, suitable dialogue characters can be created or selected based on the content of the reference dialogue data, and each character can be given a detailed personality, backstory, and appearance design. That is, after obtaining the reference dialogue data by expanding the dialogue content based on the dialogue context information, the following steps may also be included:

[0142] Role analysis is performed on the reference dialogue data to identify at least one dialogue role and the corresponding reference dialogue segment.

[0143] Generate character description information for each dialogue character based on the reference dialogue fragments corresponding to them.

[0144] Based on the character description information, generate the voice configuration information for the dialogue characters.

[0145] For example, character description information is used to describe the personality, backstory, and so on of a dialogue character. Character description information may include text character description information and image character description information, such as personality text description information, and image character description information such as an image of the dialogue character's appearance.

[0146] It should be noted that there are multiple ways to generate character description information for dialogue characters based on the reference dialogue fragments corresponding to each character. The specific method should be selected according to the actual situation, and this disclosure does not impose any limitations on this method. In one possible implementation of this disclosure, the reference dialogue fragment can be input into a text generation model to obtain text character description information. In another possible implementation of this disclosure, the reference dialogue fragment can be input into an image generation model to obtain image character description information.

[0147] In practical applications, there are various ways to generate voice configuration information for dialogue characters based on character description information. The specific method chosen depends on the actual situation, and this disclosure does not impose any limitations on this. In one possible implementation, the voice configuration information of each dialogue character can be randomly selected from multiple candidate voice configuration information. In another possible implementation, feature extraction can be performed on the character description information to obtain character description features. Feature extraction can also be performed on the candidate description information of the candidate voice configuration information to obtain candidate description features. The character description features and candidate description features are then matched, and the candidate voice configuration information corresponding to the candidate description feature that best matches the character description features is determined as the voice configuration information of the dialogue character.

[0148] By applying the solution of this disclosure embodiment, voice configuration information of each dialogue character is generated, realizing the simulation of multiple characters and the output of a humanized script, making the dialogue richer and more engaging, and enhancing the dialogue interaction and sense of participation.

[0149] In one optional embodiment of this disclosure, after expanding the dialogue content based on the dialogue context information to obtain reference dialogue data, the following steps may be further included:

[0150] The dialogue content of the reference dialogue data is reviewed using dialogue review rules to obtain the review results. The dialogue review rules include at least one of the following: grammar review rules, compliance review rules, and copyright review rules.

[0151] If the review result is "review failed", adjust the reference dialogue data to obtain the adjusted reference dialogue data, which conforms to the dialogue review rules.

[0152] It should be noted that the review result indicates whether the review of the reference dialogue data was successful or unsuccessful. After obtaining the reference dialogue data, in order to further improve the accuracy of the responses obtained from dialogue processing using the reference dialogue data, dialogue review rules can be used to review the reference dialogue data. This includes reviewing the accuracy, compliance, and grammatical errors of the reference dialogue data, whether it contains sensitive content, and ensuring that the reference dialogue data does not offend or violate copyright law. Furthermore, in the event of a failed review, the reference dialogue data can be adjusted using text proofreading tools to obtain revised reference dialogue data.

[0153] Grammar review rules are standards used to check whether reference dialogue data conforms to the grammatical rules of a specific language. Grammar review rules ensure that sentences in the reference dialogue data have correct structure, appropriate punctuation, and consistent tenses. For example, grammar review rules might check whether sentences are complete and clearly structured, whether sentences have subject-verb agreement, and whether punctuation is used correctly. By implementing grammar review rules, the readability and professionalism of reference dialogue data can be improved, avoiding misunderstandings.

[0154] Compliance review rules are standards used to ensure that content complies with the legal and regulatory requirements of a specific industry or region. These rules include, but are not limited to, data protection laws, privacy policies, advertising standards, and financial information disclosure requirements. Compliance review rules protect user rights, prevent misleading information, and maintain market order.

[0155] Copyright review rules are standards used to check whether referenced dialogue data infringes on the intellectual property rights of others. Copyright review rules include, but are not limited to, rules protecting the copyright of written works, music, images, and other creative works. Through copyright review rules, the originality and legality of content can be ensured, avoiding legal issues arising from unauthorized use of others' works.

[0156] In one optional embodiment of this disclosure, the target dialogue data includes target dialogue speech; before inputting the target dialogue data and reference dialogue data into the target dialogue model to obtain the response result of the target dialogue data, when the processing strategy is determined to be processing the target dialogue data based on reference dialogue data, the following steps may be included:

[0157] Speech recognition is performed on the target dialogue to obtain the target dialogue text.

[0158] When the processing strategy is determined to be processing the target dialogue data based on the reference dialogue data, inputting the target dialogue data and the reference dialogue data into the target dialogue model to obtain the response result of the target dialogue data may include the following steps:

[0159] If the processing strategy is determined to be processing the target dialogue text based on the reference dialogue data, the target dialogue text and the reference dialogue data are input into the target dialogue model to obtain the response result of the target dialogue voice.

[0160] It should be noted that the input to the target dialogue model is usually in text format. In order for the target dialogue model to accurately process the target dialogue data, speech recognition can be performed on the target dialogue speech to obtain the target dialogue text. Target dialogue speech refers to the voice data generated by the user during the dialogue interaction.

[0161] In practical applications, there are various methods for performing speech recognition on target dialogue speech to obtain target dialogue text. The specific method chosen depends on the actual situation, and this disclosure does not impose any limitations on this approach. In one possible implementation, the target dialogue speech can be recognized using a speech recognition tool to obtain the target dialogue text. In another possible implementation, the target dialogue speech can be input into a speech recognition model to obtain the target dialogue text.

[0162] The solution of this disclosure embodiment obtains the target dialogue text by performing speech recognition on the target dialogue speech, and then directly processes the target dialogue text using the target dialogue model. This allows the target dialogue model to acquire knowledge from the target dialogue text and, combined with the model's own semantic understanding ability, actively output content, avoid dialogue interruptions, and improve the fluency and naturalness of dialogue processing.

[0163] The following description, in conjunction with Figure 4, uses the application of the dialogue processing method provided in this disclosure in a voice chat scenario as an example to further illustrate the dialogue processing method. Figure 4 shows a flowchart of a voice dialogue method provided in an embodiment of this disclosure, specifically including the following steps:

[0164] Step 402: Obtain voice dialogue data.

[0165] Step 404: Determine the processing strategy for the voice dialogue data based on the voice dialogue data.

[0166] Step 406: If the processing strategy is determined to be processing voice dialogue data based on reference dialogue data, input the voice dialogue data and reference dialogue data into the target dialogue model to obtain the response results of the voice dialogue data. The reference dialogue data is obtained by expanding the dialogue content based on the dialogue context information.

[0167] It should be noted that the implementation methods of steps 402 to 406 are the same as those of steps 302 to 306 described above, and will not be repeated in this embodiment. Voice dialogue data refers to voice data generated by the user during voice dialogue interaction. Voice dialogue data may include at least one of voice dialogue content and voice dialogue state. The voice dialogue data processing strategy refers to the method adopted to understand and resolve the voice dialogue data. The voice dialogue data processing strategy includes at least one of the following: processing voice dialogue data based on reference dialogue data, processing voice dialogue data using a third-party plugin, and processing voice dialogue data using a target processing model; that is, at least one of the following: a processing strategy based on reference dialogue data, a processing strategy based on a third-party plugin, and a processing strategy based on a target processing model.

[0168] The solution applied in this embodiment firstly involves analyzing the voice dialogue data to determine the processing strategy for the target dialogue data, making the processing strategy more closely aligned with the voice dialogue data and ensuring the accuracy of the voice dialogue. Secondly, since the reference dialogue data is obtained by expanding the dialogue content based on dialogue context information, the reference dialogue data is more consistent with real-world dialogue scenarios. Furthermore, when the processing strategy is determined to be processing the voice dialogue data based on the reference dialogue data, the reference dialogue data is used to assist the target dialogue model in processing the voice dialogue. This allows the target dialogue model to acquire knowledge from the reference dialogue data and, combined with its own semantic understanding capabilities, actively output content, avoiding dialogue interruptions and improving the fluency and naturalness of the voice dialogue.

[0169] Referring to Figure 5, which shows a flowchart of a voice dialogue method provided in an embodiment of the present disclosure, the dialogue processing method can be implemented through active interaction of Multi-Agent. The processing can be divided into two stages: offline content creation and real-time voice chat. The two stages will be described in detail below.

[0170] Offline content creation: A Content Collection Agent collects real-time information, data, and resources from high-quality data sources, and uses natural language processing technology to identify and extract dialogue generation data. A Script Creation Agent extracts contextual information from the dialogue generation data, determines the dialogue context, expands the dialogue content based on the context, and obtains reference dialogue data. A Content Review Agent uses dialogue review rules to review the reference dialogue data, obtaining daily updated and approved reference dialogue data. A Character Design Agent parses the reference dialogue data, identifies at least one dialogue character and at least one corresponding reference dialogue segment, generates character description information for at least one dialogue character based on the corresponding reference dialogue segment, and generates voice configuration information for at least one dialogue character based on the character description information.

[0171] Real-time voice chat: The system receives target dialogue voice messages from users, performs speech recognition on the voice messages to obtain the target dialogue text, and then engages in conversation with the user through a dialogue processing system. The dialogue processing system can load reference dialogue data obtained from offline content creation as an additional knowledge base during processing, while simultaneously processing the dialogue based on the target dialogue model. When the user has no topic or remains silent, the dialogue processing system can determine the current dialogue state and recommend topics using reference dialogue data. During recommendations, multiple dialogue roles (such as dialogue role A and dialogue role B) can cooperate to deliver the voice messages. Users can interrupt at any time to join the discussion, or if a user is not interested in a topic, they can express their desire to switch topics or stop the playback.

[0172] As shown in Figure 5, the dialogue processing system includes a dialogue management unit, a dialogue response unit, and a function call unit. The dialogue management unit controls the dialogue flow, including maintaining the dialogue state and deciding on processing strategies (selecting topics for recommendation from reference dialogue data, responding to user commands, and deciding which dialogue role should reply). The dialogue response unit receives information from the dialogue management unit, calls the target dialogue model or target processing model to respond to the current round's content, and outputs text content. The dialogue response unit can also broadcast the output text content via voice through dialogue roles A and B. The function call unit receives information from the dialogue management unit, executes the corresponding function calls, and returns the function execution results to the dialogue response unit for further processing.

[0173] The solution in this disclosure uses offline content creation to obtain reference dialogue data as real-time material for voice chat. Real-time voice interaction with the user is achieved through multi-agent responses and different processing strategies. Furthermore, by combining reference dialogue data with real-time voice chat, users can interrupt discussions at any time when they can listen to the reference dialogue data. They can also switch topics, exit, or engage in other discussions using voice commands, making the voice chat process more flexible. The dialogue processing system can proactively engage in conversation with the user when they are bored or silent, which is more in line with the characteristics of natural human dialogue. It also supports voice playback of responses based on the voice configuration information of at least one target dialogue character, further enhancing user immersion and improving the fluency and naturalness of the entire voice chat conversation.

[0174] Referring to Figure 6, Figure 6 shows a flowchart of another voice dialogue method provided in an embodiment of this disclosure, which specifically includes the following steps:

[0175] Step 602: Obtain the client's voice dialogue data.

[0176] Step 604: Determine the processing strategy for the voice dialogue data based on the voice dialogue data.

[0177] Step 606: If the processing strategy is determined to be processing voice dialogue data based on reference dialogue data, input the voice dialogue data and reference dialogue data into the target dialogue model to obtain the response results of the voice dialogue data. The reference dialogue data is obtained by expanding the dialogue content based on the dialogue context information.

[0178] Step 608: Announce the response to the voice dialogue data to the client.

[0179] It should be noted that the implementation methods of steps 602 to 606 are the same as those of steps 302 to 306 described above, and will not be repeated in this embodiment.

[0180] In practical applications, when broadcasting the response results of voice dialogue data to the client, one possible implementation is to randomly select at least one dialogue role from multiple candidate dialogue roles as the target dialogue role corresponding to the response result, and then broadcast the response result based on the voice configuration information of the at least one target dialogue role. Another possible implementation is to determine the reference dialogue segment to which the response result belongs in the reference dialogue data, identify the dialogue role of the reference dialogue segment as the at least one target dialogue role corresponding to the response result, and then broadcast the response result based on the voice configuration information of the at least one target dialogue role.

[0181] The solution of this disclosure uses reference dialogue data to assist the target dialogue model in voice dialogue processing, enabling the target dialogue model to acquire knowledge from the reference dialogue data and, combined with the model's own semantic understanding ability, actively output content, avoid dialogue interruptions, improve the fluency and naturalness of voice dialogue, and further enhance the user experience by broadcasting the response results of the voice dialogue data to the client.

[0182] Corresponding to the above-described dialogue processing method embodiments, this disclosure also provides a dialogue processing apparatus embodiment. Figure 7 shows a schematic diagram of the structure of a dialogue processing apparatus provided in one embodiment of this disclosure. As shown in Figure 7, the apparatus includes:

[0183] The first acquisition module 702 is configured to acquire target dialogue data.

[0184] The first determining module 704 is configured to determine the processing strategy for the target dialogue data based on the target dialogue data.

[0185] The first input module 706 is configured to input the target dialogue data and the reference dialogue data into the target dialogue model when the processing strategy is determined to be processing the target dialogue data based on the reference dialogue data, and to obtain the response result of the target dialogue data. The reference dialogue data is obtained by expanding the dialogue content based on the dialogue context information.

[0186] Optionally, the target dialogue data includes target dialogue content; the first determining module 704 is further configured to perform intent recognition on the target dialogue content to determine the target dialogue intent; if the target dialogue intent is an instruction processing intent, the processing strategy is determined to be that a third-party plugin processes the target dialogue data or a target processing model processes the target dialogue data; if the target dialogue intent is a non-instruction processing intent, the processing strategy is determined to be that the target dialogue data is processed based on reference dialogue data.

[0187] Optionally, the target dialogue data includes the target dialogue state; the first determining module 704 is further configured to determine the processing strategy as processing the target dialogue data based on the reference dialogue data when the target dialogue state is silent.

[0188] Optionally, the device further includes: a fourth acquisition module configured to acquire dialogue generation data; extract contextual information from the dialogue generation data to determine dialogue contextual information; and expand the dialogue content based on the dialogue contextual information to obtain reference dialogue data.

[0189] Optionally, the fourth acquisition module is further configured to input the extracted prompt information and dialogue generation data into the text generation model to obtain dialogue context information, wherein the extracted prompt information is used to guide the text generation model to extract context information from the dialogue generation data; and to input the expanded prompt information and dialogue context information into the text generation model to obtain reference dialogue data, wherein the expanded prompt information is used to guide the text generation model to expand the dialogue content based on the dialogue context information.

[0190] Optionally, the device further includes: a parsing module configured to perform role parsing on the reference dialogue data, determine at least one dialogue role and a reference dialogue segment corresponding to the dialogue role; generate role description information for the dialogue role based on the reference dialogue segments corresponding to each dialogue role; and generate voice configuration information for the dialogue role based on the role description information.

[0191] Optionally, the device further includes: an auditing module configured to audit the reference dialogue data using dialogue auditing rules to obtain an audit result, wherein the dialogue auditing rules include at least one of grammar auditing rules, compliance auditing rules, and copyright auditing rules; and, in the case of an audit failure, to adjust the reference dialogue data to obtain adjusted reference dialogue data, wherein the adjusted reference dialogue data conforms to the dialogue auditing rules.

[0192] Optionally, the target dialogue data includes target dialogue speech; the device further includes: a recognition module configured to perform speech recognition on the target dialogue speech to obtain target dialogue text; and a first input module 706 further configured to, when the processing strategy is determined to be processing the target dialogue text based on reference dialogue data, input the target dialogue text and reference dialogue data into the target dialogue model to obtain the response result of the target dialogue speech.

[0193] Optionally, the device further includes: a fourth determining module, configured to determine at least one target dialogue role corresponding to the response result; and to broadcast the response result by voice according to the voice configuration information of the at least one target dialogue role.

[0194] Optionally, the device further includes: a fourth acquisition module configured to acquire updated target dialogue data, wherein the updated target dialogue data is data obtained by providing feedback on the response result; and to return to the step of determining the processing strategy of the target dialogue data based on the target dialogue data until the dialogue end condition is met and the target response result is obtained.

[0195] Optionally, the device further includes: a processing module configured to, when the processing strategy is to process the target dialogue data using a third-party plugin, call the third-party plugin to process the target dialogue data and obtain a response result for the target dialogue data; and when the processing strategy is to process the target dialogue data using a target processing model, input the target dialogue data into the target processing model and obtain a response result for the target dialogue data.

[0196] By applying the solution of this disclosure embodiment, since the dialogue processing device includes a first determining module that can determine a processing strategy based on the target dialogue data, the processing strategy is more closely aligned with the target dialogue data, ensuring the accuracy of dialogue processing. Secondly, since the reference dialogue data is obtained by expanding the dialogue content based on dialogue context information, the reference dialogue data is more consistent with real dialogue scenarios. In addition, when the processing strategy is determined to be processing the target dialogue data based on the reference dialogue data, the first input module can use the reference dialogue data to assist the target dialogue model in dialogue processing, enabling the target dialogue model to acquire knowledge from the reference dialogue data and, combined with the model's own semantic understanding ability, actively output content, avoiding dialogue interruptions and improving the fluency and naturalness of dialogue processing.

[0197] The above is an illustrative scheme of a dialogue processing device according to this embodiment. It should be noted that the technical solution of this dialogue processing device and the technical solution of the above-described dialogue processing method belong to the same concept. For details not described in detail in the technical solution of the dialogue processing device, please refer to the description of the technical solution of the above-described dialogue processing method.

[0198] Corresponding to the above-described voice dialogue method embodiments, this disclosure also provides a voice dialogue device embodiment. Figure 8 shows a schematic diagram of the structure of a voice dialogue device provided in one embodiment of this disclosure. As shown in Figure 8, the device includes:

[0199] The second acquisition module 802 is configured to acquire voice dialogue data.

[0200] The second determining module 804 is configured to determine the processing strategy for the voice dialogue data based on the voice dialogue data.

[0201] The second input module 806 is configured to input the voice dialogue data and the reference dialogue data into the target dialogue model when the processing strategy is determined to be processing the voice dialogue data based on the reference dialogue data, and to obtain the response result of the voice dialogue data. The reference dialogue data is obtained by expanding the dialogue content based on the dialogue context information.

[0202] By applying the solution of this disclosure embodiment, since the voice dialogue device includes a second determining module that can determine a processing strategy based on the voice dialogue data, the processing strategy is more closely aligned with the voice dialogue data, ensuring the accuracy of the voice dialogue. Secondly, since the reference dialogue data is obtained by expanding the dialogue content based on dialogue context information, the reference dialogue data is more consistent with real dialogue scenarios. In addition, when the processing strategy is determined to be processing the voice dialogue data based on the reference dialogue data, the second input module can use the reference dialogue data to assist the target dialogue model in performing voice dialogue, enabling the target dialogue model to acquire knowledge from the reference dialogue data and, combined with the model's own semantic understanding ability, actively output content, avoiding dialogue interruptions and improving the fluency and naturalness of the voice dialogue.

[0203] The above is an illustrative scheme of a voice dialogue device according to this embodiment. It should be noted that the technical solution of this voice dialogue device and the technical solution of the above-described voice dialogue method belong to the same concept. For details not described in detail in the technical solution of the voice dialogue device, please refer to the description of the technical solution of the above-described voice dialogue method.

[0204] Corresponding to the above-described voice dialogue method embodiments, this disclosure also provides a voice dialogue device embodiment. Figure 9 shows a schematic diagram of another dialogue processing device provided in one embodiment of this disclosure. As shown in Figure 9, the device includes:

[0205] The third acquisition module 902 is configured to acquire the client's voice dialogue data;

[0206] The third determining module 904 is configured to determine the processing strategy for the voice dialogue data based on the voice dialogue data.

[0207] The third input module 906 is configured to input the voice dialogue data and the reference dialogue data into the target dialogue model when the processing strategy is determined to be processing the voice dialogue data based on the reference dialogue data, and to obtain the response result of the voice dialogue data. The reference dialogue data is obtained by expanding the dialogue content based on the dialogue context information.

[0208] The broadcast module 908 is configured to broadcast the response results of the voice dialogue data to the client.

[0209] The solution of this disclosure uses reference dialogue data to assist the target dialogue model in voice dialogue processing, enabling the target dialogue model to acquire knowledge from the reference dialogue data and, combined with the model's own semantic understanding ability, actively output content, avoid dialogue interruptions, improve the fluency and naturalness of voice dialogue, and further enhance the user experience by broadcasting the response results of the voice dialogue data to the client.

[0210] The above is an illustrative scheme of a voice dialogue device according to this embodiment. It should be noted that the technical solution of this voice dialogue device and the technical solution of the above-described voice dialogue method belong to the same concept. For details not described in detail in the technical solution of the voice dialogue device, please refer to the description of the technical solution of the above-described voice dialogue method.

[0211] Figure 10 shows a structural block diagram of a computing device according to an embodiment of the present disclosure. The components of the computing device 1000 include, but are not limited to, a memory 1010 and a processor 1020. The processor 1020 is connected to the memory 1010 via a bus 1030, and a database 1050 is used to store data.

[0212] The computing device 1000 also includes an access device 1040, which enables the computing device 1000 to communicate via one or more networks 1060. Examples of these networks include Public Switched Telephone Network (PSTN), Local Area Network (LAN), Wide Area Network (WAN), Personal Area Network (PAN), or combinations of communication networks such as the Internet. The access device 1040 may include one or more of any type of wired or wireless network interface (e.g., Network Interface Card (NIC)), such as an IEEE 802.11 Wireless Local Area Networks (WLAN) interface, a Wi-MAX (World Interoperability for Microwave Access) interface, an Ethernet interface, a Universal Serial Bus (USB) interface, a cellular network interface, a Bluetooth interface, a Near Field Communication (NFC) interface, and so on.

[0213] In one embodiment of this disclosure, the aforementioned components of the computing device 1000, as well as other components not shown in FIG. 10, may be interconnected, for example, via a bus. It should be understood that the computing device block diagram shown in FIG. 10 is merely for illustrative purposes and is not intended to limit the scope of this disclosure. Those skilled in the art can add or replace other components as needed.

[0214] The computing device 1000 can be any type of stationary or mobile computing device, including mobile computers or mobile computing devices (e.g., tablet computers, personal digital assistants, laptop computers, notebook computers, netbooks, etc.), mobile phones (e.g., smartphones), wearable computing devices (e.g., smartwatches, smart glasses, etc.) or other types of mobile devices, or stationary computing devices such as desktop computers or personal computers (PCs). The computing device 1000 can also be a mobile or stationary server.

[0215] The processor 1020 is used to execute computer programs / instructions, which, when executed by the processor, implement the steps of the above-mentioned dialogue processing method or voice dialogue method.

[0216] The above is an illustrative scheme of a computing device according to this embodiment. It should be noted that the technical solution of this computing device belongs to the same concept as the technical solutions of the dialogue processing method and the voice dialogue method described above. For details not described in detail in the technical solution of the computing device, please refer to the description of the technical solutions of the dialogue processing method or the voice dialogue method described above.

[0217] An embodiment of this disclosure also provides a computer-readable storage medium storing a computer program / instructions that, when executed by a processor, implement the steps of the above-described dialogue processing method or voice dialogue method.

[0218] The above is an illustrative scheme of a computer-readable storage medium according to this embodiment. It should be noted that the technical solution of this storage medium belongs to the same concept as the technical solutions of the dialogue processing method and the voice dialogue method described above. For details not described in detail in the technical solution of the storage medium, please refer to the description of the technical solutions of the dialogue processing method or the voice dialogue method described above.

[0219] An embodiment of this disclosure also provides a computer program product, including a computer program / instructions that, when executed by a processor, implement the steps of the above-described dialogue processing method or voice dialogue method.

[0220] The above is an illustrative scheme of a computer program product according to this embodiment. It should be noted that the technical solution of this computer program product belongs to the same concept as the technical solutions of the above-mentioned dialogue processing method and voice dialogue method. For details not described in detail in the technical solution of the computer program product, please refer to the description of the technical solutions of the above-mentioned dialogue processing method or voice dialogue method.

[0221] The foregoing has described specific embodiments of this disclosure. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired results. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0222] The computer instructions include computer program code, which may be in the form of source code, object code, executable file, or certain intermediate forms. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording media, USB flash drive, portable hard drive, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium may be appropriately added or removed according to the requirements of patent practice. For example, in some regions, according to patent practice, computer-readable media may not include electrical carrier signals and telecommunication signals.

[0223] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that the embodiments of this disclosure are not limited to the described order of actions, because according to the embodiments of this disclosure, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to the embodiments of this disclosure.

[0224] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0225] The preferred embodiments disclosed above are merely illustrative of this disclosure. The optional embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the embodiments of this disclosure. These embodiments are selected and specifically described in this disclosure to better explain the principles and practical applications of the embodiments of this disclosure, thereby enabling those skilled in the art to better understand and utilize this disclosure. This disclosure is limited only by the claims and their full scope and equivalents.

Claims

1. A dialogue processing method, comprising: Obtain target dialogue data; Based on the target dialogue data, determine the processing strategy for the target dialogue data; When the processing strategy is determined to be processing target dialogue data based on reference dialogue data, the target dialogue data and the reference dialogue data are input into the target dialogue model to obtain the response result of the target dialogue data, wherein the reference dialogue data is obtained by expanding the dialogue content based on dialogue context information.

2. The method according to claim 1, wherein the target dialogue data includes target dialogue content; The step of determining the processing strategy for the target dialogue data based on the target dialogue data includes: The target dialogue content is subjected to intent recognition to determine the target dialogue intent; If the target dialogue intent is an instruction processing intent, the processing strategy is determined to be either a third-party plugin processing the target dialogue data or a target processing model processing the target dialogue data. If the target dialogue intent is a non-instruction processing intent, the processing strategy is determined to be to process the target dialogue data based on reference dialogue data.

3. The method according to claim 1, wherein the target dialogue data includes the target dialogue state; The step of determining the processing strategy for the target dialogue data based on the target dialogue data includes: If the target dialogue state is silent, the processing strategy is determined to be to process the target dialogue data based on the reference dialogue data.

4. The method according to claim 1, further comprising, before determining that the processing strategy is to process the target dialogue data based on reference dialogue data, inputting the target dialogue data and the reference dialogue data into the target dialogue model to obtain the response result of the target dialogue data, the method includes: Obtain the dialogue generation data; Contextual information is extracted from the generated dialogue data to determine the dialogue context information; The dialogue content is expanded based on the dialogue context information to obtain the reference dialogue data.

5. The method according to claim 4, wherein extracting contextual information from the dialogue generation data to determine the dialogue contextual information includes: The extracted prompt information and the dialogue generation data are input into the text generation model to obtain dialogue context information, wherein the extracted prompt information is used to guide the text generation model to extract context information from the dialogue generation data; The step of expanding the dialogue content based on the dialogue context information to obtain the reference dialogue data includes: The expansion prompt information and the dialogue context information are input into the text generation model to obtain the reference dialogue data, wherein the expansion prompt information is used to guide the text generation model to expand the dialogue content based on the dialogue context information.

6. The method according to claim 4, after expanding the dialogue content based on the dialogue context information to obtain the reference dialogue data, further comprising: The reference dialogue data is parsed to identify at least one dialogue character and the corresponding reference dialogue segment for that dialogue character. Generate character description information for the dialogue character based on the reference dialogue fragment corresponding to the dialogue character; Based on the character description information, the voice configuration information of the dialogue character is generated.

7. The method according to claim 4, after expanding the dialogue content based on the dialogue context information to obtain the reference dialogue data, further comprising: The reference dialogue data is reviewed using dialogue review rules to obtain review results. The dialogue review rules include at least one of grammar review rules, compliance review rules, and copyright review rules. If the review result is a failure, the reference dialogue data is adjusted to obtain adjusted reference dialogue data, wherein the adjusted reference dialogue data conforms to the dialogue review rules.

8. The method according to claim 1, wherein the target dialogue data includes target dialogue speech; Before inputting the target dialogue data and the reference dialogue data into the target dialogue model to obtain the response result of the target dialogue data, when the processing strategy is determined to be processing the target dialogue data based on the reference dialogue data, the method further includes: The target dialogue speech is subjected to speech recognition to obtain the target dialogue text; When the processing strategy is determined to be processing target dialogue data based on reference dialogue data, the step of inputting the target dialogue data and the reference dialogue data into the target dialogue model to obtain the response result of the target dialogue data includes: If the processing strategy is determined to be processing the target dialogue text based on reference dialogue data, the target dialogue text and the reference dialogue data are input into the target dialogue model to obtain the response result of the target dialogue voice.

9. The method according to any one of claims 1 to 8, wherein after determining that the processing strategy is to process the target dialogue data based on reference dialogue data, inputting the target dialogue data and the reference dialogue data into the target dialogue model to obtain the response result of the target dialogue data, the method further comprises: Identify at least one target dialogue role corresponding to the response result; The response result is broadcast aloud based on the voice configuration information of the at least one target dialogue character.

10. The method according to any one of claims 1 to 8, wherein after determining that the processing strategy is to process the target dialogue data based on reference dialogue data, inputting the target dialogue data and the reference dialogue data into the target dialogue model to obtain the response result of the target dialogue data, the method further comprises: Obtain updated target dialogue data, wherein the updated target dialogue data is data obtained by responding to the response result; Return to the step of determining the processing strategy for the target dialogue data based on the target dialogue data, until the dialogue end condition is met and the target response result is obtained.

11. The method according to any one of claims 1 to 8, wherein after determining the processing strategy for the target dialogue data based on the target dialogue data, it further comprises: When the processing strategy involves a third-party plugin processing the target dialogue data, the third-party plugin is invoked to process the target dialogue data, and a response result for the target dialogue data is obtained. When the processing strategy is to process the target dialogue data using the target processing model, the target dialogue data is input into the target processing model to obtain the response result of the target dialogue data.

12. A voice dialogue method, comprising: Acquire voice dialogue data; Based on the voice dialogue data, determine the processing strategy for the voice dialogue data; When the processing strategy is determined to be processing voice dialogue data based on reference dialogue data, the voice dialogue data and the reference dialogue data are input into the target dialogue model to obtain the response result of the voice dialogue data, wherein the reference dialogue data is obtained by expanding the dialogue content based on dialogue context information.

13. A voice dialogue method, comprising: Obtain the client's voice dialogue data; Based on the voice dialogue data, determine the processing strategy for the voice dialogue data; When the processing strategy is determined to be processing voice dialogue data based on reference dialogue data, the voice dialogue data and the reference dialogue data are input into the target dialogue model to obtain the response result of the voice dialogue data, wherein the reference dialogue data is obtained by expanding the dialogue content based on dialogue context information. The response result of the voice dialogue data is broadcast to the client via voice.

14. A computing device, comprising: Memory and processor; The memory is used to store computer programs / instructions, and the processor is used to execute the computer programs / instructions, which, when executed by the processor, implement the steps of the method described in any one of claims 1 to 11, or claim 12 or claim 13.

15. A computer-readable storage medium storing a computer program / instructions that, when executed by a processor, implement the steps of the method according to any one of claims 1 to 11, or claim 12, or claim 13.

16. A computer program product comprising a computer program / instructions that, when executed by a processor, implement the steps of the method as described in any one of claims 1 to 11, or claim 12, or claim 13.

Images