A virtual avatar automatic generation method, system and electronic device

Abstract

The application discloses a method and system for automatically extracting multi-dimensional features from historical interaction data of entities, constructing personality models independent of specific AI models, and driving multi-form outputs to realize virtual avatars. The independent personality model constructed by the application does not depend on specific AI technology, and has the advantages of more accurate and stable by combining multi-dimensional feature extraction based on real historical data. Moreover, the same personality model drives multi-form outputs to ensure the consistency of user experience. In addition, the virtual avatar can learn from new interactions, becoming more and more close to the prototype or evolving as needed. At the same time, the virtual avatar automatic generation scheme of the application can support multiple types of entities from individuals to organizations and from real to fictitious. Based on feature extraction rather than fine-tuning, the creation and update costs are significantly lower than those of traditional schemes.

Description

Technical Field

[0001] This invention relates to the field of digital asset management, and in particular to a method, system and electronic device for automatically generating virtual clones. Background Technology

[0002] With the popularization of artificial intelligence technology and the deepening of digital life, individuals and organizations have accumulated a large amount of interactive data in the digital world. Transforming these assets into a sustainable, sustainable, and usable form has become an important technological challenge.

[0003] Existing technical solutions and their shortcomings: Instant Prompt engineering cannot guarantee personality consistency in long conversations and is not sustainable; model fine-tuning is costly, transfer is difficult, and it depends on specific architectures; retrieval enhancement cannot be generalized and lacks structured modeling; dedicated systems are fragmented, have high maintenance costs, and inconsistent performance; and proprietary platform services cause users to lose data sovereignty and autonomy.

[0004] Therefore, existing virtual avatar technology has inherent flaws: it heavily relies on specific models or service providers, suffers from poor sustainability, has vague personality definitions, insufficient consistency and accuracy, and its forms are independent, unable to be driven uniformly, lacking evolutionary capabilities, and the avatars are static and do not grow. In addition, data sovereignty is unclear, assets are not guaranteed, the scope of application is limited, it is difficult to cover all scenarios, the cost is high, it is difficult to apply on a large scale, and it cannot precisely control specific personality traits. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides a method, system, and electronic device for automatically generating virtual clones.

[0006] To provide a basic understanding of some aspects of the disclosed embodiments, a brief summary is given below. This summary is not intended as a general commentary, nor is it intended to identify key / important components or to describe the scope of protection of these embodiments. Its sole purpose is to present some concepts in a simple form as a prelude to the detailed description that follows.

[0007] Firstly, a method for automatically generating virtual clones is provided, the method comprising: Extract multi-dimensional features from the entity's historical interaction data, and construct a structured personality model that is independent of specific AI models based on the multi-dimensional features. The virtual clone interacts with the interactive object based on the personality model, and the same personality model drives multiple form outputs through a form adapter. The personality model is updated based on the new interaction data.

[0008] The historical interaction data is data that can characterize entity features, including: thinking style, expression habits, knowledge structure, and personality traits.

[0009] The multidimensional features include at least a combination of any three or more of the following dimensions: language expression dimension, personality tendency dimension, knowledge structure dimension, thinking pattern dimension, interaction style dimension, and value dimension.

[0010] The personality model is a set of structured data that records feature vectors of each dimension and the relationships between dimensions.

[0011] The output forms driven by the personality model include: text dialogue, voice dialogue, visual image, video output, and physical robot; and all forms are driven by the same personality model through a form adapter.

[0012] The personality model is updated through continuous evolution; the continuous evolution refers to the virtual avatar continuously learning from new interactions with interactive objects and updating the personality model through feedback, including: interaction data collection, feedback signal collection, feature update, version management, evolution direction control, and abnormal data protection; the version management refers to the personality model generating a new version each time it is updated and retaining historical versions; the evolution direction control refers to enabling the virtual avatar to gradually approach the physical prototype or evolve autonomously in a set direction.

[0013] The entity refers to the subject that serves as the prototype for the virtual avatar, including: natural persons, teams, organizations, brands, fictional characters, and historical figures.

[0014] Secondly, a virtual avatar automatic generation system is provided, including: The data source access module is used to obtain historical interaction data of entities; The feature extraction module is used to extract multi-dimensional features from the historical interaction data of the entity. The personality modeling module is used to construct a structured personality model that is independent of specific AI models based on the multi-dimensional features. The drive output module is used to drive the virtual clone to interact with the interactive object based on the personality model, and the same personality model can drive multiple form outputs through the form adapter. An evolution module is used to update the personality model based on new interaction data.

[0015] Thirdly, embodiments of the present invention provide an electronic device, including: one or more processors; a memory; wherein the memory stores one or more executable programs, and the one or more processors read the executable program code stored in the memory to execute the virtual clone automatic generation method described in any of the first aspects.

[0016] Fourthly, embodiments of the present invention provide a computer-readable storage medium storing one or more programs, which can be executed by one or more processors to perform the virtual clone automatic generation method described in any of the first aspects.

[0017] Beneficial effects: The independent personality model constructed in this application does not rely on specific AI technologies. Combined with multi-dimensional feature extraction based on real historical data, it has the advantages of being more accurate and stable. Moreover, the same personality model drives multiple forms of output, ensuring the consistency of user experience. In addition, the virtual avatar can learn from new interactions, becoming closer to the prototype or evolving on demand. At the same time, the automatic generation scheme of virtual avatars in this application can support multiple types of entities from individuals to organizations and from real to fictional. Based on feature extraction rather than fine-tuning, the creation and updating costs are significantly lower than traditional solutions. Attached Figure Description

[0018] The accompanying drawings are provided for illustrative purposes only, and the proportions and quantities of the components in the drawings may not be consistent with the actual product.

[0019] Figure 1 This is a schematic diagram of an automatic virtual clone generation system according to the present invention; Figure 2 This is a flowchart illustrating the continuous evolution of the present invention; Figure 3 This is a schematic diagram of the architecture of an embodiment of the electronic device of the present invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] This application provides a method for automatically generating virtual avatars, specifically a method for automatically generating virtual avatars based on multi-dimensional feature extraction from multi-source historical interaction data, independent personality model construction, multi-form output driving, and continuous evolution. The method includes: Extract multi-dimensional features from the historical interaction data of the entity.

[0022] Among them, "entity" refers to the subject that serves as the prototype of the virtual avatar, including: natural persons, teams, organizations, brands, fictional characters, and historical figures.

[0023] Virtual avatars refer to AI entities that are driven by personality models and can interact with each other by simulating the characteristics of physical entities.

[0024] Personal virtual avatars are generated based on the historical data of an individual, and can be used for purposes such as digital legacy, expert-enhanced services, interaction with fans of well-known figures, and the continuation of personal digital assets. Teams, composed of multiple individuals, have virtual avatars generated based on the comprehensive historical data of all team members, reflecting the team's collective wisdom and shared style. Typical applications include new member onboarding training, cross-team collaboration consulting, and team knowledge asset management. Organizational virtual avatars, for enterprises and institutions, are generated based on the historical data of the entire organization, representing the organization's overall culture, values, and decision-making style. Typical applications include corporate decision support, organizational culture transmission, and customer consulting services.

[0025] Brand virtual avatars are generated based on all of a brand's external outputs, such as advertisements, customer service records, and social media content, reflecting the brand's tone. Typical applications include customer interaction, marketing campaigns, and brand image maintenance. Fictional character virtual avatars are generated based on characters from fictional works such as literature and film. Typical applications include entertainment, education, and creative assistance. Historical figure virtual avatars are generated based on historical documents and research, with typical applications including history education, cultural heritage preservation, and academic research.

[0026] This invention does not impose any restrictions on the type of entity.

[0027] Historical interaction data refers to data that characterizes the features of an entity, including: thought patterns, expression habits, knowledge structure, and personality traits. Specifically, historical interaction data refers to various types of data generated by an entity in the past that reflect its thought patterns, expression habits, knowledge structure, and personality traits. Its sources are not limited to: dialogue records with AI services, written documents and articles, sent emails, social media content, meeting transcripts, interview recordings, creative works, work outputs, etc. This invention does not restrict the data source; any data that can reflect the characteristics of an entity can be used as input.

[0028] Multidimensional characteristics include at least three or more of the following dimensions: language expression, personality tendencies, knowledge structure, thinking patterns, interaction style, and values.

[0029] This invention automatically extracts features covering multiple dimensions from the historical interaction data of entities, forming a multidimensional feature vector. A feature vector refers to the entity features extracted from historical interaction data and expressed in mathematical form. The number of feature dimensions, their specific definitions, and the specific extraction algorithms can all be flexibly designed and expanded according to the needs of the application scenario.

[0030] The language expression dimension is used to characterize the linguistic style features of an entity. For example, sentence structure preferences (ratio of long and short sentences, use of compound sentences), vocabulary selection (common words, professional terms, spoken or written language), rhetorical preferences (metaphor, rhetorical question, parallelism, etc.), tone features (serious, humorous, friendly, sharp, etc.), rhythm (compact or relaxed expression), punctuation and formatting habits, etc.

[0031] Personality tendency dimensions are used to characterize an entity's personality traits. For example, based on established personality theories (such as the Big Five personality traits, MBTI, etc.) or custom personality dimensions, they characterize an entity's tendencies in areas such as openness, conscientiousness, extraversion, agreeableness, and neuroticism. This dimension can also be extended to more detailed characteristics such as emotional stability, optimism / pessimism tendencies, and decision-making style (rational or emotional).

[0032] The knowledge structure dimension represents an entity's knowledge reserves and areas of expertise. For example, based on topics covered in historical data, the depth of responses, and the professional terminology used, a knowledge graph and ability score for the entity in different fields are built. This dimension determines the depth and accuracy of the virtual avatar's responses on various topics.

[0033] The mindset dimension is used to characterize an entity's thinking style. Examples include problem-solving frameworks such as systematic thinking, critical thinking, and analogical reasoning, as well as problem-solving habits, a tendency to use examples, inductive or deductive biases, and attention to detail.

[0034] Interaction style dimension is used to characterize the behavioral patterns of entities in interactions. Examples include response length preferences, tendency to ask questions proactively, differences in attitudes towards different types of interlocutors, openness to emotional expression, and methods of handling conflict.

[0035] The value dimension is used to characterize and reflect an entity's core value orientation. For example, its stance on important issues, ethical preferences, and priority judgments.

[0036] After extracting multi-dimensional features, a structured personality model that is independent of specific AI models is constructed based on these features.

[0037] A personality model is a set of structured data that records feature vectors in various dimensions and the relationships between dimensions. Specifically, personality models can be serialized and stored as structured data, such as JSON, Protobuf, or other formats, allowing them to be completely serialized into files or data records for easy persistent storage, version management, and cross-system transmission.

[0038] The personality model in this application does not depend on any specific model. That is, the definition, storage, and use of the personality model do not depend on any specific AI model architecture, specific AI service provider, or specific training method. When the personality model is needed, it is connected to a specific AI model through an adaptation layer. Therefore, the personality model can be migrated and used between different AI models. That is, when the underlying AI model changes, such as being upgraded, replaced, or discontinued, the personality model can be reconnected to the new AI model, and the virtual clone remains uninterrupted.

[0039] Multiple personality models can be combined to form new virtual avatars, with combinations such as overlay and weighted fusion. The performance of the virtual avatar can be fine-tuned by adjusting the parameters of each dimension of the personality model.

[0040] As structured data, the ownership and control of personality models belong entirely to the user or organization and are not locked by any service provider. This application adopts a fine-grained control approach, allowing users or authorized parties to precisely control the data scope for feature extraction, the usage scenarios of virtual avatars, permitted interaction objects, and specific parameters of personality performance. For example, the creation of virtual avatars must be based on explicit authorization from the original entity or its legal heirs or authorized persons; users can specify which historical data can be used for personality extraction and which should be excluded; users can specify in which scenarios virtual avatars can be used and in which scenarios they are prohibited; users can specify who can interact with virtual avatars; virtual avatars should clearly identify themselves as AI in all interactions to avoid fraud; virtual avatars should not generate content that violates laws, regulations, or social ethics, even if there is controversial content in the historical data; users can revoke or permanently delete virtual avatars at any time; and inheritance rules for virtual avatars are defined when the original entity no longer exists.

[0041] After constructing the personality model, the virtual avatar interacts with the interactive object based on the personality model, and the personality model drives various forms of output.

[0042] The forms of output driven by personality models include: text dialogue, voice dialogue, visual images, video output, and physical robots.

[0043] All forms are driven by the same personality model through a form adapter. This means that the same personality model can drive the output of multiple different forms without maintaining an independent model for each form, and switching between forms does not affect the personality model itself. The driving principle is that the personality model serves as the core representation, the form adapter converts it into the parameters required for the specific form, and each form engine generates the output.

[0044] Text-based dialogue refers to converting personality model features into constraints and style guidelines for text generation, and then using this interface to generate AI-generated text that matches the personality. Voice-based dialogue involves mapping the language style and emotional features of the personality model to parameters in a speech synthesis system, such as timbre, speech rate, intonation, and pauses, to generate speech that matches the personality. Visual representation refers to two-dimensional or three-dimensional virtual human models, ensuring their appearance, expressions, and movements conform to personality traits. Video output refers to a video synthesis system that generates video content featuring a virtual human character. Physical robots are the dialogue, movement, and expression systems of physical robots, enabling them to exhibit personality traits.

[0045] Multiple forms can be output simultaneously to create a consistent virtual avatar experience. When new interaction forms emerge in the future, such as brain-computer interfaces and augmented reality, they can be supported by adding new form adapters without changing the personality model itself.

[0046] This also includes updating the personality model based on new interaction data. Specifically, such as... Figure 2 As shown, the personality model is updated through continuous evolution. Continuous evolution means that the virtual avatar learns continuously from new interactions with interactive objects and updates the personality model through feedback, including: interaction data collection, feedback signal collection, feature updates, version management, evolution direction control, and abnormal data protection.

[0047] Interaction data acquisition refers to recording the complete dialogue between the virtual avatar and the interactive object as a new source of training data; feedback signal collection refers to obtaining interaction quality signals through explicit feedback, such as likes, ratings, and comments, and implicit feedback, such as dialogue duration and topic switching patterns; feature updating refers to periodically recalculating feature vectors based on new data and integrating them with the existing personality model; version management refers to generating a new version of the personality model with each update and retaining historical versions; evolution direction control refers to making the virtual avatar gradually approach the physical prototype or evolve autonomously in a set direction, such as gradually becoming more patient and concise; abnormal data protection refers to automatically identifying and excluding abnormal data, such as malicious attacks and test data, to prevent the personality model from being contaminated.

[0048] like Figure 1 As shown, the present invention also provides a virtual clone automatic generation system, comprising: The data source access module is used to obtain historical interaction data of entities. It can access historical data from various sources, such as AI dialogue records, document files, emails, social media content, and transcripts of interview recordings.

[0049] The data cleaning module is used to remove noise, standardize the format, desensitize privacy data, and assess the quality of the incoming data.

[0050] The feature extraction module is used to extract multi-dimensional features from the historical interaction data of entities, that is, to extract multi-dimensional features from cleaned data, such as natural language processing, statistical analysis, pre-trained models, etc.

[0051] The personality modeling module is used to construct structured personality models that are independent of specific AI models based on multi-dimensional features.

[0052] The driver output module is used to drive the virtual clone to interact with interactive objects based on the personality model, and the same personality model can drive multiple form outputs through the form adapter.

[0053] The model storage module is used to persistently store personality models and their historical versions; The morphology adaptation module is used to implement adapters for each supported output morphology, converting the personality model into the parameters required for that morphology, and connecting to specific morphology implementation engines, such as text generation, speech synthesis, and virtual character rendering.

[0054] The evolution module is used to update the personality model based on new interaction data.

[0055] It is understood that the structures illustrated in the embodiments of this specification do not constitute a specific limitation on the above-described apparatus. In other embodiments of this specification, there may be more or fewer components than illustrated, or some components may be combined, some components may be separated, or different component arrangements may be used. The illustrated components may be implemented in hardware, software, or a combination of both.

[0056] The information interaction and execution process between the modules in the above-mentioned device are based on the same concept as the method embodiments in this specification, and the specific details can be found in the descriptions in the method embodiments in this specification, so they will not be repeated here.

[0057] Figure 3 This is a schematic diagram of the architecture of an electronic device according to an embodiment of the present invention, which can implement the methods described in any embodiment of the present invention, such as... Figure 3 As shown, as an optional embodiment, the above-mentioned electronic device may include: a housing 41, a processor 42, a memory 43, a circuit board 44, and a power supply circuit 45, wherein the circuit board 44 is disposed inside the space enclosed by the housing 41, and the processor 42 and the memory 43 are disposed on the circuit board 44; the power supply circuit 45 is used to supply power to various circuits or devices of the above-mentioned electronic device; the memory 43 is used to store executable program code; the processor 42 is used to execute any method provided in the embodiments of the present invention by reading the executable program code stored in the memory 43.

[0058] The specific execution process of the above steps by the processor 42, as well as the steps further executed by the processor 42 by running executable program code, will not be described in detail here.

[0059] (1) Mobile communication devices: These devices are characterized by their mobile communication capabilities and are primarily designed to provide voice and data communication. These terminals include smartphones (such as iPhones), multimedia phones, feature phones, and low-end phones.

[0060] (2) Ultra-mobile personal computer devices: These devices fall under the category of personal computers, possessing computing and processing capabilities, and generally also have mobile internet access features. These terminals include PDAs, MIDs, and UMPCs, such as the iPad.

[0061] (3) Portable entertainment devices: These devices can display and play multimedia content. This category includes: audio and video players (such as iPods), handheld game consoles, e-books, as well as smart toys and portable car navigation devices.

[0062] (4) Server: A device that provides computing services. The components of a server include a processor, hard disk, memory, system bus, etc. Servers are similar to general computer architectures, but because they need to provide highly reliable services, they have higher requirements in terms of processing power, stability, reliability, security, scalability, and manageability.

[0063] (5) Other electronic devices with data interaction functions.

[0064] Accordingly, embodiments of this application also provide a computer-readable storage medium storing one or more programs, which can be executed by one or more processors to implement any of the methods provided in the embodiments of the present invention, and thus can also achieve the corresponding technical effects. This has been described in detail above and will not be repeated here.

[0065] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. The storage medium embodiments for providing the program code include floppy disks, hard disks, magneto-optical disks, optical disks (such as CD-ROM, CD-R, CD-RW, DVD-ROM, DVD-RAM, DVD-RW, DVD+RW), magnetic tapes, non-volatile memory cards, and ROMs. Optionally, the program code can be downloaded from a server computer via a communication network. Furthermore, it should be understood that not only can the program code read by the computer be executed, but also the operating system or other components operating on the computer can be instructed based on the program code to perform some or all of the actual operations, thereby realizing the function of any of the embodiments described above.

[0066] It is understood that the program code read from the storage medium is written to the memory set in the expansion board inserted into the computer or to the memory set in the expansion module connected to the computer. Then, based on the instructions of the program code, the CPU or other components installed on the expansion board or expansion module execute some and all of the actual operations, thereby realizing the function of any of the above embodiments.

[0067] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0068] The various embodiments in this specification are described in a related manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0069] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this specification, and are not intended to limit them. Although this specification has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this specification.

Claims

1. A method for automatically generating virtual clones, characterized in that, The method includes: Extract multi-dimensional features from the entity's historical interaction data, and construct a structured personality model that is independent of specific AI models based on the multi-dimensional features. The virtual clone interacts with the interactive object based on the personality model, and the same personality model drives multiple form outputs through a form adapter. The personality model is updated based on the new interaction data.

2. The method for automatically generating virtual clones according to claim 1, characterized in that, The historical interaction data is data that can characterize entity features, including: thinking style, expression habits, knowledge structure, and personality traits.

3. The method for automatically generating virtual clones according to claim 2, characterized in that, The multidimensional features include at least a combination of any three or more of the following dimensions: language expression dimension, personality tendency dimension, knowledge structure dimension, thinking pattern dimension, interaction style dimension, and value dimension.

4. The method for automatically generating virtual clones according to claim 3, characterized in that, The personality model is a set of structured data that records feature vectors of each dimension and the relationships between dimensions.

5. The method for automatically generating virtual clones according to claim 4, characterized in that, The output forms driven by the personality model include: text dialogue, voice dialogue, visual image, video output, and physical robot; and all forms are driven by the same personality model through a form adapter.

6. The method for automatically generating virtual clones according to claim 5, characterized in that, The personality model is updated through continuous evolution. The continuous evolution refers to the virtual avatar continuously learning from new interactions with interactive objects and updating the personality model through feedback, including: interaction data collection, feedback signal collection, feature update, version management, evolution direction control, and abnormal data protection. Version management refers to the generation of a new version of the personality model with each update and the retention of historical versions; evolution direction control refers to making the virtual clone gradually approach the physical prototype or evolve autonomously in a set direction.

7. The method for automatically generating virtual clones according to claim 6, characterized in that, The entity refers to the subject that serves as the prototype for the virtual avatar, including: natural persons, teams, organizations, brands, fictional characters, and historical figures.

8. A virtual clone automatic generation system, characterized in that, include: The data source access module is used to obtain historical interaction data of entities; The feature extraction module is used to extract multi-dimensional features from the historical interaction data of the entity. The personality modeling module is used to construct a structured personality model that is independent of specific AI models based on the multi-dimensional features. The drive output module is used to drive the virtual clone to interact with the interactive object based on the personality model, and the same personality model can drive multiple form outputs through the form adapter. An evolution module is used to update the personality model based on new interaction data.

9. An electronic device, characterized in that, include: One or more processors; Memory; The memory stores one or more executable programs, and the one or more processors read the executable program code stored in the memory to execute the virtual clone automatic generation method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores one or more programs, which can be executed by one or more processors to perform the virtual clone automatic generation method according to any one of claims 1 to 7.

Images