Real-time communication assisting method and system based on digital personality model

Abstract

The invention discloses a real-time communication assisting method and system based on a digital personality model, and the method comprises the steps: receiving scene setting inputted by a user, and obtaining multi-modal data in a communication scene in real time; performing data preprocessing on the multi-modal data, and extracting multi-modal personality features from the preprocessed multi-modal data; aggregating the multi-modal personality features, and inputting the aggregated multi-modal personality features into a pre-constructed digital personality model for preliminary personality assessment to obtain a personality traits prediction result of the user; and matching and generating a real-time communication strategy from a pre-constructed strategy database according to the scene setting and the personality characteristic prediction result. According to the method provided by the invention, the psychological state and personality characteristics of the opposite side can be observed in real time in the communication process, the optimal communication strategy is generated and transmitted in time to reduce potential conflicts, and the problems of single function, serious lagging and lack of personality of the existing communication auxiliary method are solved.

Description

Technical Field

[0001] This application relates to the field of human-computer interaction technology, specifically to a real-time communication assistance method and system based on a digital personality model. Background Technology

[0002] With the rapid development of technologies such as artificial intelligence, natural language processing (NLP), and affective computing, communication assistance technologies have been widely applied in various fields, including customer service, sales, psychological counseling, and enterprise management. Existing technologies mainly focus on identifying users' emotional states through single-modal data such as text, voice, or video, and providing basic feedback based on this.

[0003] However, current mainstream communication assistance technologies still have significant shortcomings and are unable to meet the deep support needs in real-world, complex interpersonal interaction scenarios. These shortcomings are specifically reflected in the following four aspects:

[0004] 1. Limited Functionality: Most existing communication support tools only have emotion recognition capabilities and lack proactive guidance on communication strategies. Even if they can accurately determine whether the other party's emotion is "dissatisfaction" or "hesitation," they often cannot generate targeted response suggestions or optimized communication strategies, forcing users to still rely on their own experience to make decisions.

[0005] 2. Significantly Delayed Response: Currently, most sentiment analysis and communication assessment methods operate offline, generating debriefing reports only after the conversation has ended. This "hindsight" approach fails to provide real-time intervention at critical communication junctures (such as impending customer churn or negotiation deadlock), missing the optimal opportunity to influence the direction of communication and diminishing the practical effectiveness of technology in dynamic interactions.

[0006] 3. Lack of personalized adaptation: General-purpose chatbots or auxiliary plugins are usually trained on mass-market corpora and do not model the personality traits, historical interaction preferences, cultural background, or current situation of specific communication targets. Therefore, the suggestions they provide are often superficial and template-based, and may even exacerbate communication friction due to their "one-size-fits-all" approach, failing to achieve true "personalized solutions."

[0007] 4. Lack of insights into growth and relationship evolution: While existing CRM or communication record systems can store large amounts of interaction logs, they generally lack the ability to fuse and analyze multimodal (e.g., voice tone, facial expressions, text semantics, behavioral patterns) micro-emotional signals, and even more so, they lack long-term learning mechanisms to track the changing trends of implicit indicators such as relationship intimacy and trust. This prevents the system from extracting deep patterns from historical interactions, and from dynamically adjusting communication strategies as relationships develop, making it difficult to support long-term, high-value interpersonal relationship management.

[0008] In summary, current communication assistance technologies have significant shortcomings in terms of functionality, timeliness, personalization, and continuous evolution capabilities. There is an urgent need for a new generation of intelligent communication assistance methods that can integrate multimodal perception, real-time intervention, personalized modeling, and long-term learning capabilities to truly achieve closed-loop communication support that "understands emotions, knows the target audience, provides advice, and grows." Summary of the Invention

[0009] Therefore, this application provides a real-time communication assistance method and system based on a digital personality model to solve the problems of existing communication assistance methods being single-function, severely lagging, and lacking in personalization.

[0010] To achieve the above objectives, this application provides the following technical solution:

[0011] Firstly, a real-time communication assistance method based on a digital personality model includes:

[0012] Step 1: Receive the scenario settings input by the user and acquire multimodal data in the communication scenario in real time; the multimodal data includes video data, audio data and text data;

[0013] Step 2: Perform data preprocessing on the multimodal data and extract multimodal personality features from the preprocessed multimodal data;

[0014] Step 3: Aggregate the multimodal personality features and input them into a pre-built digital personality model for preliminary personality assessment to obtain the user's personality trait prediction results; the digital personality model continuously receives new aggregated multimodal personality features and calibrates and corrects the personality trait prediction results;

[0015] Step 4: Based on the scenario setting and the personality trait prediction results, match and generate real-time communication strategies from the pre-built strategy database.

[0016] Preferably, it also includes: real-time analysis of the matching degree between the other party's behavior pattern and the digital personality model; when emotional fluctuations, resistant body language or conflicting remarks are detected, a conflict warning is triggered in real time, and the generated real-time communication strategy is adjusted in real time according to the conflict warning.

[0017] As an alternative, this also includes: anonymizing each communication data with a registered user and continuously storing it in the corresponding user profile to obtain a long-term historical profile of the registered user.

[0018] Preferably, the method further includes: receiving communication goals input by the user, retrieving the user's long-term historical records based on the communication goals, performing scenario rehearsals, and generating a set of communication key points and risk warnings based on long-term observations based on the scenario rehearsals results.

[0019] As an option, it also includes: automatically calculating the health score of the relationship with each person based on the user's long-term historical records, and comprehensively considering the display of the user's relationship network through a visual graph.

[0020] Preferably, in step 1, the scenario setting includes: the personality label of the communication object, our role and bottom line of demands, the topic and goal of the conversation, and scenario information.

[0021] Preferably, in step 2, computer vision technology, speech signal processing technology, and NLP technology are used to extract multimodal personality features from the preprocessed multimodal data.

[0022] Preferably, in step 4, the strategy database is constructed using a knowledge graph approach.

[0023] Preferably, in step 4, the real-time communication strategy includes: the words to be said, the topics to be avoided, the suggested speaking speed and tone, and the suggested body language.

[0024] Secondly, a real-time communication assistance system based on a digital personality model includes:

[0025] The multimodal data acquisition module is used to receive the scenario settings input by the user and acquire multimodal data in the communication scenario in real time; the multimodal data includes video data, audio data and text data;

[0026] A multimodal personality feature extraction module is used to preprocess the multimodal data and extract multimodal personality features from the preprocessed multimodal data.

[0027] The personality recognition module is used to aggregate the multimodal personality features and input them into a pre-built digital personality model for preliminary personality assessment to obtain the user's personality trait prediction results; the digital personality model continuously receives new aggregated multimodal personality features and calibrates and corrects the personality trait prediction results;

[0028] The communication strategy generation module is used to match and generate real-time communication strategies from a pre-built strategy database based on the scenario settings and the personality trait prediction results.

[0029] Compared with the prior art, this application has at least the following beneficial effects:

[0030] 1. Based on further analysis and research of existing technical problems, this application provides a real-time communication assistance method based on a digital personality model. This method receives user-input scenario settings and acquires multimodal data from the communication scenario in real time. It preprocesses the multimodal data and extracts multimodal personality features from the preprocessed data. These features are then aggregated and input into a pre-constructed digital personality model for preliminary personality assessment, yielding a prediction of the user's personality traits. Based on the scenario settings and personality trait predictions, a real-time communication strategy is generated by matching against a pre-constructed strategy database. The method provided in this application can gain real-time insight into the other party's psychological state and personality traits during communication, instantly generating and delivering the optimal communication strategy to reduce potential conflicts. This solves the problems of existing communication assistance methods being functionally limited, severely lagging, and lacking personalization.

[0031] 2. This application anonymizes each communication with a registered user and continuously stores it in the corresponding user profile, creating a long-term historical archive of the registered users. Based on this long-term historical archive, continuous learning can be conducted to optimize relationship management and communication suggestions. Attached Figure Description

[0032] To more intuitively illustrate the prior art and this application, exemplary drawings are provided below. It should be understood that the specific shapes and structures shown in the drawings should not generally be regarded as limiting conditions for implementing this application; for example, based on the technical concept disclosed in this application and the exemplary drawings, those skilled in the art are able to easily make conventional adjustments or further optimizations to the addition / reduction / classification, specific shapes, positional relationships, connection methods, size ratios, etc. of certain units (components).

[0033] Figure 1 A flowchart illustrating a real-time communication assistance method based on a digital personality model, provided in Embodiment 1 of this application;

[0034] Figure 2 This is a schematic diagram of the strategy database structure provided in Embodiment 1 of this application. Detailed Implementation

[0035] The present application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0036] In the description of this application: unless otherwise stated, "a plurality of" means two or more. The terms "first," "second," "third," etc., in this application are intended to distinguish the objects referred to and do not have any special meaning in terms of technical connotation (e.g., they should not be construed as an emphasis on importance or order). Expressions such as "including," "comprising," and "having" also mean "not limited to" (certain units, components, materials, steps, etc.).

[0037] The terms used in this application, such as "upper," "lower," "left," "right," and "middle," are generally used to indicate the general relative positional relationship for the purpose of intuitive understanding by referring to the accompanying drawings, and are not absolute limitations on the positional relationship in the actual product.

[0038] Example 1

[0039] Please see Figure 1 This embodiment provides a real-time communication assistance method based on a digital personality model, the method including:

[0040] S1: Receive the scenario settings input by the user and acquire multimodal data in the communication scenario in real time; the multimodal data includes video data, audio data and text data;

[0041] Specifically, the user-input scenario settings include: the personality label of the communication partner (such as "decisive and efficient"), the role and bottom line of "our side", the topic and goal of the conversation (such as "project budget negotiation"), and scenario information (such as "one-on-one meeting").

[0042] After users input their scenario settings, they can engage in human-computer interaction. During this interaction, multimodal data from the communication scenario can be acquired in real time and comprehensively, including video data, audio data, and text data. Specifically, video data can capture facial expressions, eye contact, micro-expressions, and body language (such as gestures and posture); audio data can capture the semantic content of speech, tone of voice, volume, speech rate, and pause frequency; and text data includes received chat logs, meeting minutes, and other textual information.

[0043] S2: Perform data preprocessing on the multimodal data and extract multimodal personality features from the preprocessed multimodal data;

[0044] Specifically, this step employs computer vision, speech signal processing, and natural language processing (NLP) techniques to extract multimodal personality features from preprocessed multimodal data. This can be described by the following formula:

[0045] Assumptions: Multimodal data input:

[0046] Video data is represented as V: facial expressions and body language of the tested subject;

[0047] Audio data is represented as A: spoken recording of the test subject;

[0048] Text data is represented as T: the transcript of the test subject's speech, i.e.:

[0049] T = AudioToText(A);

[0050] It can also include other information O: other information related to the conversation.

[0051] Using three AI models (i.e., computer vision model, speech signal processing model, and NLP model) v、 f a f t Automatically extract canonical mapping-related features (i.e., personality traits) of personality traits from multimodal input {V, A, T}:

[0052]

[0053] in, This includes, but is not limited to, facial expressions (such as eye contact, head and neck micro-movements, facial expressions, and micro-expressions) and body language (posture and body orientation, gestures, and adaptive movements); This includes, but is not limited to, speech rate, intonation, intensity, rhythm, and extra-vocal signals; This includes, but is not limited to, context, semantics, lexical structure, syntax, structure, logic, and emotion; O includes other information related to the dialogue, such as the scene.

[0054] S3: Aggregate multimodal personality traits and input them into a pre-built digital personality model for preliminary personality assessment to obtain the user's personality trait prediction results; the digital personality model continuously receives new aggregated multimodal personality traits and calibrates and corrects the personality trait prediction results;

[0055] Specifically, this step aggregates multimodal personality traits and inputs them into a pre-built digital personality model to quickly conduct a preliminary personality assessment of the communication subject (e.g., scores based on the five dimensions of the Big Five personality theory), obtaining personality trait prediction results. This can be described by the formula:

[0056] An AI model After aggregating multimodal relevant features, the aggregation results are used for prediction and classification to obtain the user personality trait prediction result P:

[0057]

[0058] in, This includes, but is not limited to, specific personality traits, and can integrate multiple personality assessment systems, such as agreeableness in the Big Five personality traits and the Enneagram.

[0059] Build This can be achieved through the following methods: 1) pre-training; 2) fine-tuning training; 3) context engineering or cue word engineering.

[0060] In this embodiment, the digital personality model does not make a one-time judgment, but continuously receives new multimodal data as communication progresses, and calibrates and corrects the preliminary assessment results to form a dynamic and evolving personality profile.

[0061] Specifically, given that the actual testing process uses sampling to collect the user's V, A, T metrics, a single collection may be affected by various factors inherent to the tested object and external interference, leading to some deviation in the test results. Therefore, in engineering implementation, the tested object can be tested multiple times, and the results of these multiple tests, combined with other information, can be used to call the AI ​​model. Make a comprehensive judgment to obtain the user's personality characteristics. .

[0062]

[0063] A simple implementation example is to statistically analyze different personality traits. The frequency of occurrence is taken as the highest frequency; however, in practical engineering implementation, a comprehensive consideration can be taken into account. To mitigate the influence of external interference factors and eliminate biases in personality prediction.

[0064] The above method requires historical prediction input information. and results Record storage.

[0065] S4: Based on the scenario setting and personality trait prediction results, match and generate real-time communication strategies from a pre-built strategy database.

[0066] Specifically, in this embodiment, the strategy database is actually a goal-oriented strategy database. This strategy database associates different personality traits and different communication goals with optimal communication techniques and behavioral suggestions (for example, for cautious personalities, it is recommended to use more data support; for impatient personalities, it is recommended to state the conclusion first).

[0067] In this embodiment, the policy database can be constructed using a knowledge graph approach (e.g., Figure 2 (As shown): Taking the Big Five personality traits as an example, the actual chart is not limited to the following elements:

[0068] Persona:

[0069] Dimensions: Big Five (openness, conscientiousness, extraversion, agreeableness, emotional stability / neuroticism), risk attitude (avoidance / neutrality / preference), time orientation (short / medium / long term), motivation (achievement / affinity / power), conflict style (competitive / cooperative / compromise / avoidance / accommodation), decision-making style (intuitive / analytical), and pace preference (fast / medium / slow).

[0070] Attributes: intensity score, source (self-assessment / other-assessment / behavioral inference), and confidence level.

[0071] Goal (Communication Objective):

[0072] Examples include: reaching an agreement, persuading adoption, joint decision-making, risk consensus, relationship repair, information clarification, and finalizing commitments.

[0073] Phases: Opening / Exploration / Solution Development / Concession Exchange / Finalization / Follow-up Commitments

[0074] Context:

[0075] Variables: closeness of relationship, power / status equality, culture / language, time pressure, magnitude of interests, focus of dispute, risk uncertainty, historical trust level, and on-site / remote.

[0076] Strategy (Blueprint for Strategy)

[0077] Description: Purpose, applicable prerequisites, contraindications, and key mechanisms.

[0078] Structure: It consists of a series of steps linked together by several tactic, with decision branches and triggering conditions.

[0079] Tactics (specific tactics / rhetoric module):

[0080] Language types: Questioning, restatement / mirroring, risk acknowledgment and quantification, choosing an architecture, phased commitment, alternative solutions, framework transition, summary alignment, and setting checkpoints.

[0081] Fields: Applicable traits, applicable goals and stages, text script template (with parameter placeholders), nonverbal suggestions (speech rate / pauses / gestures), triggering conditions, other party signals, transfer conditions, and risk of adverse effects.

[0082] Outcome (Results and Feedback):

[0083] Record metrics for using a certain strategy in a certain context, and use these metrics to feed back into weights and recommendation scores.

[0084] Therefore, this step can match and generate one or more sets of real-time suggestions from the strategy database based on the scenario setting and personality trait prediction results. The content includes: the words to say, the topics to avoid, the suggested speaking speed and tone, and the suggested body language.

[0085] This embodiment provides a real-time communication assistance method based on a digital personality model, which further includes: analyzing the matching degree between the other party's behavior pattern and the digital personality model in real time; triggering a conflict warning in real time when emotional fluctuations, resistant body language, or conflicting remarks are detected; and adjusting the generated real-time communication strategy in real time according to the conflict warning.

[0086] This embodiment provides a real-time communication assistance method based on a digital personality model, which further includes: anonymizing each communication data with a registered user and continuously storing it in the corresponding person's profile to obtain a long-term historical profile of the registered user.

[0087] Specifically, this embodiment maintains a dynamic profile (i.e., a long-term historical profile) for each registered individual, recording key information for each communication: the scenario in which the conversation took place (such as a formal meeting or a private conversation), the communication goals set by the user, the strategies recommended by AI, the actual actions taken by the user, and feedback on the effectiveness of the communication (whether the goals were achieved and the satisfaction of both parties). All sensitive information is stored after being anonymized.

[0088] Based on this long-term historical archive, this implementation can automatically calculate the health score of relationships with each person, comprehensively considering dimensions such as communication frequency, goal achievement rate, satisfaction, and conflict resolution efficiency. A visual graph displays the user's relationship network: healthy relationships are shown as green nodes, while relationships requiring attention are marked in yellow or red, helping users to clearly understand their interpersonal status at a glance.

[0089] This embodiment can also automatically analyze historical conflict records based on long-term historical archives to identify high-frequency conflict points. For example, it may find that "there is an 80% conflict rate with someone on budget topics" or "Monday morning meetings are prone to disagreements." Based on these patterns, the method will provide preventative advice before similar scenarios occur, such as "suggest sending detailed information in advance" or "avoid discussing sensitive topics when time is tight."

[0090] This embodiment can also track the actual effects of different strategies on different individuals based on long-term historical archives, generating personalized strategy effectiveness reports. For example, it may find that "active listening" has a 90% success rate for a particular person, while "data persuasion" has a mediocre effect, thereby continuously optimizing the accuracy of AI recommendations.

[0091] As data accumulates, this embodiment can identify the stages of relationship development (strangeness period, stable period, etc.) and proactively remind users to maintain the relationship when its health declines, thus truly achieving scientific interpersonal relationship management.

[0092] Therefore, this embodiment will anonymize each communication with the documented person, including multimodal data, personality model evolution, and strategy adoption results, and continuously store them in the corresponding person's profile. After long-term accumulation, a "relationship health map" between the user and different people can be constructed, and high-frequency conflict points can be identified.

[0093] This embodiment provides a real-time communication assistance method based on a digital personality model, which further includes: receiving the communication goal input by the user, retrieving the user's long-term historical file according to the communication goal, performing scenario rehearsal, and generating a set of communication key points and risk warnings based on long-term observation based on the scenario rehearsal results.

[0094] Specifically, this embodiment allows users to "rehearse" upcoming conversations before important communications take place. A simple description is sufficient: "Tomorrow I need to discuss next quarter's budget adjustments with Manager Zhang. I'm worried he might question the data sources, and I hope the plan can be approved smoothly."

[0095] This embodiment immediately retrieves the personality profiles and historical communication records of both parties for intelligent analysis. Based on the other party's personality traits (such as "data-driven decision-maker" and "sensitive to time pressure"), combined with the parties' past interaction patterns (such as "three disagreements on budget topics" and "85% success rate of using data visualization strategies"), a customized communication plan is generated.

[0096] The output communication plan includes, but is not limited to:

[0097] Strategy Recommendation: Recommend the communication tactics best suited to the other party's personality and the current objective;

[0098] Key points of the speech: specific opening remarks, key argumentative logic, and wording for responding to questions;

[0099] Risk warning: Based on historical conflict points, highlight potential pitfalls (e.g., "Avoid scheduling on Monday mornings" and "Prepare alternative solutions to address the issue").

[0100] Psychological preparation: anticipate the other party's possible reactions and emotional changes.

[0101] This feature is like having a private consultant who understands the history of the relationship between the two parties, helping users to prepare thoroughly before key conversations, increasing the success rate of communication, and reducing unnecessary friction.

[0102] It should be noted that the optimal implementation scheme for the real-time communication assistance method based on a digital personality model provided in this embodiment is AI glasses (carrying a camera and microphone to collect dialogue information and feeding back communication strategies to the user through visual / auditory means); and wearable brain-computer interface devices (collecting dialogue information through cameras, microphones, etc., and feeding back communication strategies through brain-computer interfaces).

[0103] This embodiment provides a real-time communication assistance method based on a digital personality model. This method is an integrated, closed-loop, and real-time intelligent communication assistance method. It can gain real-time insight into the other party's psychological state and personality traits during the communication process, generate and deliver the optimal communication strategy in an instant to reduce potential conflicts, assist users in conducting efficient and low-conflict communication, and continuously optimize relationship management and communication suggestions through long-term learning, ultimately facilitating the achievement of the user's communication goals.

[0104] Example 2

[0105] This embodiment provides a real-time communication assistance system based on a digital personality model, including:

[0106] The multimodal data acquisition module is used to receive the scenario settings input by the user and acquire multimodal data in the communication scenario in real time; the multimodal data includes video data, audio data and text data;

[0107] A multimodal personality feature extraction module is used to preprocess the multimodal data and extract multimodal personality features from the preprocessed multimodal data.

[0108] The personality recognition module is used to aggregate the multimodal personality features and input them into a pre-built digital personality model for preliminary personality assessment to obtain the user's personality trait prediction results; the digital personality model continuously receives new aggregated multimodal personality features and calibrates and corrects the personality trait prediction results;

[0109] The communication strategy generation module is used to match and generate real-time communication strategies from a pre-built strategy database based on the scenario settings and the personality trait prediction results.

[0110] For details on the specific implementation of each module in a real-time communication assistance system based on a digital personality model, please refer to the above description of the limitations of a real-time communication assistance method based on a digital personality model, which will not be repeated here.

[0111] The technical features of the above embodiments can be combined in any way (as long as there is no contradiction in the combination of these technical features). For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described; these embodiments not explicitly written should also be considered to be within the scope of this specification.

Claims

1. A real-time communication assistance method based on a digital personality model, characterized in that, include: Step 1: Receive the scenario settings input by the user and acquire multimodal data in the communication scenario in real time; the multimodal data includes video data, audio data and text data; Step 2: Perform data preprocessing on the multimodal data and extract multimodal personality features from the preprocessed multimodal data; Step 3: Aggregate the multimodal personality features and input them into a pre-built digital personality model for preliminary personality assessment to obtain the user's personality trait prediction results; the digital personality model continuously receives new aggregated multimodal personality features and calibrates and corrects the personality trait prediction results; Step 4: Based on the scenario setting and the personality trait prediction results, match and generate real-time communication strategies from the pre-built strategy database.

2. The real-time communication assistance method based on a digital personality model according to claim 1, characterized in that, Also includes: The system analyzes the matching degree between the other party's behavior patterns and the digital personality model in real time. When emotional fluctuations, resistant body language, or conflicting remarks are detected, a conflict warning is triggered in real time, and the generated real-time communication strategy is adjusted in real time based on the conflict warning.

3. The real-time communication assistance method based on a digital personality model according to claim 1, characterized in that, Also includes: After anonymizing each communication with a registered user, the data is continuously stored in the corresponding user profile to obtain a long-term historical archive of the registered user.

4. The real-time communication assistance method based on a digital personality model according to claim 3, characterized in that, Also includes: The system receives the communication goals input by the user, retrieves the user's long-term historical records based on the communication goals, performs scenario rehearsals, and generates a set of communication key points and risk warnings based on long-term observations based on the scenario rehearsal results.

5. The real-time communication assistance method based on a digital personality model according to claim 3, characterized in that, Also includes: The system automatically calculates the health score of the user's relationship with each person based on the user's long-term historical records, and comprehensively considers these scores to display the user's relationship network through a visual graph.

6. The real-time communication assistance method based on a digital personality model according to claim 1, characterized in that, In step 1, the scenario setting includes: the personality label of the communication object, our role and bottom line of demands, the topic and goal of the conversation, and scenario information.

7. The real-time communication assistance method based on a digital personality model according to claim 1, characterized in that, In step 2, computer vision technology, speech signal processing technology, and NLP technology are used to extract multimodal personality features from the preprocessed multimodal data.

8. The real-time communication assistance method based on a digital personality model according to claim 1, characterized in that, In step 4, the policy database is constructed using a knowledge graph approach.

9. The real-time communication assistance method based on a digital personality model according to claim 1, characterized in that, In step 4, the real-time communication strategy includes: the words to be said, the topics to be avoided, the suggested speaking speed and tone, and the suggested body language.

10. A real-time communication assistance system based on a digital personality model, characterized in that, include: The multimodal data acquisition module is used to receive the scenario settings input by the user and acquire multimodal data in the communication scenario in real time; the multimodal data includes video data, audio data and text data; A multimodal personality feature extraction module is used to preprocess the multimodal data and extract multimodal personality features from the preprocessed multimodal data. The personality recognition module is used to aggregate the multimodal personality features and input them into a pre-built digital personality model for preliminary personality assessment to obtain the user's personality trait prediction results; the digital personality model continuously receives new aggregated multimodal personality features and calibrates and corrects the personality trait prediction results; The communication strategy generation module is used to match and generate real-time communication strategies from a pre-built strategy database based on the scenario settings and the personality trait prediction results.

Images