Child aptitude area excavation method and companion robot

Abstract

The application provides a child's strong field excavation method and a companion robot, and relates to the technical field of child education. The target child is scored in multiple preset ability dimensions through the obtained interactive record data, so that the quantitative characterization of the target child in each preset ability dimension can be realized. The target strong field is determined by using each current dimension score, valuable educational insights are extracted from the quantitative data, and the whole process from data collection to strong field discovery is completed. The method can automatically and objectively excavate the target strong field of the target child from the daily interaction between the target child and the companion robot, without manual intervention, so that the excavation efficiency and accuracy of the target strong field can be improved, and an explicit direction is provided for subsequent personalized education guidance. The method solves the problems of low efficiency and poor accuracy caused by manual observation or subjective judgment in the prior art, and has the beneficial effects of objectivity, efficiency and quantification.

Description

Technical Field

[0001] This invention relates to the field of children's education technology, and in particular to a method for identifying children's strengths and a companion robot. Background Technology

[0002] As educational reforms continue to deepen, parents are increasingly emphasizing the discovery of children's strengths, hoping to better identify these areas and provide focused training for their children's growth.

[0003] However, discovering a child's strengths requires long-term companionship and meticulous observation, which is quite difficult for most busy parents. Summary of the Invention

[0004] This invention provides a method for identifying children's areas of expertise and a companion robot, thereby addressing the deficiencies in related technologies.

[0005] This invention provides a method for identifying children's areas of expertise, comprising: Acquire interaction records between the target child and the companion robot within the current period; Based on the interaction record data, the target child is scored on each preset ability dimension to obtain the current dimension score of the target child on each preset ability dimension; Based on the scores of each current dimension, at least one target area of ​​expertise is determined from each of the preset capability dimensions.

[0006] According to the present invention, a method for identifying a child's areas of expertise is provided, wherein each current dimension score corresponds to a scoring reason; the step of determining at least one target area of ​​expertise from each preset ability dimension based on each current dimension score includes: Based on the areas of expertise of each of the aforementioned targets, generate parent feedback and suggestions; The robot pushes the areas of expertise of each target, the scoring reasons for each area of ​​expertise, and the feedback and suggestions from the parents to the parent devices associated with the companion robot.

[0007] According to the present invention, a method for identifying a child's areas of expertise includes interactive record data comprising the number of interactions; the step of scoring the target child on each preset ability dimension based on the interactive record data to obtain the target child's current dimension score on each preset ability dimension includes: Based on the interaction record data, the target child is scored on each preset ability dimension to obtain the target child's current initial score on each preset ability dimension; Determine the statistical data of the first dimension scores of the target children in each of the preset ability dimensions within multiple historical periods; Based on the score statistics of each first dimension and the number of interactions, calculate the confidence weight of each preset capability dimension; Based on the confidence weights of each preset capability dimension and the initial scores, the scores of each current dimension are determined.

[0008] According to a method for identifying children's areas of expertise provided by the present invention, the step of generating parental feedback suggestions based on each of the target areas of expertise includes: For any of the target areas of expertise, if the confidence weight corresponding to any target area of ​​expertise is greater than or equal to the confidence weight threshold, then the parent feedback suggestion corresponding to the advanced challenge task is generated. If the confidence weight corresponding to any of the target's areas of expertise is less than the confidence weight threshold, then the parent feedback suggestion corresponding to the reinforcement task is generated.

[0009] According to a method for identifying a child's areas of expertise provided by the present invention, determining at least one target area of ​​expertise from each of the preset ability dimensions based on the scores of each current dimension includes: The current dimension score is compared with the corresponding dimension score threshold to obtain the comparison results; Based on the comparison results, the target's areas of expertise are determined from the preset capability dimensions. The score threshold for each dimension is determined based on the weight coefficient of each preset ability dimension, and each weight coefficient is used to characterize the degree of importance that the target parents of the target child attach to each preset ability dimension.

[0010] According to a method for identifying children's areas of expertise provided by the present invention, the steps for determining the score thresholds for each dimension include: Obtain a norm database, which is used to store statistical data on the second dimension scores of peers who are of the same age, gender and region as the target child in each of the preset ability dimensions; Based on the score statistics of each of the second dimensions and the weight coefficients of each dimension, the score threshold of each dimension is determined.

[0011] According to a method for identifying children's areas of expertise provided by the present invention, the step of pushing each target area of ​​expertise, the scoring reasons corresponding to each target area of ​​expertise, and the parent's feedback suggestions to the parent device associated with the companion robot includes: Receive feedback operation data sent by the parent device; Based on the feedback operation data, the weight coefficients of each target's area of ​​expertise are updated.

[0012] According to a method for identifying a child's areas of expertise provided by the present invention, the interaction record data includes the number of interactions and target interaction data; the step of acquiring the interaction record data between the target child and the companion robot within the current period includes: Acquire the initial interaction data of each interaction collected by the sensors mounted on the companion robot; Based on the initial interaction data, the emotional state of the target child is identified; Based on the emotional state, the initial interaction data with negative emotional states are removed from each of the initial interaction data to obtain the target interaction data.

[0013] The present invention also provides a device for children to explore their areas of expertise, comprising: The acquisition module is used to acquire interaction record data between the target child and the companion robot within the current period; The scoring module is used to score the target child on each preset ability dimension based on the interaction record data, and obtain the current dimension score of the target child on each preset ability dimension; The determination module is used to determine at least one target area of ​​expertise from each of the preset capability dimensions based on the scores of each of the current dimensions.

[0014] The present invention also provides a companion robot, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the child's area of ​​expertise discovery method as described above.

[0015] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for identifying children's areas of expertise as described above.

[0016] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the method for identifying children's areas of expertise as described above.

[0017] The present invention provides a method for identifying children's areas of expertise and a companion robot. By acquiring interaction record data, the method scores target children across multiple preset ability dimensions, enabling quantitative representation of each child's abilities in these dimensions. The scores for each dimension are used to determine the target child's areas of expertise, extracting valuable educational insights from the quantitative data, thus completing the entire process from data collection to area of ​​expertise discovery. This method can automatically and objectively identify a child's target areas of expertise from their daily interactions with the companion robot, without human intervention. It improves the efficiency and accuracy of identifying these areas of expertise and provides a clear direction for subsequent personalized educational guidance. This solves the problems of low efficiency and poor accuracy caused by reliance on manual observation or subjective judgment in existing technologies, offering objective, efficient, and quantifiable benefits. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in this invention or related technologies, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating the method for identifying children's areas of expertise provided by this invention.

[0020] Figure 2 This is a schematic diagram of the structure of the child's area of ​​expertise exploration device provided by the present invention.

[0021] Figure 3 This is a structural schematic diagram of the companion robot provided by the present invention. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0023] Figure 1 This is a flowchart illustrating a method for identifying children's areas of expertise provided in an embodiment of the present invention, such as... Figure 1 As shown, the method includes: S1, acquire the interaction record data between the target child and the companion robot within the current period; S2, based on the interaction record data, score the target child on each preset ability dimension to obtain the current dimension score of the target child on each preset ability dimension; S3, based on the scores of each current dimension, determine at least one target area of ​​expertise from each of the preset capability dimensions.

[0024] Specifically, the child's area of ​​expertise discovery method provided in this embodiment of the invention is executed by a child's area of ​​expertise discovery device. This device can be configured within a companion robot, a server or cloud computing platform that communicates with the companion robot. The companion robot can be a child companion robot, or other types of companion robots, such as pet robots or early childhood education robots. The companion robot is an intelligent interactive device that provides emotional companionship, educational guidance, and safety monitoring through functions such as voice dialogue, facial expression interaction, motion sensing, and content playback. The companion robot can integrate artificial intelligence (AI) technologies such as voice recognition, sensors, and cameras, enabling it to tell stories, teach knowledge, remind users of schedules, and provide simple emotional responses. The companion robot can be connected to parental devices, enabling remote monitoring and interaction through interaction with these devices.

[0025] First, step S1 is executed to obtain the interaction record data between the target child and the companion robot within the current period. Here, the current period refers to a pre-set time interval, such as a day, a week, or a month, which is used to limit the time range for collecting interaction record data, so that the scores of each current dimension can reflect the performance of the target child within the current period, avoiding inaccurate scoring due to a time span that is too long or too short.

[0026] Interaction recording data refers to various types of data generated during the process of the target child interacting with the companion robot through voice dialogue, touch interaction, and game interaction. This includes interaction data and the number of interactions. Interaction data can be multimodal, including but not limited to voice data, visual data, and behavioral data. Voice data can be voice data of the target child collected through a microphone, which can be used to determine the target child's voice commands or dialogue content, and also to identify the target child's emotional state. Visual data can be image data of the target child captured by a camera, which can be used to determine the target child's facial expressions and body movements. Behavioral data can include the target child's touch behavior collected through a touch sensor. In addition, interaction recording data may also include dialogue turns, question frequency, task completion time, and content selection preferences.

[0027] The number of interactions refers to the total number of effective interactions between the target child and the companion robot within the current period. It can be directly obtained from the interaction record data and reflects the target child's activity level and data richness within the current period. The number of interactions directly affects the reliability of the scores for each current dimension. More interactions usually mean a larger amount of data on which the scoring is based, and more stable the scores for each current dimension.

[0028] Interaction data can be recorded in real time and stored in the companion robot's local memory, or retrieved from the companion robot's server at the end of each cycle. Retrieval can be done by reading directly from local memory or downloading from the server's database.

[0029] Then, step S2 is executed, using interactive recorded data to score the target child on each preset ability dimension, and obtain the target child's current dimension score on each preset ability dimension.

[0030] Among them, the preset ability dimensions refer to the pre-set classification indicators used to measure children's abilities in different aspects. For example, they may include interest exploration ability, logical thinking ability, emotional and social ability, creativity, language expression ability, etc. The types and number of preset ability dimensions can be adjusted according to the actual application scenario and the child's developmental stage.

[0031] Each target child has a corresponding current dimension score for each preset ability dimension. This current dimension score is a quantitative assessment of the target child's performance in the corresponding preset ability dimension, usually expressed numerically. For example, the scoring criteria can include 0 to 3 points or 0 to 100 points. The higher the current dimension score, the stronger the ability in the corresponding preset ability dimension. Taking 0 to 3 points as an example, 0 points represent a pure statement with no cognitive or emotional value, such as the target child's interaction record data "I went to the zoo"; 1 point represents expressing facts combined with slight interest, such as the target child's interaction record data "I like elephants"; 2 points represent reasons, details, or emotions, such as the target child's interaction record data "I like elephants because they have long trunks and can spray water"; 3 points represent demonstrating active thinking, creativity, and logical reasoning, such as the target child's interaction record data "If an elephant's trunk becomes shorter, can it still spray water?".

[0032] The scoring process can be implemented using various strategies, such as leveraging pre-trained machine learning models, rule engines, or knowledge graphs. For the language expression ability dimension, the companion robot can analyze features such as the target child's vocabulary richness, sentence complexity, and the number of times they initiate conversations in their interaction records, and calculate the dimensional score for that dimension using a machine learning model. For the logical thinking ability dimension, the companion robot can analyze features such as the target child's reasoning paths and causal expressions when answering questions in their interaction records, and calculate the current dimensional score for that dimension using a machine learning model.

[0033] This step can transform the interaction record data into a structured current dimension score, which allows the target child's various preset abilities to be quantitatively presented, providing comparable and analyzable numerical basis for subsequent identification of areas of expertise.

[0034] Finally, step S3 is executed to determine at least one target area of ​​expertise from each preset ability dimension using the current dimension scores. A target area of ​​expertise refers to an area where the target child excels in the corresponding preset ability dimension, with a current dimension score significantly higher than the target child's other preset ability dimensions or higher than the age-appropriate norm. The target area of ​​expertise can be determined by setting a target score threshold; for example, if the scoring standard is 0 to 3 points, the target score threshold could be set to 2 points, and the preset ability dimensions with a current dimension score greater than or equal to 2 points would be considered the target area of ​​expertise. Alternatively, a dimension score threshold can be set for each preset ability dimension, and the current dimension score can be compared with the corresponding preset ability dimension's dimension score threshold. The preset ability dimensions with a current dimension score greater than or equal to the corresponding preset ability dimension's dimension score threshold would be considered the target area of ​​expertise.

[0035] Furthermore, a relative comparison can be used, selecting the top one or more preset ability dimensions with the highest current scores as the target area of ​​strength. For example, if the target child's current score in the language expression ability dimension is 3 points, the current score in the logical thinking ability dimension is 2 points, and the current score in the creativity dimension is 1 point, and the target score threshold is 2 points, then language expression ability and logical thinking ability are identified as the target areas of strength. This step completes the transformation from quantitative scoring to ability strength identification, directly outputting the target child's most prominent ability direction, providing parents or educators with a clear focus.

[0036] The method for identifying children's areas of expertise provided in this invention scores target children across multiple preset ability dimensions using acquired interaction record data, enabling quantitative representation of the target child's abilities in each dimension. The method then uses the scores for each dimension to determine the target child's areas of expertise, extracting valuable educational insights from the quantitative data and completing the entire process from data collection to area of ​​expertise discovery. This method can automatically and objectively identify the target child's areas of expertise from their daily interactions with the companion robot, without human intervention. It improves the efficiency and accuracy of identifying areas of expertise and provides a clear direction for subsequent personalized educational guidance. This solves the problems of low efficiency and poor accuracy caused by reliance on manual observation or subjective judgment in existing technologies, offering objective, efficient, and quantifiable benefits.

[0037] Based on the above embodiments, each current dimension score corresponds to a scoring reason; the step of determining at least one target area of ​​expertise from each preset ability dimension based on each current dimension score then includes: Based on the areas of expertise of each of the aforementioned targets, generate parent feedback and suggestions; The robot pushes the areas of expertise of each target, the scoring reasons for each area of ​​expertise, and the feedback and suggestions from the parents to the parent devices associated with the companion robot.

[0038] Specifically, while scoring the target child on each preset ability dimension to obtain the dimension score, a scoring reason corresponding to the current dimension score is also generated. This scoring reason provides a specific explanation or basis for the corresponding current dimension score, explaining why the target child obtained the corresponding current dimension score.

[0039] The scoring reasons can be generated from key behavioral features extracted from the interaction record data. For example, in the dimension of language expression ability, the scoring reason could be "the target child actively told three complete stories in the current period and used more than twenty different adjectives"; in the dimension of logical thinking ability, the scoring reason could be "the target child successfully solved five reasoning questions in the current period and was able to clearly explain the thought process of each step".

[0040] The generation of scoring reasons can be achieved through template filling, that is, multiple scoring reason templates are pre-set for each preset ability dimension, and the appropriate scoring reason templates are selected or combined according to the actual interactive performance of the target child; or personalized scoring reasons can be automatically generated by a natural language generation model based on interaction record data. The existence of scoring reasons makes the current dimension score no longer an isolated number, but has traceable and understandable evidence, thereby enhancing the credibility and transparency of the current dimension score.

[0041] After identifying each child's strengths, parent feedback suggestions can be generated based on these areas. Parent feedback suggestions are specific action recommendations or guidance strategies provided to parents regarding the child's performance in their target strengths. These suggestions can include methods of encouragement, recommendations for extension activities, and topics for parent-child interaction. They should avoid vague praise, be simple and friendly, and avoid using technical jargon. For example, if the target strength is language expression, parent feedback suggestions could be, "Parents are advised to spend ten minutes each day having a themed conversation with their child, encouraging them to describe interesting things that happened that day." If the target strength is logical thinking, parent feedback suggestions could be, "Try playing Sudoku or puzzles with your child, guiding them to explain their reasoning process."

[0042] Parent feedback suggestions can be generated based on a preset suggestion library, matching the type of the target's area of ​​expertise and the current dimension score; or they can be personalized by combining information such as the target child's age and gender, without specific limitations here.

[0043] Afterwards, the areas of expertise of each target child, the corresponding scoring reasons for each area of ​​expertise, and parental feedback and suggestions will be pushed to the parent's device associated with the companion robot. The parent's device refers to a smart terminal that has been linked to the companion robot, such as the target child's parent's smartphone, tablet, or smartwatch.

[0044] Push notifications can send data to applications or WeChat mini-programs installed on the parent's device via wireless network, presenting it in the form of notification messages, graphic reports, or voice broadcasts.

[0045] The push notification can be triggered automatically at the end of each current cycle, or it can be sent when a parent makes a request.

[0046] The content pushed to parents includes information on their child's strengths, allowing them to quickly understand their child's advantages; the scoring rationale helps them understand the basis for the current score; and parental feedback provides concrete action guidance. Together, these three elements form a complete and actionable family education reference report.

[0047] In this embodiment of the invention, the steps of scoring the target child on each preset ability dimension based on interaction record data, determining at least one target area of ​​expertise from each preset ability dimension based on the current dimension score, and generating parent feedback suggestions based on each target area of ​​expertise can all be achieved through a forward inference of a machine learning model. That is, the interaction record data between the target child and the companion robot in the current period is input into the machine learning model, and the machine learning model can output the target child's current dimension score on each preset ability dimension, the reason for the score, the target area of ​​expertise, and parent feedback suggestions. Here, the output of the machine learning model can be displayed in a table.

[0048] In this embodiment of the invention, the scoring reasons can provide a traceable explanation for the current dimension score, enhancing its credibility; parental feedback and suggestions can transform the target child's area of ​​expertise into an actionable guide for the target parents, improving the practical application value of identifying the child's strengths; and pushing the information to the parent's device ensures that the information is delivered to the target parents in a timely and convenient manner, allowing them to adjust their educational strategies accordingly. This solution not only identifies the target child's area of ​​expertise but also presents the scoring reasons and parental feedback and suggestions to the target parents in a structured manner, achieving a complete closed loop from data collection to educational guidance. It addresses the problem of insufficient interpretability and operability of assessment results in existing technologies, demonstrating high credibility and strong practicality.

[0049] Based on the above embodiments, the interaction record data includes the number of interactions; the step of scoring the target child on each preset ability dimension based on the interaction record data to obtain the target child's current dimension score on each preset ability dimension includes: Based on the interaction record data, the target child is scored on each preset ability dimension to obtain the initial score of the target child on each preset ability dimension; Determine the statistical data of the first dimension scores of the target children in each of the preset ability dimensions within multiple historical periods; Based on the score statistics of each first dimension and the number of interactions, calculate the confidence weight of each preset capability dimension; Based on the confidence weights of each preset capability dimension and the initial scores, the scores of each current dimension are determined.

[0050] Specifically, in the process of scoring the target child on each preset ability dimension, the initial scores can be obtained by first using interactive record data to score the target child on each preset ability dimension. Each initial score refers to the raw score calculated directly based on the interactive record data of the current period, without any correction or adjustment. Each initial score can reflect the target child's immediate performance in the current period, but may be affected by accidental factors or emotional fluctuations, so it needs to be corrected in subsequent steps.

[0051] It is also necessary to determine the first-dimensional score statistics for the target children across multiple historical periods for each preset ability dimension. First-dimensional score statistics refer to indicators obtained through statistical analysis of the target children's historical scores for each preset ability dimension across multiple historical periods. These statistics may include the mean, standard deviation, maximum, minimum, or trend of each historical score for each preset ability dimension. The number of historical periods can be set as needed, such as 4, 8, or 10. First-dimensional score statistics can reflect the long-term performance level and stability of the target children across each preset ability dimension, providing a reference benchmark for the reliability of the current dimension score for the corresponding preset ability dimension in the current period. For example, if the target children have a high mean and a small standard deviation in their historical scores for the logical thinking ability dimension, it indicates that their performance in that logical thinking ability dimension is stable, and the current dimension score for that logical thinking ability dimension in the current period is highly reliable. Conversely, if the target children's historical scores for the logical thinking ability dimension fluctuate drastically, the current dimension score for that logical thinking ability dimension in the current period may be affected by random factors.

[0052] Next, using the statistical data of the scores for each first dimension and the number of interactions, the confidence weight for each preset capability dimension is calculated. The confidence weight is a value between 0 and 1, used to characterize the reliability of the initial score for the corresponding preset capability dimension within the current period.

[0053] The calculation of confidence weight takes into account two factors: first, the number of interactions, the more interactions, the higher the confidence weight; and second, the stability of historical performance, the smaller the fluctuation of the scores of each historical dimension on the preset capability dimension, the higher the confidence weight.

[0054] For example, the confidence weight can be calculated as follows: First, using the number of interactions, the first component is calculated through a first mapping function. For instance, when the number of interactions is less than the lower limit, the first component is 0; when the number of interactions reaches the upper limit, the first component is 1. When the number of interactions is greater than or equal to the lower limit but less than the upper limit, a linear interpolation algorithm is used to calculate the first component. This first component is used to characterize the sufficiency of the data, that is, whether the number of interactions used for scoring in the current period is sufficient to support reliable scores for each current dimension. The more interactions, the more sufficient the interaction record data on which the scoring is based, the less likely the scores for each current dimension are affected by random factors, and the closer the first component is to 1. Conversely, the fewer interactions, the sparser the interaction record data, and the more random the scores for each current dimension may be, the closer the first component is to 0 or even 0.

[0055] The second component is calculated using historical score statistics through a second mapping function. For example, volatility can be measured by the ratio of the standard deviation to the mean of historical scores; the lower the volatility, the higher the second component. This second component characterizes the stability of historical performance, i.e., whether the target child's long-term performance across all preset ability dimensions is consistent and predictable. The smaller the volatility of historical scores, the more stable and repeatable the target child's performance in the corresponding preset ability dimension, and the higher the reliability of the score for the current dimension in the current period, the closer the second component is to 1. Conversely, if historical scores fluctuate wildly, it indicates that the performance in that preset ability dimension is easily affected by external factors such as emotions and environment, and the lower the reliability of the score for the current dimension in the current period, the closer the second component is to 0.

[0056] Finally, the confidence weights can be obtained by multiplying the first component by the second component or by weighted summation.

[0057] Furthermore, the scores for each preset ability dimension can be determined using the confidence weights for each preset ability dimension and the initial scores. Here, the corresponding initial score can be multiplied by the confidence weight to obtain the corresponding dimension score. For example, if the initial score for a preset ability dimension is 3 points and the confidence weight for that preset ability dimension is 0.8, then the current dimension score for that preset ability dimension is 2.4 points; if the confidence weight for that preset ability dimension is only 0.4, then the current dimension score for that preset ability dimension is 1.2 points. In this way, initial scores with low confidence are significantly reduced, thereby reducing the interference of random factors on the current dimension score, making the final output current dimension score more reflective of the target child's stable ability level.

[0058] In this embodiment of the invention, the number of interactions and historical statistical data provide an objective basis for the calculation of confidence weights, enabling the scores of each current dimension to be adaptively adjusted according to data quality and historical performance. As a correction factor, confidence weights can effectively suppress scoring deviations caused by insufficient data or accidental fluctuations, improve the stability and reliability of the scores of each current dimension, and solve the problem that single evaluation results are easily affected by noise in the prior art.

[0059] Based on the above embodiments, the step of generating parent feedback suggestions based on the areas of expertise of each target includes: For any of the target areas of expertise, if the confidence weight corresponding to any target area of ​​expertise is greater than or equal to the confidence weight threshold, then the parent feedback suggestion corresponding to the advanced challenge task is generated. If the confidence weight corresponding to any of the target's areas of expertise is less than the confidence weight threshold, then the parent feedback suggestion corresponding to the reinforcement task is generated.

[0060] Specifically, the confidence weight threshold is a preset numerical limit used to distinguish between high and low confidence weights. For example, it can be set to 0.7 or 0.8, and the specific value can be adjusted according to the actual application scenario and the requirements for the rigor of the assessment. Advanced challenge tasks refer to activities or exercises that are more difficult than the target child's current ability level in the target area of ​​expertise, aiming to further stimulate and develop the child's potential in that area. For example, if the target area of ​​expertise is language expression ability and the confidence weight is greater than or equal to the confidence weight threshold, it indicates that the target area of ​​expertise is reliable, and the target child does indeed excel in this preset ability dimension. The parent feedback suggestion generated in this case is an advanced challenge task, which could be, "It is recommended that parents guide their children to try writing short stories or participating in role-playing dialogues to further improve their narrative ability and imagination." The parent feedback suggestions corresponding to advanced challenge tasks focus on expansion and deepening, helping the target child achieve higher-level ability development based on their existing strengths.

[0061] A confidence weight less than the confidence weight threshold indicates insufficient reliability of the current score in the corresponding preset ability dimension within the current period. This may be due to insufficient interaction frequency or significant historical performance fluctuations. In this case, it is not advisable to rush into advanced training. Instead, reinforcement tasks should be used to verify and stabilize the target child's performance in that preset ability dimension. Reinforcement tasks refer to practice activities that are equivalent to or slightly lower than the target child's current ability level in that area of ​​expertise, aiming to solidify the foundation through repetition and reinforcement. For example, if the target area of ​​expertise is logical thinking ability but the confidence weight is less than the confidence weight threshold, the generated parent feedback suggestion would be a reinforcement task, such as "It is recommended that parents complete basic classification games or simple mazes with their children, observe the child's problem-solving process, and provide encouragement." The parent feedback suggestion corresponding to the reinforcement task focuses on stabilization and verification, helping the target child build a more solid foundation of ability based on their existing foundation, while accumulating more reliable interaction record data for the next period's scoring.

[0062] In this embodiment of the invention, a comparison between confidence weights and confidence weight thresholds is introduced to achieve differentiated generation of parent feedback suggestions. For areas of expertise with reliable assessment results, advanced challenge tasks are pushed to promote ability improvement; for areas of expertise where assessment results still need verification, consolidation and reinforcement tasks are pushed to ensure a solid foundation. This differentiated feedback strategy enables refined customization of parent feedback suggestions, making the guidance received by target parents more closely aligned with the actual situation of the target child and the reliability of the assessment results. It solves the problems of lack of specificity and difficulty in adapting to different rating reliability in existing technologies, and has the beneficial effects of strong personalization and precise guidance.

[0063] Based on the above embodiments, determining at least one target area of ​​expertise from each of the preset capability dimensions based on the scores of each current dimension includes: The current dimension score is compared with the corresponding dimension score threshold to obtain the comparison results; Based on the comparison results, the areas of expertise of each target are determined from each of the preset capability dimensions; The score threshold for each dimension is determined based on the weight coefficient of each preset ability dimension, and each weight coefficient is used to characterize the degree of importance that the target parents of the target child attach to each preset ability dimension.

[0064] Specifically, a corresponding score threshold can be set for each preset ability dimension. The score threshold is a threshold value set separately for each preset ability dimension, used to judge whether the child's performance in that preset ability dimension meets the standard of proficiency.

[0065] Each dimension's score threshold can be determined using the corresponding preset ability dimension's weight coefficient. Each weight coefficient represents the degree of importance the target child's parents attach to the corresponding preset ability dimension. Target parents are the child's parents or other guardians, and also users of the parental devices associated with the companion robot. The degree of importance reflects the target parents' subjective judgment of the significance of different preset ability dimensions; for example, target parents may value language expression ability more, or they may value logical thinking ability more. The weight coefficient can be a single numerical value, such as between 0 and 1. The sum of the weight coefficients for all preset ability dimensions can be 1, or they can be independent of each other.

[0066] The relationship between the score threshold and weight coefficient for each preset ability dimension can be negatively correlated. This means the more the target parents value a particular preset ability dimension, the lower its score threshold, making it easier to identify that dimension as a strength and encouraging parents to focus more on it. Alternatively, it can be positively correlated, where the more the target parents value a preset ability dimension, the higher its score threshold, indicating higher requirements for it to be considered a strength. The specific relationship can be set according to actual needs. For example, with a negative correlation, the score threshold can be set to a baseline threshold multiplied by (1 - weight coefficient), so a larger weight coefficient results in a lower score threshold. The initial value of the weight coefficient can be based on general educational philosophies or the child's age group and can be dynamically adjusted based on parental feedback. For instance, if parents frequently view or like the assessment results for a preset ability dimension, the weight coefficient for that dimension can be automatically increased.

[0067] Furthermore, during scoring, the score for each current dimension is compared with its corresponding dimension score threshold to obtain the comparison result. The comparison result can be a Boolean value, such as true if greater than or equal to the dimension score threshold, and false if less than the threshold; or it can be a difference, such as the current dimension score minus the dimension score threshold, used for subsequent sorting or filtering. This step transforms the continuous current dimension scores into discrete comparison results, providing a clear basis for subsequent selection of target areas of expertise.

[0068] By utilizing the comparison results, each target's area of ​​expertise can be determined from each preset ability dimension. For example, if the current score of the language expression ability dimension is 2.5 points, the corresponding dimension score threshold is 2.0 points, and the comparison result is true, then the language expression ability dimension is determined to be the target's area of ​​expertise; if the current score of the logical thinking ability dimension is 1.8 points, the corresponding dimension score threshold is 2.0 points, and the comparison result is false, then the logical thinking ability dimension is not determined to be the target's area of ​​expertise.

[0069] In this embodiment of the invention, by introducing a dimensional score threshold determined based on weighting coefficients, the determination of the target's area of ​​expertise not only depends on the objective performance of the target child but also incorporates the subjective educational preferences of the target parents, thus achieving personalized scoring. The introduction of weighting coefficients allows the target parents' educational philosophy to be reflected in the scoring process, making it easier to identify the preset ability dimensions valued by the target parents as the target's area of ​​expertise, thereby gaining more attention and resource allocation in the parent feedback suggestions. This method considers both the objective current dimensional score and the target parents' subjective preferences when determining the target's area of ​​expertise, making the determined area of ​​expertise more aligned with the actual needs of family education, achieving personalized adaptation of the target's area of ​​expertise, and solving the problem of existing technologies having a single evaluation standard that cannot reflect parents' educational preferences.

[0070] Based on the above embodiments, the steps for determining the score thresholds for each dimension include: Obtain a norm database, which is used to store statistical data on the second dimension scores of peers who are of the same age, gender and region as the target child in each of the preset ability dimensions; Based on the score statistics of each of the second dimensions and the weight coefficients of each dimension, the score threshold of each dimension is determined.

[0071] Specifically, when determining the score thresholds for each dimension, a norm database can be obtained first. This norm database is a pre-built and continuously updated standardized reference dataset. Its data comes from a large-scale peer group with the same age, gender, and region as the target child, and is used to store statistical data on the second dimension scores of the peer group in each preset ability dimension. Among them, the same age group refers to a group with the same or similar birth year and month as the target child. For example, children can be divided into six-month age groups, such as three to three and a half years old, three and a half to four years old, etc., to ensure that the developmental levels of the comparison subjects are comparable.

[0072] The same region refers to the same or similar group as the target child's region, such as the same province or the same urban cluster, because different regions may have different educational resources and parenting styles, which may lead to differences in the target child's ability development.

[0073] The second-dimensional score statistics are indicators obtained by statistically calculating the current dimension scores of all children in the norm database for the corresponding preset ability dimension within the current period. For example, they may include the mean, standard deviation, and percentile distribution of the current dimension scores of all children for the corresponding preset ability dimension. The second-dimensional score statistics can provide an objective group reference benchmark for determining the dimension score thresholds for the target child's corresponding preset ability dimension, ensuring that the score thresholds for each dimension are not fixed values ​​set arbitrarily, but standardized values ​​calculated based on large-scale real data.

[0074] Then, using the statistical data of the scores of each second dimension and the weight coefficients of each dimension, the score thresholds of each dimension are determined. For any preset capability dimension, the average value of the statistical data of the scores of the second dimension of that preset capability dimension is used as the benchmark value. The weight coefficients of that preset capability dimension are multiplied by the benchmark value to obtain the dimension score threshold of that preset capability dimension.

[0075] For example, if the average score of the second dimension of a certain preset ability dimension is 1.5, the standard deviation is 0.5, and the weight coefficient is 0.8, then the dimension score threshold of this preset ability dimension can be set as: average + weight coefficient × standard deviation, i.e., 1.5 + 0.8 × 0.5 = 1.9. Alternatively, other calculation methods can be used, such as the dimension score threshold of this preset ability dimension being: average / weight coefficient, so that the larger the weight coefficient, the smaller the corresponding dimension score threshold.

[0076] In this embodiment of the invention, the introduction of a normative database ensures that the dimensional score threshold is no longer an isolated, arbitrarily set value, but rather a standardized reference value calculated based on statistical data of the second-dimensional scores of a large-scale peer group. This significantly improves the objectivity and comparability of the selected target areas of expertise. The introduction of weighting coefficients preserves the adjustment space for the personalized preferences of the target parents, allowing the dimensional score threshold to be flexibly adjusted according to the educational philosophies of the target parents while maintaining objective standards. This method considers both group statistical standards and individual preferences when determining the dimensional score threshold, ensuring that the determination of the target's areas of expertise is both scientifically based and meets the personalized needs of family education.

[0077] Based on the above embodiments, the step of pushing each of the target areas of expertise, the scoring reasons corresponding to each of the target areas of expertise, and the parent feedback suggestions to the parent device associated with the companion robot includes: Receive feedback operation data sent by the parent device; Based on the feedback operation data, the weight coefficients of each target's area of ​​expertise are updated.

[0078] Specifically, feedback action data refers to the action instructions or evaluation information returned by the target parent through an application or mini-program on the parent's device after receiving the target's area of ​​expertise, the rating reasons, and parent feedback suggestions. Feedback action data can include various types, such as the target parent's confirmation action for a target's area of ​​expertise indicating that the target parent approves of that area of ​​expertise, the target parent's rejection action indicating that the parent does not approve of that area of ​​expertise, the target parent's acceptance or disregard of parent feedback suggestions, and supplementary opinions or remarks manually entered by the target parent.

[0079] Feedback data can be received via wireless network. The companion robot's server or cloud platform listens for interface requests from the parent's device and parses the operation type and related parameters carried in the request. This step provides the target parent with a two-way interaction channel with the companion robot, enabling the target parent to move beyond being a passive recipient in their area of ​​expertise and become a participant in the selection process within that area.

[0080] Next, the weight coefficients of each target's area of ​​expertise are updated using feedback data. The specific method for updating these weight coefficients can be differentiated based on the type of feedback data. For example, if a parent confirms a target's area of ​​expertise, it indicates that the parent approves of and pays attention to that area, and the weight coefficient of the corresponding preset ability dimension can be appropriately increased. Conversely, if a parent rejects a target's area of ​​expertise, it indicates that the parent believes that the area does not align with their observations, and the weight coefficient of the corresponding preset ability dimension can be appropriately decreased. The adjustment range of the weight coefficients can be set according to a preset step size, such as increasing by 0.05 for each confirmation and decreasing by 0.05 for each rejection, while limiting the range of weight coefficient values ​​to prevent over-adjustment. The updated weight coefficients are stored and applied to the calculation of the dimension score thresholds in the next cycle, thereby achieving dynamic optimization of the scoring criteria. Through this feedback-driven update mechanism, the companion robot's evaluation preferences gradually align with the actual educational philosophies of the target parents, thus achieving personalized adaptive adjustments.

[0081] In this embodiment of the invention, the reception of feedback operation data establishes a two-way communication channel from the companion robot to the target parent and back to the companion robot, enabling the companion robot to perceive and utilize the target parent's subjective judgment. The updating of the weight coefficient transforms the target parent's feedback into an adjustment action of the weight coefficient, realizing dynamic optimization and personalized adaptation of the scoring standard. This allows the companion robot to continuously adjust the scoring standard based on the target parent's feedback, thereby better aligning with the target parent's educational philosophy and observation perspective in subsequent scoring. This solves the problem in the prior art where the scoring standard is fixed and cannot be dynamically adjusted based on the target parent's feedback.

[0082] Based on the above embodiments, the interaction record data includes the number of interactions and the target interaction data; The acquisition of interaction record data between the target child and the companion robot within the current period includes: Acquire the initial interaction data of each interaction collected by the sensors mounted on the companion robot; Based on the initial interaction data, the emotional state of the target child is identified; Based on the emotional state, the initial interaction data with negative emotional states are removed from each of the initial interaction data to obtain the target interaction data.

[0083] Specifically, initial interaction data refers to the raw data directly collected by the sensors on the companion robot during each interaction, without any filtering or processing. For example, when the target child speaks to the companion robot, the microphone records the complete voice waveform; when the target child makes facial expressions or gestures, the camera captures the corresponding video frame sequence. This initial interaction data contains all the information from the interaction process, including both effective content that helps in the evaluation and noisy data that may interfere with the evaluation.

[0084] Then, based on the initial interaction data, the target child's emotional state is identified. Emotional state refers to the emotional tendency exhibited by the target child during the interaction, such as pleasure, calmness, curiosity, frustration, irritability, or fatigue. Emotional state identification can be achieved by analyzing multimodal signals in the initial interaction data. For example, for speech signals, acoustic features such as pitch, speech rate, and volume can be extracted, and a pre-trained emotion recognition model can be used to determine the target child's emotional tendency; for facial images, facial muscle movement patterns can be detected, and facial expressions such as smiling, frowning, and pursing can be identified to infer the emotional state. Emotion recognition can employ machine learning models, such as convolutional neural networks or recurrent neural networks, or rule-based expert systems. The purpose of this step is to determine the target child's emotional state at each interaction, providing a basis for subsequent data filtering.

[0085] Next, based on emotional states, initial interaction data with negative emotional states is removed from each initial interaction data set to obtain the target interaction data. Negative emotional states refer to emotion types that are unfavorable to objective scoring, such as frustration, irritability, fatigue, and anger. When the target child is in a negative emotional state, their interactive behavior may deviate from normal levels, such as giving perfunctory answers to questions, refusing to interact, or showing impatience. Interaction data generated under these circumstances cannot accurately reflect the target child's ability level, and including them in the scoring would distort the scores for each current dimension. Therefore, initial interaction data collected under these negative emotional states is removed, and only initial interaction data collected under positive or neutral emotional states such as joy, calmness, and curiosity is retained as the target interaction data. Removal can be achieved by directly deleting the corresponding data files or marking them as invalid, skipping these data in subsequent processing. Through this emotional filtering mechanism, the final target interaction data used for scoring all come from interaction scenarios where the child's emotions are normal, thus ensuring the accuracy and reliability of the scores for each current dimension.

[0086] In this embodiment of the invention, an emotion filtering mechanism is introduced during the acquisition of interaction record data, so that subsequent scoring and target strength domain mining are based on high-quality and high-reliability data. By identifying the emotional state of the target child and removing interaction data in a negative emotional state, the quality control and noise filtering of the initial interaction data are achieved, solving the problem of evaluation distortion caused by mixed emotional interference in the interaction data in the prior art.

[0087] like Figure 2 As shown, based on the above embodiments, this embodiment of the invention provides a device for identifying children's areas of expertise, comprising: The acquisition module 21 is used to acquire the interaction record data between the target child and the companion robot within the current period; The scoring module 22 is used to score the target child on each preset ability dimension based on the interaction record data, and obtain the current dimension score of the target child on each preset ability dimension; The determination module 23 is used to determine at least one target area of ​​expertise from each of the preset ability dimensions based on the scores of each current dimension.

[0088] Based on the above embodiments, the children's strengths area discovery device provided in this embodiment of the invention includes a scoring reason for each current dimension score; it also includes a push module for: Based on the areas of expertise of each of the aforementioned targets, generate parent feedback and suggestions; The robot pushes the areas of expertise of each target, the scoring reasons for each area of ​​expertise, and the feedback and suggestions from the parents to the parent devices associated with the companion robot.

[0089] Based on the above embodiments, the child's strengths area discovery device provided in this embodiment of the invention includes interaction record data including the number of interactions; the scoring module is specifically used for: Based on the interaction record data, the target child is scored on each preset ability dimension to obtain the target child's current initial score on each preset ability dimension; Determine the statistical data of the first dimension scores of the target children in each of the preset ability dimensions within multiple historical periods; Based on the score statistics of each first dimension and the number of interactions, calculate the confidence weight of each preset capability dimension; Based on the confidence weights of each preset capability dimension and the initial scores, the scores of each current dimension are determined.

[0090] Based on the above embodiments, the children's strengths area discovery device provided in this embodiment of the invention, wherein the push module is specifically used for: For any of the target areas of expertise, if the confidence weight corresponding to any target area of ​​expertise is greater than or equal to the confidence weight threshold, then the parent feedback suggestion corresponding to the advanced challenge task is generated. If the confidence weight corresponding to any of the target's areas of expertise is less than the confidence weight threshold, then the parent feedback suggestion corresponding to the reinforcement task is generated.

[0091] Based on the above embodiments, the children's strengths area discovery device provided in this embodiment of the invention, wherein the determining module is specifically used for: The current dimension score is compared with the corresponding dimension score threshold to obtain the comparison results; Based on the comparison results, the areas of expertise of each target are determined from each of the preset capability dimensions; The score threshold for each dimension is determined based on the weight coefficient of each preset ability dimension, and each weight coefficient is used to characterize the degree of importance that the target parents of the target child attach to each preset ability dimension.

[0092] Based on the above embodiments, in the children's area of ​​expertise discovery device provided in this embodiment of the invention, the determining module is further used for: Obtain a norm database, which is used to store statistical data on the second dimension scores of peers who are of the same age, gender and region as the target child in each of the preset ability dimensions; Based on the score statistics of each of the second dimensions and the weight coefficients of each dimension, the score threshold of each dimension is determined.

[0093] Based on the above embodiments, the child's strength area discovery device provided in this embodiment of the invention further includes a feedback module, used for: Receive feedback operation data sent by the parent device; Based on the feedback operation data, the weight coefficients of each target's area of ​​expertise are updated.

[0094] Based on the above embodiments, the child's area of ​​expertise discovery device provided in this embodiment of the invention includes interaction record data including the number of interactions and target interaction data; The acquisition module is specifically used for: Acquire the initial interaction data of each interaction collected by the sensors mounted on the companion robot; Based on the initial interaction data, the emotional state of the target child is identified; Based on the emotional state, the initial interaction data with negative emotional states are removed from each of the initial interaction data to obtain the target interaction data.

[0095] Specifically, the functions of each module in the children's area of ​​expertise discovery device provided in this embodiment of the invention correspond one-to-one with the operation flow of each step in the above method-like embodiments, and the achieved effects are also the same. For details, please refer to the above embodiments, and this will not be repeated in this embodiment of the invention.

[0096] Figure 3 An example is a schematic diagram of the physical structure of a companion robot, such as... Figure 3 As shown, the companion robot may include a processor 810, a communications interface 820, a memory 830, and a communication bus 840. The processor 810, communications interface 820, and memory 830 communicate with each other via the communication bus 840. The processor 810 can call logical instructions in the memory 830 to execute the child's area of ​​expertise discovery method provided in the above embodiments.

[0097] Furthermore, the logical instructions in the aforementioned memory 830 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to related technologies, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0098] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a computer-readable storage medium. When the computer program is executed by a processor, the computer is able to execute the children's strengths area discovery method provided in the above embodiments.

[0099] In another aspect, the present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to perform the children's strengths area discovery method provided in the above embodiments. This computer-readable storage medium can be either a non-transitory computer-readable storage medium or a transient computer-readable storage medium, and is not specifically limited herein.

[0100] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0101] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the parts that contribute to the related technology, can be embodied in the form of software products. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0102] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention 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; and these 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 the present invention.

Claims

1. A method for identifying children's areas of expertise, characterized in that, include: Acquire interaction records between the target child and the companion robot within the current period; Based on the interaction record data, the target child is scored on each preset ability dimension to obtain the current dimension score of the target child on each preset ability dimension; Based on the scores of each current dimension, at least one target area of ​​expertise is determined from each of the preset capability dimensions.

2. The method for identifying children's areas of expertise according to claim 1, characterized in that, Each current dimension score has a corresponding scoring reason; the step of determining at least one target area of ​​expertise from each preset ability dimension based on the current dimension scores includes: Based on the areas of expertise of each of the aforementioned targets, generate parent feedback and suggestions; The robot pushes the areas of expertise of each target, the scoring reasons for each area of ​​expertise, and the feedback and suggestions from the parents to the parent devices associated with the companion robot.

3. The method for identifying children's areas of expertise according to claim 2, characterized in that, The interaction record data includes the number of interactions; the step of scoring the target child on each preset ability dimension based on the interaction record data to obtain the target child's current dimension score on each preset ability dimension includes: Based on the interaction record data, the target child is scored on each preset ability dimension to obtain the target child's current initial score on each preset ability dimension; Determine the statistical data of the first dimension scores of the target children in each of the preset ability dimensions within multiple historical periods; Based on the score statistics of each first dimension and the number of interactions, calculate the confidence weight of each preset capability dimension; Based on the confidence weights of each preset capability dimension and the initial scores, the scores of each current dimension are determined.

4. The method for identifying children's areas of expertise according to claim 3, characterized in that, The process of generating parent feedback suggestions based on the areas of expertise of each target includes: For any of the target areas of expertise, if the confidence weight corresponding to any target area of ​​expertise is greater than or equal to the confidence weight threshold, then the parent feedback suggestion corresponding to the advanced challenge task is generated. If the confidence weight corresponding to any of the target's areas of expertise is less than the confidence weight threshold, then the parent feedback suggestion corresponding to the reinforcement task is generated.

5. The method for identifying children's areas of expertise according to claim 2, characterized in that, The step of determining at least one target area of ​​expertise from each of the preset ability dimensions based on the scores of each current dimension includes: The current dimension score is compared with the corresponding dimension score threshold to obtain the comparison results; Based on the comparison results, the areas of expertise of each target are determined from each of the preset capability dimensions; The score threshold for each dimension is determined based on the weight coefficient of each preset ability dimension, and each weight coefficient is used to characterize the degree of importance that the target parents of the target child attach to each preset ability dimension.

6. The method for identifying children's areas of expertise according to claim 5, characterized in that, The steps for determining the score thresholds for each dimension include: Obtain a norm database, which is used to store statistical data on the second dimension scores of peers who are of the same age, gender and region as the target child in each of the preset ability dimensions; Based on the score statistics of each of the second dimensions and the weight coefficients of each dimension, the score threshold of each dimension is determined.

7. The method for identifying children's areas of expertise according to claim 6, characterized in that, The step of pushing each of the target's areas of expertise, the corresponding scoring reasons for each of the target's areas of expertise, and the parent's feedback and suggestions to the parent's device associated with the companion robot, then includes: Receive feedback operation data sent by the parent device; Based on the feedback operation data, the weight coefficients of each target's area of ​​expertise are updated.

8. The method for identifying children's areas of expertise according to any one of claims 1-7, characterized in that, The interaction record data includes the number of interactions and the target interaction data; obtaining the interaction record data between the target child and the companion robot within the current period includes: Acquire the initial interaction data of each interaction collected by the sensors mounted on the companion robot; Based on the initial interaction data, the emotional state of the target child is identified; Based on the emotional state, the initial interaction data with negative emotional states are removed from each of the initial interaction data to obtain the target interaction data.

9. A device for exploring a child's area of ​​expertise, characterized in that, include: The acquisition module is used to acquire interaction record data between the target child and the companion robot within the current period; The scoring module is used to score the target child on each preset ability dimension based on the interaction record data, and obtain the current dimension score of the target child on each preset ability dimension; The determination module is used to determine at least one target area of ​​expertise from each of the preset capability dimensions based on the scores of each of the current dimensions.

10. A companion robot, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method for identifying children's areas of expertise as described in any one of claims 1-8.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the method for identifying children's areas of expertise as described in any one of claims 1-8.

Images