Method and system for evaluating cognitive competence and psychological development of teenagers based on artificial intelligence

Abstract

The invention discloses a teenager cognitive competence and psychological development evaluation method and system based on artificial intelligence, and relates to the technical field of teenager cognitive competence and psychological development evaluation. Progressive development and stepped transition of cognitive competence are effectively captured, and misjudgment of a traditional time sequence model on discontinuous changes is avoided. Based on a continuous coherence characteristic monitoring mechanism of topological data analysis, cognitive representation dimension mutation caused by strategy recombination and neuroplasticity is accurately recognized, and the delay bottleneck of a traditional method on burst mode recognition is broken through; and the incremental learning triggering module and the evaluation path optimization module cooperatively operate, so that online evolution of model parameters is realized while the calculation overhead is reduced, and continuity and consistency of long-term tracking are kept.

Description

Technical Field

[0001] The present invention relates to the technical field of assessment of adolescent cognitive ability and psychological development, and in particular to an artificial intelligence-based assessment method and system for adolescent cognitive ability and psychological development. Background Art

[0002] Currently, the assessment of adolescent cognitive and psychological development is gradually shifting from single static evaluations to long-term dynamic tracking, relying on wearable devices and multimodal data fusion technology to build digital twin models of individual cognitive development; the new generation of systems generally adopts an adaptive assessment framework, dynamically adjusts task difficulty through reinforcement learning, and introduces federated learning to achieve cross-institutional data sharing; this type of solution performs outstandingly in capturing short-term cognitive fluctuations and can identify instantaneous state changes such as temporary decline in executive function caused by exam pressure.

[0003] Adolescents' cognitive strategies will undergo discontinuous transitions due to neural development and accumulation of social experience, but mainstream time series models (such as LSTM and Transformer) assume that changes in cognitive characteristics conform to a stationary random process, resulting in systematic biases in long-term tracking. Typical problems include: misjudging the improvement in impulse control caused by the maturation of the prefrontal cortex as the effect of intervention measures; there is a 3-6 month recognition delay in the reconstruction of cognitive patterns caused by drastic changes in peer relationships; it is impossible to distinguish between temporary strategy adjustments caused by fluctuations in hormone levels and permanent changes in cognitive structure; the current solution relies on regular full-scale model retraining, but faces bottlenecks such as high computational costs and interruptions to the continuous tracking process.

[0004] Cutting-edge research attempts to use neural differential equations to construct continuous models of cognitive evolution, mapping discrete observation data to underlying dynamic systems. For example, the NeuralCDE framework captures the differential characteristics of development trajectories through interpolation control nodes, reducing the missed reporting rate of burst modes. Another team uses online Bayesian knowledge transfer technology to achieve gradual adaptation of cognitive mutations through weight space elasticity while ensuring the memory of historical patterns. However, the problem of dimensional mutations in cognitive representation caused by neural plasticity has not yet been solved, and experts are still required to regularly expand the model dimensions. Summary of the Invention

[0005] In view of the above existing problems, the present invention is proposed.

[0006] The present invention provides an artificial intelligence-based method and system for evaluating adolescent cognitive ability and psychological development. Although existing solutions turn to dynamic tracking, they are limited by the stationary assumption and static modeling, making it difficult to cope with the discontinuous transitions of adolescent cognitive strategies, leading to misjudgment and delays.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0008] In a first aspect, an embodiment of the present invention provides an artificial intelligence-based method for assessing adolescent cognitive abilities and psychological development, which includes:

[0009] Step S1, continuously collecting the subject's physiological signals and behavioral operation sequences through the wearable device's galvanic skin response sensor and touch screen interaction module;

[0010] Step S2: Using neural differential equations to construct a continuous model of cognitive evolution, mapping multimodal observation data at discrete time points to the underlying dynamical system;

[0011] Step S3: Detecting phase mutation events of cognitive strategies based on a topological data analysis method. When a change in the dimension of a persistent coherent feature is identified, triggering a model incremental learning mechanism. Detecting the change in the dimension of a persistent coherent feature specifically involves constructing a Vietoris-Rips complex in the cognitive feature manifold space, calculating the length distribution of the persistent intervals in each dimension, and determining a phase mutation when the persistent interval of a one-dimensional ring structure suddenly increases.

[0012] Step S4: Generate a digital twin of cognitive development containing a timestamp and dynamically update the dimensional space and evolution path of individual cognitive characteristics.

[0013] As a preferred solution of the method for assessing the cognitive ability and psychological development of adolescents based on artificial intelligence described in the present invention, data processing is performed in step S1, and the steps include:

[0014] Perform complex domain time-frequency decomposition on the original physiological signal to separate the environmental noise and micro-motion feature components;

[0015] Construct a multi-scale time alignment module to calibrate the millisecond delay between behavioral manipulation events and physiological responses;

[0016] Use an adversarial training framework to eliminate the interference of culture-specific behavioral patterns on the underlying cognitive feature representation;

[0017] The adversarial training framework includes a cascade structure of a gradient reversal layer and an attribute discriminator, where the discriminator distinguishes samples from Eastern and Western cultural backgrounds through a binary classification loss function, and the generator optimizes feature extraction by maximizing the discriminator confusion rate.

[0018] As a preferred solution of the method for assessing cognitive ability and psychological development of adolescents based on artificial intelligence described in the present invention, in step S1, the original multimodal input is denoted as x, and the feature extractor mapping is z=F(x;θ1),

[0019] On the discriminator side, the Wasserstein distance is used to estimate the difference in sample distribution between the East and the West, and the discriminator loss is defined as:

[0020]

[0021] Among them, E x～P1 [·] represents the expectation operation of random variables under distribution P1, E x～P2 [·] represents the expectation operation of random variables under distribution P2, Represents the interpolation samples , θ1 represents the feature extractor parameter, θ2 represents the discriminator parameter, P1 represents the distribution of Western culture samples, P2 represents the distribution of Eastern culture samples, F(·;θ1) represents the feature extraction mapping, D(·;θ2) represents the attribute discriminant mapping, λ1 represents the gradient penalty coefficient, represents the interpolated sample of the discriminator input, represents the gradient operator for the discriminator input, and |·|2 represents the two-norm;

[0022] The interpolation samples The calculation formula is:

[0023]

[0024] Among them, ∈1 represents the interpolation coefficient, which obeys the uniform distribution z W represents the characteristics of Western samples, z E Indicates characteristics of oriental samples;

[0025] On the feature extractor side, cognitive task classification and cultural confrontation are simultaneously incorporated into the total loss:

[0026] L F =L task +γ1(E x～P2 [D(F(x;θ1);θ2)]-E x～P1 [D(F(x;θ1);θ2)]),

[0027] Among them, L task represents the cognitive task classification loss, γ1 represents the adversarial loss weight;

[0028] The classification loss for cognitive tasks is defined as:

[0029] L task =E (x,y) [l(C(F(x;θ1)),y)],

[0030] Here, (x, y) represents samples and their labels, C(·) represents the cognitive task classifier, and l(·, ·) represents the cross-entropy loss function.

[0031] As a preferred embodiment of the method for assessing the cognitive ability and psychological development of adolescents based on artificial intelligence described in the present invention, the cognitive evolution continuous model is constructed in the following manner:

[0032] The historical cognition assessment data were encoded into a continuous time latent variable sequence;

[0033] Use controlled differential equations to model the dynamics of latent variables along neural development;

[0034] Introducing the branch of counterfactual reasoning to distinguish endogenous cognitive transitions from exogenous intervention influence paths;

[0035] The neural differential equation adopts the form of a controlled differential equation, in which the control term is generated by the interpolation function of the historical cognitive feature trajectory, the differential operator is constructed by a multi-layer perceptron, and the time step is dynamically adjusted according to the subject's evaluation frequency.

[0036] As a preferred embodiment of the method for assessing the cognitive ability and psychological development of adolescents based on artificial intelligence described in the present invention, in step S2, a continuous model of cognitive evolution is constructed using a neural differential equation, and the steps include:

[0037] Let the initial latent variable state be mapped by the encoder:

[0038]

[0039] Among them, z0 represents the initial latent variable, represents the multimodal observation at the initial time t0, g represents the initial encoding map, and β1 represents the encoding map parameter;

[0040] The NeuralCDE standard form is used to describe the continuous evolution of latent variables over time, which is expressed as:

[0041]

[0042] Where z(t) represents the latent variable at time t, f represents the neural network differential function, α1 represents the function parameter, U(s) represents the control path constructed by discrete observations, and s is the integral variable;

[0043] The control path U(t) is defined by the interpolation basis function and is expressed as:

[0044]

[0045] Among them, N1 represents the total number of observation points, represents the i-th observation input, φ i (t) represents the corresponding interpolation basis function.

[0046] As a preferred embodiment of the method for assessing the cognitive ability and psychological development of adolescents based on artificial intelligence described in the present invention, the method for detecting the phase mutation event of cognitive strategy in step S3 is:

[0047] Construct the Vietoris-Rips complex in the cognitive feature manifold space and calculate the persistence graph entropy of the kth dimension. The formula is:

[0048]

[0049] Among them, H (k) (t) represents the kth maintained long-term entropy at time t, N k (t) represents the number of continuous intervals in the kth dimension, k represents the dimension index of the homology group, represents the length of the i-th interval, Indicates the corresponding normalized weight;

[0050] For a one-dimensional ring with k=1, the change of its persistent entropy over time is monitored, and the average persistent entropy is calculated by sliding the time window:

[0051]

[0052] Among them, τ1 represents the length of the time window. If And it continues, δ1 represents the persistent entropy threshold for determining phase mutation, then it is determined to be a phase mutation and the model incremental learning mechanism is triggered.

[0053] In a second aspect, the present invention provides an artificial intelligence-based system for assessing adolescent cognitive abilities and psychological development, comprising:

[0054] Multimodal data acquisition module, integrating flexible electronic skin sensors and adaptive touch interface;

[0055] Edge computing unit, deploying lightweight time series feature extraction network and privacy protection mechanism;

[0056] Dynamic assessment path generator, which adjusts task types and difficulty parameters according to real-time cognitive status;

[0057] Cognitive development visualization engine generates three-dimensional spatiotemporal evolution maps and intervention recommendations.

[0058] As a preferred solution of the artificial intelligence-based adolescent cognitive ability and psychological development assessment system described in the present invention, the privacy protection mechanism is specifically implemented as follows:

[0059] A federated learning architecture is used for distributed model training, with the original data retained on the local terminal.

[0060] Add controllable noise to cognitive feature vectors through differential privacy technology;

[0061] Use homomorphic encryption to transmit sensitive intermediate data during the fingerprint matching process.

[0062] As a preferred solution of the artificial intelligence-based adolescent cognitive ability and psychological development assessment system of the present invention, the dynamic assessment path generator further includes:

[0063] A strategy fingerprint feature library stores a feature transfer matrix of typical cognitive development patterns. This feature transfer matrix is constructed by calculating the cosine similarity variance of cross-task operational features and includes a 52-dimensional transfer relationship map between three types of tasks: risk decision-making, pattern recognition, and social simulation.

[0064] Online Bayesian optimizer, dynamically selecting evaluation dimensions based on the sparsity of the current cognitive space;

[0065] The sudden jump event response module inserts a cross-domain verification task when a sudden phase change is detected.

[0066] As a preferred solution of the artificial intelligence-based adolescent cognitive ability and psychological development assessment system described in the present invention, the visualization engine includes:

[0067] a cultural adaptability calibration layer to eliminate the bias in autonomy ratings caused by the collectivist decision-making model;

[0068] Developmental trajectory prediction interface, simulating the branching paths of cognitive evolution under different intervention measures;

[0069] Hidden risk warning unit detects early psychological development deviations based on abnormal strategy migration.

[0070] The beneficial effects of the present invention are as follows: the neural differential equation of the present invention maps discrete observations into a continuous latent variable dynamic system, effectively capturing the progressive development and step-by-step transition of cognitive abilities, and avoiding the misjudgment of discontinuous changes by traditional time series models; the continuous coherent feature monitoring mechanism based on topological data analysis accurately identifies strategy reorganization and sudden changes in cognitive representation dimensions caused by neural plasticity, breaking through the delay bottleneck of traditional methods for sudden pattern recognition; the incremental learning trigger and evaluation path optimization modules work together to achieve online evolution of model parameters while reducing computing overhead, maintaining continuity and consistency in long-term tracking; the adversarial training framework and the cultural calibration layer deeply decouple regional behavioral preferences, eliminating the systematic bias of collectivist decision-making models in autonomy assessment; federated learning and homomorphic encryption technology support cross-institutional cognitive development model mining without leaking original data, solving the problem of data silos. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0072] Figure 1 This is a flow chart of the method for assessing adolescent cognitive abilities and psychological development based on artificial intelligence in Example 1.

[0073] Figure 2 This is a schematic diagram of the framework of the artificial intelligence-based adolescent cognitive ability and psychological development assessment system in Example 1. DETAILED DESCRIPTION

[0074] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0075] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0076] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0077] Example 1, with reference to Figure 1 and Figure 2 This embodiment provides an artificial intelligence-based method for assessing adolescent cognitive abilities and psychological development, comprising the following steps:

[0078] Step S1, continuously collecting the subject's physiological signals and behavioral operation sequences through the wearable device's galvanic skin response sensor and touch screen interaction module;

[0079] Data processing is performed in step S1, and the steps include:

[0080] Perform complex domain time-frequency decomposition on the original physiological signal to separate the environmental noise and micro-motion feature components;

[0081] Construct a multi-scale time alignment module to calibrate the millisecond delay between behavioral manipulation events and physiological responses;

[0082] Use an adversarial training framework to eliminate the interference of culture-specific behavioral patterns on the underlying cognitive feature representation;

[0083] The adversarial training framework consists of a cascade structure of a gradient reversal layer and an attribute discriminator. The discriminator distinguishes between samples from Eastern and Western cultural backgrounds using a binary classification loss function, and the generator optimizes feature extraction by maximizing the discriminator's confusion rate.

[0084] In step S1, let the original multimodal input be x, and the feature extractor mapping is z = F(x; θ1),

[0085] On the discriminator side, the Wasserstein distance is used to estimate the difference in sample distribution between the East and the West, and the discriminator loss is defined as:

[0086]

[0087] Among them, E x～P1 [·] represents the expectation operation of random variables under distribution P1, E x～P2 [·] represents the expectation operation of random variables under distribution P2, Represents the interpolation samples , θ1 represents the feature extractor parameter, θ2 represents the discriminator parameter, P1 represents the distribution of Western culture samples, P2 represents the distribution of Eastern culture samples, F(·;θ1) represents the feature extraction mapping, D(·;θ2) represents the attribute discriminant mapping, λ1 represents the gradient penalty coefficient, represents the interpolated sample of the discriminator input, represents the gradient operator for the discriminator input, and |·|2 represents the two-norm;

[0088] The interpolation samples The calculation formula is:

[0089]

[0090] Among them, ∈1 represents the interpolation coefficient, which obeys the uniform distribution z W represents the characteristics of Western samples, z E Indicates characteristics of oriental samples;

[0091] On the feature extractor side, cognitive task classification and cultural confrontation are simultaneously incorporated into the total loss:

[0092] L F =L task +γ1(E x～P2 [D(F(x;θ1);θ2)]-E x～P1 [D(F(x;θ1);θ2)]),

[0093] Among them, L taskrepresents the cognitive task classification loss, γ1 represents the adversarial loss weight;

[0094] The classification loss for cognitive tasks is defined as:

[0095] L task =E (x,y) [l(C(F(x;θ1)),y)],

[0096] Where (x, y) represents the sample and its label, C(·) represents the cognitive task classifier, and l(·,·) represents the cross entropy loss function;

[0097] Specifically, this adversarial training framework imposes explicit adversarial constraints on the feature extractor by measuring the Wasserstein distance between the distributions of Eastern and Western cultural samples. The discriminator estimates the distribution distance and combines it with a gradient penalty to improve the Lipschitz continuity of the discriminant function, thereby avoiding training instability and gradient vanishing problems. The extractor introduces an adversarial confusion loss to effectively eliminate the interference of culturally specific behavioral patterns on cognitive representations, while retaining the cognitive task classification loss to ensure that the extracted features still contain information related to cognitive ability.

[0098] Step S2: Using neural differential equations to construct a continuous model of cognitive evolution, mapping multimodal observation data at discrete time points to the underlying dynamical system;

[0099] The cognitive evolution continuum model is constructed in the following way:

[0100] The historical cognition assessment data were encoded into a continuous time latent variable sequence;

[0101] Use controlled differential equations to model the dynamics of latent variables along neural development;

[0102] Introducing the branch of counterfactual reasoning to distinguish endogenous cognitive transitions from exogenous intervention influence paths;

[0103] The neural differential equation adopts the form of a controlled differential equation, in which the control term is generated by the interpolation function of the historical cognitive feature trajectory, the differential operator is constructed by a multi-layer perceptron, and the time step is dynamically adjusted according to the subject's evaluation frequency;

[0104] In step S2, a neural differential equation is used to construct a continuous model of cognitive evolution, which includes the following steps:

[0105] Let the initial latent variable state be mapped by the encoder:

[0106]

[0107] Among them, z0 represents the initial latent variable, represents the multimodal observation at the initial time t0, g represents the initial encoding map, and β1 represents the encoding map parameter;

[0108] The NeuralCDE standard form is used to describe the continuous evolution of latent variables over time, which is expressed as:

[0109]

[0110] Where z(t) represents the latent variable at time t, f represents the neural network differential function, α1 represents the function parameter, U(s) represents the control path constructed by discrete observations, and s is the integral variable;

[0111] The control path U(t) is defined by the interpolation basis function and is expressed as:

[0112]

[0113] Among them, N1 represents the total number of observation points, represents the i-th observation input, φ i (t) represents the corresponding interpolation basis function;

[0114] Specifically, the neural CDE model maps discrete multimodal observations into continuous latent variable trajectories, constructs control paths through initial encoding and interpolation functions, and enables the model to capture the subtle dynamics of cognitive evolution. The parameter α1 of the function f learns the latent dynamics in the neural network and forms a controlled differential equation together with the control term U to ensure adaptive modeling of irregular evaluation frequencies. This method uses integral operations to accumulate historical events, providing a smooth evolution curve, which is convenient for the intervention effect analysis of subsequent counterfactual reasoning branches. The interpolation basis function φ i It ensures signal continuity, reduces information loss due to downsampling, and uses the Adjoint method to calculate gradients, reducing computational overhead while ensuring model trainability.

[0115] Step S3: Detecting phase mutation events of cognitive strategies based on topological data analysis methods. When a change in the dimension of a persistent homology feature is identified, the model's incremental learning mechanism is triggered. Detecting changes in the dimension of a persistent homology feature specifically involves constructing a Vietoris-Rips complex in the cognitive feature manifold space, calculating the length distribution of the persistent intervals in each dimension, and determining a phase mutation when the persistent interval of a one-dimensional ring structure suddenly increases.

[0116] The method for detecting the phase mutation event of the cognitive strategy in step S3 is:

[0117] Construct the Vietoris-Rips complex in the cognitive feature manifold space and calculate the persistence graph entropy of the kth dimension. The formula is:

[0118]

[0119] Among them, H (k) (t) represents the kth maintained long-term entropy at time t, N k (t) represents the number of continuous intervals in the kth dimension, k represents the dimension index of the homology group, represents the length of the i-th interval, Indicates the corresponding normalized weight;

[0120] For a one-dimensional ring with k=1, the change of its persistent entropy over time is monitored, and the average persistent entropy is calculated by sliding the time window:

[0121]

[0122] Among them, τ1 represents the length of the time window. If And it continues, δ1 represents the persistent entropy threshold for determining phase mutation, then it is determined to be a phase mutation and the model incremental learning mechanism is triggered;

[0123] Specifically, the persistence graph entropy metric, combined with VR complexes and entropy calculations, provides a clear quantification of the persistence changes of high-dimensional homological features. A one-dimensional ring structure is selected as the key homological dimension, and the length of its persistence interval directly reflects the strategy synergy and the stability of the loop structure. The entropy value considers both the number and length distribution of intervals, making it more robust than simple counting. Monitoring the average entropy over a sliding window can reduce misjudgments caused by sudden noise. When the average persistence entropy exceeds the threshold δ1 and the duration τ1 meets the standard, the cognitive strategy is considered to have entered a new phase. The incremental learning mechanism then updates the feature extraction and prediction modules online, avoiding complete retraining and reducing latency.

[0124] Step S4: Generate a digital twin of cognitive development containing a timestamp and dynamically update the dimensional space and evolution path of individual cognitive characteristics.

[0125] This embodiment also provides an evaluation system for the above-mentioned AI-based evaluation method for adolescent cognitive ability and psychological development, including:

[0126] Multimodal data acquisition module, integrating flexible electronic skin sensors and adaptive touch interface;

[0127] Edge computing unit, deploying lightweight time series feature extraction network and privacy protection mechanism;

[0128] The privacy protection mechanism is specifically implemented as follows:

[0129] A federated learning architecture is used for distributed model training, with the original data retained on the local terminal.

[0130] Add controllable noise to cognitive feature vectors through differential privacy technology;

[0131] Use homomorphic encryption to transmit sensitive intermediate data during the fingerprint matching process;

[0132] Dynamic assessment path generator, which adjusts task types and difficulty parameters according to real-time cognitive status;

[0133] The dynamic assessment path generator further includes:

[0134] The strategy fingerprint feature library stores the feature transfer matrix of typical cognitive development patterns. The feature transfer matrix is constructed by calculating the cosine similarity variance of cross-task operation features and contains a 52-dimensional transfer relationship map between three types of tasks: risk decision-making, pattern recognition, and social simulation.

[0135] Online Bayesian optimizer, dynamically selecting evaluation dimensions based on the sparsity of the current cognitive space;

[0136] The sudden jump event response module inserts a cross-domain verification task when a phase mutation is detected;

[0137] The strategy fingerprint feature library is used to store the feature transfer matrix between tasks in a typical cognitive development model;

[0138] Assume that the total number of tasks is N, and the normalized fingerprint feature vector of the i-th task is Construct a third-order tensor:

[0139]

[0140] in, represents the feature tensor, Represents the tensor product value of task i and task j in the kth dimension;

[0141] Compute the tensor product mean μ of the pair ij :

[0142]

[0143] Finally, define the characteristic mobility matrix M∈R N×N element:

[0144]

[0145] Among them, N represents the total number of tasks, d represents the dimension of the fingerprint feature vector, represents the normalized feature vector of task i, i, j∈{1,…,N} represents the task index, k∈{1,…,d} represents the feature dimension index, μ ij represents the mean tensor product of task pair (i, j), M ij Indicates the feature transfer value from task i to task j;

[0146] Specifically, the migration matrix definition makes full use of tensor product operations to refine the interactions of task fingerprint feature vectors in each dimension into tensor elements, and captures the consistency and difference of cross-task features through variance calculation. Compared with directly calculating the variance of cosine similarity, this method more comprehensively preserves the intrinsic mutual information of features through the high-dimensional tensor structure, improving the accuracy and robustness of migration relationship modeling. The normalized vector ensures the comparability of metrics between different tasks, while the variance metric effectively highlights the coordinated changes of features under the dynamic evolution stage.

[0147] Cognitive development visualization engine, generating three-dimensional spatiotemporal evolution maps and intervention recommendations;

[0148] The visualization engine includes:

[0149] a cultural adaptability calibration layer to eliminate the bias in autonomy ratings caused by the collectivist decision-making model;

[0150] Developmental trajectory prediction interface, simulating the branching paths of cognitive evolution under different intervention measures;

[0151] Hidden risk warning unit detects early psychological development deviations based on abnormal strategy migration.

[0152] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for assessing adolescent cognitive ability and psychological development based on artificial intelligence, characterized by: The method comprises: step S1, continuously collecting physiological signals and behavioral operation sequences of the subject through the skin galvanic response sensor and the touch screen interaction module of the wearable device; Step S2: Using neural differential equations to construct a continuous model of cognitive evolution, mapping multimodal observation data at discrete time points to the underlying dynamical system; Step S3: Detecting phase mutation events of cognitive strategies based on a topological data analysis method. When a change in the dimension of a persistent coherent feature is identified, triggering a model incremental learning mechanism. Detecting the change in the dimension of a persistent coherent feature specifically involves constructing a Vietoris-Rips complex in the cognitive feature manifold space, calculating the length distribution of the persistent intervals in each dimension, and determining a phase mutation when the persistent interval of a one-dimensional ring structure suddenly increases. Step S4: Generate a digital twin of cognitive development containing a timestamp and dynamically update the dimensional space and evolution path of individual cognitive characteristics.

2. The method for assessing adolescent cognitive ability and psychological development based on artificial intelligence according to claim 1, characterized in that: Data processing is performed in step S1, and the steps include: Perform complex domain time-frequency decomposition on the original physiological signal to separate the environmental noise and micro-motion feature components; Construct a multi-scale time alignment module to calibrate the millisecond delay between behavioral manipulation events and physiological responses; Use an adversarial training framework to eliminate the interference of culture-specific behavioral patterns on the underlying cognitive feature representation; The adversarial training framework includes a cascade structure of a gradient reversal layer and an attribute discriminator, where the discriminator distinguishes samples from Eastern and Western cultural backgrounds through a binary classification loss function, and the generator optimizes feature extraction by maximizing the discriminator confusion rate.

3. The method for assessing adolescent cognitive ability and psychological development based on artificial intelligence according to claim 2, characterized in that: In step S1, let the original multimodal input be x, and the feature extractor mapping is z = F(x; θ1), On the discriminator side, the Wasserstein distance is used to estimate the difference in sample distribution between the East and the West, and the discriminator loss is defined as: Among them, E x～P1 [·] represents the expectation operation of random variables under distribution P1, E x～P2 [·] represents the expectation operation of random variables under distribution P2, Represents the interpolation samples , θ1 represents the feature extractor parameter, θ2 represents the discriminator parameter, P1 represents the distribution of Western culture samples, P2 represents the distribution of Eastern culture samples, F(·;θ1) represents the feature extraction mapping, D(·;θ2) represents the attribute discriminant mapping, λ1 represents the gradient penalty coefficient, represents the interpolated sample of the discriminator input, represents the gradient operator for the discriminator input, and |·|2 represents the two-norm; The interpolation samples The calculation formula is: Among them, ∈1 represents the interpolation coefficient, which obeys the uniform distribution z W represents the characteristics of Western samples, z E Indicates characteristics of oriental samples; On the feature extractor side, cognitive task classification and cultural confrontation are simultaneously incorporated into the total loss: L F =L task +γ1(E x～P2 [D(F(x;θ1);θ2)]-E x～P1 [D(F(x;θ1);θ2)]), Among them, L task represents the cognitive task classification loss, γ1 represents the adversarial loss weight; The classification loss for cognitive tasks is defined as: L task =E (x,y) [l(C(F(x;θ1)),y)], Here, (x, y) represents samples and their labels, C(·) represents the cognitive task classifier, and l(·, ·) represents the cross-entropy loss function.

4. The method for assessing adolescent cognitive ability and psychological development based on artificial intelligence according to claim 1, characterized in that: The cognitive evolution continuum model is constructed in the following way: The historical cognition assessment data were encoded into a continuous time latent variable sequence; Use controlled differential equations to model the dynamics of latent variables along neural development; Introducing the branch of counterfactual reasoning to distinguish endogenous cognitive transitions from exogenous intervention influence paths; The neural differential equation adopts the form of a controlled differential equation, in which the control term is generated by the interpolation function of the historical cognitive feature trajectory, the differential operator is constructed by a multi-layer perceptron, and the time step is dynamically adjusted according to the subject's evaluation frequency.

5. The method for assessing adolescent cognitive ability and psychological development based on artificial intelligence according to claim 4, characterized in that: In step S2, a neural differential equation is used to construct a continuous model of cognitive evolution, which includes the following steps: Let the initial latent variable state be mapped by the encoder: Among them, z0 represents the initial latent variable, represents the multimodal observation at the initial time t0, g represents the initial encoding map, and β1 represents the encoding map parameter; The NeuralCDE standard form is used to describe the continuous evolution of latent variables over time, which is expressed as: Where z(t) represents the latent variable at time t, f represents the neural network differential function, α1 represents the function parameter, U(s) represents the control path constructed by discrete observations, and s is the integral variable; The control path U(t) is defined by the interpolation basis function and is expressed as: Among them, N1 represents the total number of observation points, represents the i-th observation input, φ i (t) represents the corresponding interpolation basis function.

6. The method for assessing adolescent cognitive ability and psychological development based on artificial intelligence according to claim 1, characterized in that: The method for detecting the phase mutation event of the cognitive strategy in step S3 is: Construct the Vietoris-Rips complex in the cognitive feature manifold space and calculate the persistence graph entropy of the kth dimension. The formula is: Among them, H (k) (t) represents the kth maintained long-term entropy at time t, N k (t) represents the number of continuous intervals in the kth dimension, k represents the dimension index of the homology group, represents the length of the i-th interval, Indicates the corresponding normalized weight; For a one-dimensional ring with k=1, the change of its persistent entropy over time is monitored, and the average persistent entropy is calculated by sliding the time window: Among them, τ1 represents the length of the time window. If And it continues, δ1 represents the persistent entropy threshold for determining phase mutation, then it is determined to be a phase mutation and the model incremental learning mechanism is triggered.

7. An artificial intelligence-based system for assessing the cognitive ability and psychological development of adolescents, based on an artificial intelligence-based method for assessing the cognitive ability and psychological development of adolescents according to any one of claims 1 to 6, characterized in that: include: Multimodal data acquisition module, integrating flexible electronic skin sensors and adaptive touch interface; Edge computing unit, deploying lightweight time series feature extraction network and privacy protection mechanism; Dynamic assessment path generator, which adjusts task types and difficulty parameters according to real-time cognitive status; Cognitive development visualization engine generates three-dimensional spatiotemporal evolution maps and intervention recommendations.

8. The artificial intelligence-based adolescent cognitive ability and psychological development assessment system according to claim 7, characterized in that: The privacy protection mechanism is specifically implemented as follows: A federated learning architecture is used for distributed model training, with the original data retained on the local terminal. Add controllable noise to cognitive feature vectors through differential privacy technology; Use homomorphic encryption to transmit sensitive intermediate data during the fingerprint matching process.

9. The artificial intelligence-based adolescent cognitive ability and psychological development assessment system according to claim 7, characterized in that: The dynamic evaluation path generator further includes: A strategy fingerprint feature library stores a feature transfer matrix of typical cognitive development patterns. This feature transfer matrix is constructed by calculating the cosine similarity variance of cross-task operational features and includes a 52-dimensional transfer relationship map between three types of tasks: risk decision-making, pattern recognition, and social simulation. Online Bayesian optimizer, dynamically selecting evaluation dimensions based on the sparsity of the current cognitive space; The sudden jump event response module inserts a cross-domain verification task when a sudden phase change is detected.

10. The artificial intelligence-based adolescent cognitive ability and psychological development assessment system according to claim 7, characterized in that: The visualization engine includes: a cultural adaptability calibration layer to eliminate the bias in autonomy ratings caused by the collectivist decision-making model; Developmental trajectory prediction interface, simulating the branching paths of cognitive evolution under different intervention measures; Hidden risk warning unit detects early psychological development deviations based on abnormal strategy migration.

Images