44 results about "Stochastic differential equation" patented technology

The invention discloses an engine state estimation and system modeling correction method based on double-layer variational inference, which relates to the field of engine state estimation and comprises a variational inference stage aiming at component performance states and kinetic model parameters; a system output prediction stage based on an observation equation; and a solving stage of performing objective function optimization through an evidence lower bound. The structure clearly presents information flow and key calculation links of the proposed algorithm in state estimation and model learning. According to the method, combined reasoning of state variables and model parameters is achieved by building a probability modeling structure, the modeling problem when system dynamics is partially or completely unknown is solved by combining a modeling method of a stochastic differential equation, and while the state variables and the model parameters are optimized, the modeling efficiency is improved. Precise inference of component states and reliable identification of fault features are achieved, the fault detection accuracy of sudden gas circuit abnormity reaches the standard, and meanwhile the performance is better in the aspect of tracking long-term performance degradation.

The invention provides a hockey team tactical position early warning and correction method and system fusing spatio-temporal features, and relates to the technical field of hockey tactical analysis, and the method comprises the steps: generating tactical feature vectors through employing a self-adaptive diffusion probability graph algorithm, and comparing the tactical feature vectors with a standard tactical library to generate an early warning; calculating position deviation according to tactical rules; a stochastic differential equation is constructed, a plurality of groups of candidate trajectories are generated, and the trajectory with the minimum energy value is selected as the correction path, so that real-time early warning and scientific correction of the hockey tactics are realized, and the tactical execution accuracy is improved.

The invention discloses a refrigerator life prediction method based on an adaptive physical information recurrent neural network, and the method comprises the steps: extracting multi-source statistical health features through a steady-state control window, capturing time sequence dependence through a recurrent neural network, and generating a degeneration state estimator; embedding the estimator into a stochastic differential equation driven by a Wiener stochastic process to enable a drift term to reflect a deterministic degradation mechanism and a diffusion term to quantify the uncertainty of a working condition; based on data deviation, equation residual and monotonicity violation, multi-constraint joint loss is constructed, Bayesian uncertainty estimation is adopted to dynamically optimize each loss weight, and adaptive balance of physical consistency and observation fitting is realized. According to the method, residual life prediction with probability distribution is output, and the precision and robustness of refrigerator health management under complex working conditions are effectively improved.

The invention relates to the technical field of carbon emission prediction and early warning, and provides a multi-scale neural distribution prediction and hierarchical migration early warning method, which comprises the following steps: acquiring historical carbon emission data, respectively inputting a historical sequence and a to-be-predicted sequence into an energy consumption stochastic differential equation model and a carbon factor stochastic differential equation model, generating a multi-scale carbon emission path sample set through an independent random disturbance term; calculating a path-level suitability score of each path sample based on the standard-exceeding risk integral, the first standard-exceeding moment and the path fluctuation variance; layering the calibration data set into a plurality of working condition layers according to working condition labels, sharing distribution shape parameters among the working condition layers through a hierarchical Bayesian method, and regularizing quantiles of small sample working condition layers to obtain an early warning threshold value of each working condition layer; and selecting a corresponding early warning threshold value according to the current working condition label to compare and trigger early warning. According to the method, the accuracy of carbon emission distribution prediction and the robustness of an early warning system are improved, and the problem that the early warning threshold value is unstable under the small sample working condition is relieved.

The invention discloses a random stability analysis method for an electromechanical composite transmission system, and belongs to the technical field of electromechanical transmission. The method comprises the following steps: acquiring operation data of the electromechanical composite transmission system, and estimating drift term parameters and diffusion term parameters of an Irat stochastic differential equation representing a random road load by adopting a maximum likelihood method; establishing a random state space model according to the current balance relationship, and judging the random stability of the model by adopting a moment stability analysis method to obtain a system stability analysis result; and performing numerical simulation according to the random state space model, verifying the accuracy of the system stability analysis result, and outputting a random stability verification conclusion. According to the method, random excitation can be processed by estimating the Italian process parameters based on the maximum likelihood method and adopting the moment stability analysis method, and the method has the advantages of being good in adaptability, high in calculation efficiency and high in stability.

This invention discloses a visual reconstruction method and system for remote sensing interpretation based on a generative diffusion model. The method includes: acquiring high-resolution and low-resolution remote sensing image data; adding different levels of Gaussian noise to the training data using a forward stochastic differential equation until pure Gaussian noise data is obtained; training a noise conditional scoring network to predict the scores corresponding to these noisy data; adding noise to the low-resolution image using a forward stochastic differential equation to finally obtain pure Gaussian noise; using a trained neural network to guide the random noise to gradually converge and generate a super-resolution remote sensing image; rapidly identifying land cover types on the generated remote sensing image; and delineating land cover patches on the original remote sensing image and assigning patch information based on the identified land cover categories. This invention achieves a super-resolution effect from low resolution without changing the land cover types and patch boundaries, thereby reducing interpretation costs and improving interpretation efficiency.

The invention discloses an investment fund distribution method based on value network updating and strategy network updating, and the method comprises the steps: obtaining market data of a plurality of financial assets, constructing the processed market data into a high-dimensional state vector representing a market environment, and enabling the state vector to obey a controlled stochastic differential equation, the equation depends on a current moment, a state vector and an asset weight distribution action, and the asset weight distribution action is generated by a strategy network; the total expected revenue which can be obtained in the future is modeled as an entropy regularization value function, the entropy regularization value function comprises an entropy regularization instantaneous return rate, and the entropy regularization instantaneous return rate comprises an external reward and an internal reward for encouraging exploration; the entropy regularization value function is updated and optimized through a weak yoke confrontation value network and a yoke residual strategy network, a strategy network used for asset configuration is obtained, then an asset weight distribution action is generated, and a transaction instruction is generated based on the action and sent to an automatic transaction execution system.

The invention provides a composite degraded image fusion method, system and device and a storage medium, and belongs to the technical field of image processing, and the method comprises the steps: mapping a multi-modal image to a potential space through a pre-trained automatic encoder; performing information recovery in the potential space by adopting a potential diffusion framework based on a mean regression stochastic differential equation to obtain clean potential features of the multi-modal image after degeneration is removed; and based on the clean potential features, performing feature fusion through a potential diffusion fusion module of a U-net architecture, and according to a feature fusion result, decoding to obtain a fused image. According to the method, the applicability of image fusion in an extreme composite degradation scene is effectively improved, on the basis, preference modulation of a specific target object in the image can be achieved according to a language instruction input by a user, and various visual and semantic requirements of the user are flexibly met.

The application provides a kind of based on knowledge and data fusion driven aero-engine life prediction method, belong to aero-engine life prediction technical field;The technical problem to be solved is to provide a kind of based on knowledge and data fusion driven aero-engine life prediction method;The technical scheme for solving the technical problem is that the overall architecture of encoder-decoder is used, the encoder extracts high-level feature representation from past observations, and the decoder summarizes past information and continuously optimizes the prediction result;A GRU-SDE module is designed in the encoder, and a stochastic differential equation SDE is used to simulate the engine degradation process, which more accurately simulates the uncertainty and randomness of degradation while extracting timing dependence;A knowledge graph is constructed for the inherent knowledge of the engine to obtain a knowledge-based multi-sensor relationship matrix;The data-based relationship matrix is randomly initialized, and the two are adaptively fused during the training process;The application is applied to aero-engine life prediction.

The invention discloses a brain simulation-oriented high-performance numerical differential solver method and system. The method comprises the following steps of: receiving a neuron cluster model and simulation parameters thereof; the differential equation definition of the model is analyzed, and when the input is in a character string form, the input is converted into a function form; setting a current moment t, a simulation step length dt and simulation time simt; the simulation parameters are analyzed and classified; judging whether the model contains noise or not, if yes, initializing a stochastic differential equation solver, and otherwise, initializing an ordinary differential equation solver; adaptive operators of different neuron models are called in each step length to calculate numerical solutions, and variables are subjected to parallel calculation and updating in the calculation process; updating the variable yn and the current moment t when tlt; and when t is greater than or equal to simt, entering the next step of length calculation until t is greater than or equal to simt. According to the method, the high-performance differential solver supporting parallel solving of the whole neuron cluster is provided, the parallel acceleration characteristic of hardware such as a GPU is fully utilized, the simulation speed is increased, and large-scale brain simulation is achieved.

The invention discloses a stochastic differential equation-based building method for connecting a physical rendering and diffusion model, which is characterized by comprising the following steps of: on the basis of a central limit theorem, converting discrete Monte Carlo integral into continuous time SDE expression; the method comprises the following steps: constructing a Monte Carlo stochastic differential equation, establishing mapping alignment of MC-SDE variance time tau and diffusion model time step t through uniform hemisphere sampling of a unified noise source and separation of diffuse reflection and specular reflection components, pointing out a denoising earlier stage corresponding to a low sampling path tracking image, and the like. Experiments prove that the method can apply physical control to a diffusion model generation result, the low-sampling path tracking image is input into the pre-training diffusion model to complete denoising, and a high-quality rendering result is generated. Compared with the prior art, the method has the advantages that the physical control capability of the PBR and the generation flexibility of the diffusion model are effectively combined, and the technical problems that the physical controllability of the diffusion model is insufficient, the physical rendering prompt driving flexibility is insufficient, and the two are difficult to fuse are solved.

The invention relates to a multi-target model prediction control method for a wind power access multi-terminal flexible direct current system. Comprising the following steps: constructing a grey box stochastic differential equation model based on wind power randomness and multi-terminal flexible direct current system operation data, and applying Lamperti transformation to the model to obtain an equivalent state space prediction model; establishing a multi-target optimization model considering wind power real-time tracking, DC voltage stabilization and multi-converter station collaborative power distribution in the prediction domain; setting a random stability constraint by taking moment stability as a criterion, and determining the probability constraint of a system state quantity and a control quantity into an equivalent deterministic constraint; and solving the multi-objective optimization model at each sampling moment based on the random stability constraint and the deterministic constraint, obtaining the optimal control quantity of the multiple converter stations, and applying and executing the optimal control quantity. Robust stability, real-time power tracking and multi-station cooperative distribution can be achieved under the conditions of wind power random disturbance and power electronic nonlinearity, and high control precision and real-time performance are achieved.

The invention relates to the technical field of artificial intelligence, can be applied to the fields of financial science and technology and medical science and technology, and discloses a speech synthesis method and device based on direct preference optimization, equipment and a medium. An ordinary differential equation sampling process is converted into an equivalent stochastic differential equation sampling process containing random noise introduced in each integral step, and a randomized model is obtained; obtaining an input text and a corresponding reference voice, and inputting the randomization model to output a plurality of candidate voice samples; grading and screening out preference samples and non-preference samples through a preset grading strategy, and constructing a preference data pair; based on a direct preference optimization algorithm, training the two to-be-trained randomized models by using the preference data pair to obtain a preference model and a non-preference model; and finally, receiving a to-be-synthesized text, synthesizing output results through the two models, and fusing to obtain a target voice. And model optimization is guided through preference data, so that the speech synthesis quality is remarkably improved.

The invention discloses a watershed runoff simulation method fusing a random Xinanjiang model and machine learning, and belongs to the technical field of watershed runoff simulation. The method comprises the following steps: collecting hydro-meteorological data in a research basin; constructing a stochastic three-water-source Xinanjiang model based on a stochastic differential equation; utilizing Monte Carlo simulation to generate a runoff probability trajectory and extracting statistical characteristics; extracting a multi-scale feature vector of the runoff sequence based on discrete wavelet transform; and constructing machine learning models with different architectures, inputting a mean trajectory or a full feature set subjected to wavelet decomposition, and simulating the drainage basin runoff by adopting different information combinations. According to the method, the three-water-source Xinanjiang model based on the stochastic differential equation is constructed, and noise reduction and multi-scale decomposition of the runoff random trajectory are realized in combination with Monte Carlo simulation and discrete wavelet transform, so that a machine learning model can more accurately capture multi-scale hydrological signals; and the runoff simulation effect of the model is improved by improving the low and high flow simulation precision of the drainage basin.

This invention relates to the field of edge computing technology, specifically disclosing a multi-task optimization scheduling method and system based on edge computing. The method involves: acquiring the original observation reward sequence; calculating its Hearst exponent and setting the state space of a stochastic differential equation, outputting the prior distribution of the implicit true reward; updating the posterior distribution through particle filtering, extracting the posterior mean and posterior variance, and calculating the effective diversity index; training a deep reinforcement learning strategy using the posterior mean instead of the original reward, and calculating the exploration adjustment factor by combining the posterior variance and the effective diversity index; periodically reestimating the parameters of the stochastic differential equation and feeding them back to the state space. This invention can stabilize the scheduling strategy under sparse and noisy reward conditions, achieving an adaptive balance between exploration and utilization.

The invention belongs to the technical field of stability analysis of a permanent magnet synchronous generator system, and aims to solve the problems that a current wind power system cannot avoid noise generation under complex operation conditions, but the research on the stability of the system is less for the noise in the PMSG system. According to the method for researching the random stability and the random bifurcation behavior of the permanent magnet synchronous generator system under the excitation of the color noise, disturbance of random factors such as airflow fluctuation on mechanical torque of the permanent magnet synchronous generator system is described by introducing the color noise, and then the permanent magnet synchronous generator system is established; an Itstochastic differential equation of the system is obtained through a color noise unified approximation principle, a center manifold theory, polar coordinate transformation and a random average method; the random stability and the random bifurcation behavior of the system are discussed according to a maximum Lyapunov index and a probability density function method. A numerical simulation result further verifies the effectiveness and reliability of the research method.

The invention relates to the technical field of short-wave communication, and discloses an ionosphere model-based short-wave positioning method, which comprises the following steps of: acquiring multi-site ionosphere observation data, and generating a spatial-temporal characteristic tensor and a noise covariance matrix; calculating a long-range dependency relationship of the spatial-temporal feature sequence by using a Transform encoder, and generating a global feature vector; and inputting the global feature vector into a neural stochastic differential equation network to generate a state-dependent diffusion coefficient function. According to the invention, the long-range dependence modeling capability of the time sequence Transform, the randomness description capability of the neural stochastic differential equation and the probability trajectory optimization capability of the flow matching algorithm are fused, so that the defect that the deterministic propagation rule and the random fluctuation characteristic of the ionospheric disturbance cannot be processed at the same time in the traditional method is overcome; the technical problem that short-wave positioning lacks reliability evaluation under the disturbance condition is solved.

The application discloses a kind of based on big data analysis's construction engineering progress risk prediction evaluation method, comprising: S1, synchronous acquisition construction engineering field multi-source heterogeneous big data, space-time alignment and construct five-order progress tensor space;S2, complete denoising and pass through Riemann manifold learning projection to low-dimensional feature space, calculate geodesic trajectory and dynamic risk potential field;S3, construct heterogeneous progress correlation diagram and map manifold characteristics to vertex attribute;S4, decouple noise and management factors using counterfactual learning mechanism, mine causal characteristics;S5, use the improved neural stochastic differential equation of introducing heterogeneous graph neural network drift term guide mechanism to construct differential equation, solve risk evolution trajectory;S6, matching knowledge graph generates comprehensive risk evaluation report.The application realizes the accurate prediction of risk evolution trend, improves the accuracy of complex engineering progress risk control.

The present invention provides a bankruptcy risk factor identification device and a bankruptcy risk factor identification method that identify bankruptcy risk factors based on the probability of bankruptcy calculated using stochastic differential equations. [Solution] The bankruptcy risk factor identification device according to the present invention comprises: a data storage unit that stores accounting data of a target company for a predetermined period provided by an accounting system; a variable parameter calculation unit that calculates the drift and volatility of the target company's assets and liabilities using the accounting data for the predetermined period; a bankruptcy probability calculation unit that calculates the bankruptcy probability by applying the calculated drift and volatility of the assets and liabilities to a stochastic differential equation that represents the target company's net assets in a stochastic process; a loop operation instruction unit that changes the value of the drift and / or volatility of any of the assets and liabilities applied to the stochastic differential equation by a predetermined step size and transmits it to the bankruptcy probability calculation unit, causing the bankruptcy probability calculation unit to repeatedly recalculate the bankruptcy probability; and a risk factor determination unit that identifies the drift and / or volatility value at a position where the bankruptcy probability exceeds a predetermined threshold in a probability distribution plotted against the changed drift and / or volatility value as a bankruptcy risk factor for the target company.

The invention relates to the technical field of power equipment management, and provides a wire and cable delivery tray automatic selection calculation method and system, and the method comprises the steps: employing a multi-scale cavity convolutional neural network to extract cable layered structure features; carrying out three-dimensional to two-dimensional disk loading mapping calculation based on a spiral constraint graph neural network; adopting a multi-time-scale long-short-term memory network to predict delivery tray performance degradation; a multi-agent reinforcement learning network is adopted to carry out inventory scheduling optimization; performing dynamic stability calculation through a stochastic differential equation reinforcement learning algorithm; and carrying out adaptive optimization by adopting a Bayesian element learning algorithm, and outputting an optimal use scheme. The selection strategy is adaptively reconstructed according to the stability evaluation result, and the accuracy of selection calculation of the wire and cable delivery tray in the marine environment is improved.

The invention discloses a wireless communication power control method based on a fractional calculus theory, which comprises the following steps of: firstly, constructing a continuous time stochastic differential equation channel model driven by a symmetric-stable Levy process, and respectively depicting heavy tail jump behaviors in long-term path loss and short-term multipath fading; then, on the basis of the channel model, a generalized Riesz fractional order derivative operator is introduced, a class of fractional order HJB (FHJB) control equations with non-local operators are deduced under the dynamic programming principle, and the fractional order HJB (FHJB) control equations are used for describing an optimal power control strategy under the infinite variance interference condition; and finally, in a multi-base-station multi-user same-frequency interference scene, constructing a fractional order FHJB power control solution framework, obtaining an optimal transmitting power distribution strategy by using a value function iteration method, and verifying the robustness and communication quality assurance capability of the optimal transmitting power distribution strategy under a jump fading channel through numerical simulation. According to the invention, adaptive control of the transmitting power in a non-stationary and heavy-tailed fading environment can be realized.

According to a diffusion-based training optimization device and method for a smart vision inspection system proposed in the present invention, the device comprises an optimizer for performing stochastic optimization to estimate parameters that minimize an objective function in training of an artificial intelligence model. The optimizer may convert a stochastic optimization problem into a sampling problem on the basis that an invariant measure of a Langevin SDE includes a global solution of the objective function, and by applying a sampling algorithm, the optimizer may apply an optimization algorithm for which global convergence is theoretically proven, and can ensure sufficiently high accuracy by training the artificial intelligence model with time and computational resources usable even for a highly complex objective function.

The invention provides an electric power data virtual acquisition method based on a stochastic differential equation, and the method constructs a diffusion model network structure with a continuous time distribution modeling capability aiming at the problem that data missing is easy to occur in a new energy load data acquisition process. The method comprises the following steps: firstly, masking an input sequence through a random mask strategy in a training process so as to simulate a potential data missing scene, and carrying out supervised training on observed data; and then accurate virtual acquisition of missing values is realized through learning conditional probability distribution. On the basis, a stochastic differential equation (SDE) is adopted, noise is slowly injected, complex data distribution is smoothly converted into known prior distribution, and reverse time SDE of a time correlation gradient field which only depends on disturbance data distribution is adopted, and the prior distribution is converted into data distribution by slowly removing noise. Experimental verification on two real world data sets shows that the method is superior to an existing method in the aspect of virtual acquisition accuracy, and the superiority of the method is proved.

This application relates to a method, apparatus, and computer equipment for determining wind speed forecasts along high-speed railway lines. The method includes: collecting historical monitoring data from meteorological stations along the high-speed railway line and normalizing the historical monitoring data to determine normalized data; determining the target decomposition level of a multivariate variational mode decomposition algorithm (MMD), and performing MMD on the normalized data based on the target decomposition level to determine the target intrinsic mode function matrix; dynamically modeling the kernel parameters of the radial basis functions based on stochastic differential equations to determine the dynamic kernel function; constructing a wind speed prediction model; using the wind speed prediction model to perform variational inference on the characteristic data of each intrinsic mode component based on the normalized data to determine the point-to-point and interval-to-interval wind speed prediction results along the high-speed railway line; constructing a multi-step prediction interval using a multi-step prediction strategy and determining early warning information. The above scheme improves the accuracy of point-to-point wind speed prediction and the stability of interval-to-interval prediction.

This invention relates to the fields of marine meteorology and navigation, and discloses a method for predicting meteorological factors for ship navigation based on marine meteorological information. This method is used to predict meteorological factors for ship navigation, assess navigation risks, and plan alternative routes. The method includes processing meteorological data and ship data for a target sea area to obtain historical meteorological time series and ship motion response sequences; constructing state vector trajectories of meteorological factors using historical meteorological data, establishing a dynamic model library, and obtaining the ship motion response function to meteorology based on meteorological and motion response data; generating deterministic meteorological factor prediction sequences by embedding the current meteorological state into the dynamic model; and constructing stochastic differential equations to generate probabilistic prediction distributions of meteorological factors. This invention integrates data-driven and physical constraints, providing more comprehensive navigation decision support.

The application discloses a kind of unmanned aerial vehicle cluster target joint tracking method, first, according to the three interactive rules of unmanned aerial vehicle cluster separation, aggregation and adjustment, the cooperative movement of unmanned aerial vehicle cluster is effectively described using stochastic differential equation;Second, based on the definition of new distance measure of target motion characteristics, the division of cluster in field of view is realized using DBSCAN clustering algorithm based on the measure;Finally, in the framework of Bayesian filtering, DBSCAN clustering algorithm is combined with JPDA algorithm, and the cluster division and target joint state estimation in clutter environment are realized simultaneously;In simulation experiment, the behavior of cluster is simulated and state estimation is carried out, and the results show that the algorithm can effectively describe the interactive motion process of unmanned aerial vehicle cluster target, and the tracking algorithm can effectively track the joint tracking of cluster target, which proves that the method has the characteristics of good tracking and classification effect.

This invention relates to a method for cross-domain migration diagnosis of aero-engine test faults based on continuous diffusion. Existing diagnostic models suffer from low accuracy and insufficient reliability under varying operating conditions due to inconsistencies in feature distributions between the source and target domains. Aero-engine test signals are severely affected by environmental interference, and large differences in cross-domain features result in weak generalization ability of traditional models, making them prone to misdiagnosis. To address this issue, this invention proposes a fractional gradient modeling method based on continuous diffusion. First, forward diffusion of stochastic differential equations is used to construct a global fractional network model of the source and target domains. Then, dual-path gradient modeling is performed on the signal, and the cross-domain divergence degree is calculated. Finally, the diagnostic results and confidence evaluation are output. This method can effectively extract long-range time-series features and achieve distribution alignment, improving diagnostic accuracy under varying operating conditions. This invention is applied to the cross-domain migration diagnosis of aero-engine test faults.

The invention relates to the technical field of mental health assessment, in particular to a mental electrophysiology assessment method based on a video follow-up visit system, which comprises the following steps of: acquiring doctor-patient double-channel audio and video data through the video follow-up visit system; preprocessing the collected data and extracting multi-modal features including facial behavior features, voice acoustic features and motion dynamics features; and constructing a spatial model of the mental pathological state of the patient based on the extracted multi-modal features, wherein the spatial model describes a dynamic evolution process of a multi-dimensional state variable through a stochastic differential equation. According to the method, a multi-modal feature extraction system of doctor-patient dual-channel audio and video data is constructed, and then a psychiatric pathology state space model based on a stochastic differential equation is established, so that the problems that subjective scale and static observation are mostly adopted in a traditional mental assessment method, and due to lack of objective quantitative indexes and a dynamic tracking mechanism, the accuracy of mental assessment is poor are solved. Therefore, the evaluation result is greatly influenced by subjective experience of doctors, and the dynamic change of symptoms cannot be captured.

The invention discloses a protein generation method, device and equipment based on a flow matching Transform diffusion model, and relates to the field of biomedicine. The method comprises the steps of performing data conversion on to-be-combined electron density to obtain continuous features corresponding to the to-be-combined electron density, and partitioning the continuous features corresponding to the to-be-combined electron density based on a preset partitioning condition to obtain input lexical element features; inputting the time step information, a preset diffusion condition and the input lexical element features into a stream matching Transform diffusion model to obtain a velocity field prediction result; and according to the diffusion path trajectory and a reverse stochastic differential equation of a flow matching Transform diffusion model, sampling and denoising processing is carried out on a velocity field prediction result to obtain binding protein corresponding to the to-be-bound electron density. The flow matching Transform diffusion model uses the electron density information of the target protein to guide the transformation of the specific flow, so that the initial Gaussian noise is transformed to the electron density of the binding protein possibly bound with the target protein.

The application provides an AIGC content self-adaptive method and system based on user feedback, comprising: acquiring and preprocessing a multi-modal user feedback dataset, event modeling and causal discovery, acquiring and performing feedback event quality evaluation according to a first feedback event dataset and a causal influence dataset, generating a second feedback event dataset, synchronously acquiring event exploration intensity, performing counterfactual reasoning based on a preset neural jump differential equation and the causal influence dataset, acquiring a corrected event exploration intensity and a third feedback event dataset, performing AIGC self-adaptive generation based on a preset multi-scale stochastic differential equation, synchronously combining the second and third feedback event datasets and the corrected event exploration intensity, acquiring an original AIGC content dataset, performing quality evaluation on the original AIGC content dataset, and combining a Bayesian optimization algorithm to perform optimization fine-tuning to generate an optimized AIGC content dataset, thereby improving the accuracy of the generated content.