350 results about "Adaptive strategies" patented technology

The invention relates to the technical field of data security management, in particular to a self-adaptive data security management and risk early warning system based on intelligent analysis under a cloud platform. Comprising a multi-dimensional data acquisition module; an intelligent analysis module; a self-adaptive strategy generation module; a risk early warning module; and a user behavior portrait construction module. In the design, the security policy can be dynamically adjusted along with the risk situation of the cloud platform, the problem that a static policy cannot adapt to real-time change is solved, and dynamic mapping of risk characteristics-policy parameters is realized; according to the design, the one-sidedness of single-dimension analysis is broken through, multi-modal feature association modeling of user behaviors is achieved, an abnormal behavior triggering threshold value is accurately recognized, and the integrity and accuracy of risk feature analysis are improved; the security policy can be continuously optimized through historical event data, so that protection efficiency attenuation caused by long-term static operation is avoided, and an autonomous lifting link of data driving, algorithm optimization and policy evolution is realized.

A system for UAV collaborative coverage path planning based on a Q-learning adaptive ant colony algorithm including a memory, an image collection device, and a plurality of UAVs loaded with a path planning module configured to: construct a 3D model in a collaborative coverage environment, by performing a cell division on the 3D model based on a scanning range of an airborne radar of each UAV, obtain one or more sub-regions; by establishing constraints of the UAV and the environment based on the determined 3D model of the region to be searched, establish a problem total cost model; perform a plurality of rounds of iterations, calculate a reward value of each ant colony and determine whether a maximum iteration count is reached, if the maximum iteration count is reached, enter a new round of iteration, otherwise, output a path corresponding to a current round of iteration as a final path.

The invention provides a software supply chain risk detection protection method and system, and the method comprises the steps: obtaining a third-party component from a software supply chain, and analyzing the multi-level dynamic behaviors of the third-party component; predicting a risk behavior pattern by using a Bayesian network algorithm, and constructing a behavior pattern library; monitoring a behavior of the third-party component during operation, and recording a behavior data stream; identifying an abnormal behavior deviating from a normal behavior mode by applying a time sequence anomaly detection algorithm of deep learning; evaluating risk levels of the abnormal behaviors by using a graph neural network, and generating an abnormal behavior list in combination with relevance and propagation paths among the behaviors; according to the list, a response strategy in the safety control measure library is dynamically adjusted through an adaptive strategy selection algorithm, and safety control measures are formulated for the third-party component. According to the invention, the identification and prediction capabilities of potential risk behaviors are improved.

The invention discloses a data center machine room AI energy-saving control method and system, a digital twin model of a machine room operation state is constructed through a holographic perception and heterogeneous data fusion technology, centimeter-level monitoring of an equipment state and environmental parameters is realized, and the system integrates a laser radar array, an acoustic sensor and a gas sensor network. The time-space alignment of multi-modal data is completed by combining edge computing nodes, holographic mapping including thermodynamic characteristics, vibration characteristics and gas leakage risks is formed, historical temperature control strategy characteristics are extracted by adopting a variational auto-encoder based on a dynamic strategy generation mechanism of generative artificial intelligence, and a load trend is predicted by combining a long-short-term memory network. Constructing a self-adaptive strategy pool; the multi-agent reinforcement learning framework enables temperature control, equipment scheduling and power grid response to form game optimization, the strategy robustness in a complex scene is improved, and the system innovatively fuses power grid real-time electricity price and carbon transaction data so as to establish a multi-target decision system.

The invention relates to the technical field of elevator equipment state monitoring and fault diagnosis, and discloses an elevator traction machine running state real-time diagnosis method and system based on high-frequency sampling. According to the method, vibration (larger than or equal to 20 kHz), current (larger than or equal to 10 kHz), sound / sound emission, temperature and rotating speed signals of a traction machine are synchronously collected through a high-frequency multi-mode sensor array; capturing early weak fault transient characteristics; the edge computing unit completes data preprocessing, time synchronization, feature extraction and anomaly detection, and uploads key data to a cloud end through cloud-edge collaboration; the cloud end adopts a working condition self-adaptive strategy and a multi-modal fusion model to carry out deep diagnosis, and outputs fault types, positions and grades; and combining incremental learning and a degradation model to realize health quantification and residual life prediction. Through fusion of high-frequency data capture and an intelligent algorithm, the early fault detection capability, variable working condition adaptability and diagnosis real-time performance of the traction machine are improved, and a solution is provided for predictive maintenance of an elevator.

The invention discloses a reinforcement learning-based virtual power plant response optimization scheduling system and method, and relates to the technical field of virtual power plant intelligent scheduling. The system comprises an environment modeling module, an intelligent agent module, a multi-agent coordination module and a self-adaptive optimization module which are respectively used for constructing a multi-dimensional state space and a layered action space, generating and optimizing an action strategy based on an Actor-Critic network, executing a scheduling instruction through a layered multi-agent structure and realizing conflict consensus, and dynamically adapting to state space change in combination with incremental learning and meta-learning mechanisms. The system and the method have the advantages of fine state modeling, efficient action response, adaptive strategy updating, stable agent coordination and the like, and can keep the continuity, the stability and the optimality of a scheduling strategy in an operation environment in which multi-source heterogeneous power resources participate in scheduling cooperatively, market rules change frequently and load fluctuation is violent.

The invention discloses a unified multi-modal alignment method and system based on mixed experts and low-rank adaptation, and belongs to the field of artificial intelligence and multi-modal learning. According to the method, firstly, a CLIP model serves as a basic model, and a unified modal encoder is constructed in combination with a modal perception hybrid expert strategy and a low-rank adaptive strategy; the unified modal encoder comprises an anchor modal marker corresponding to each anchor modal, an extended modal marker corresponding to each extended modal and a plurality of stacked multi-modal hybrid expert modules; then, on the basis of universal modal alignment measurement and knowledge distillation and cross-modal optimization strategies, multi-modal data samples are sampled in batches, and fine tuning training is conducted on the unified modal encoder. Through a single encoder and single training, alignment of any number of anchor modes and extension modes is realized, a special model does not need to be trained for each mode independently, the transferability of cross-domain and downstream tasks can be improved, and the method is suitable for tasks such as multi-mode understanding, generation and reasoning.

The embodiment of the invention provides an intelligent agent memory indexing method and system based on intention recognition. The method is applied to the technical field of artificial intelligence and comprises the steps of obtaining real-time question-answer data, and performing preliminary intention classification on the real-time question-answer data by utilizing a domain knowledge rule library; extracting a structured description from the real-time question and answer data after the preliminary intention classification, performing deep intention analysis in stages, and outputting a standardized intention description text; according to the standardized intention description text, acquiring an Agent operation context, performing multi-dimensional retrieval to obtain an adaptive strategy, executing the adaptive strategy, and returning a strategy evaluation result; according to a strategy evaluation result, carrying out microscopic feedback and macroscopic feedback to update a strategy library; the Agent operation context is obtained through the following steps that semantic features of a standardized intention description text are captured, and the Agent operation context corresponding to the deep semantic features is recorded based on a fine-grained metadata labeling system. According to the invention, a complete closed loop from intention identification to strategy multiplexing to strategy optimization is realized.

The invention relates to the technical field of artificial intelligence, can be applied to business scenes such as financial science and technology and medical health, and discloses a generation strategy optimization method, device and equipment based on a dynamic environment and a medium. Correcting the generated action vector by combining a domain constraint strategy to obtain a compliant action vector, constructing a multi-dimensional reward vector according to feedback after execution, scaling the reward vector into a reward signal, and finally updating the pre-training generative model by adopting a self-adaptive strategy optimization module based on the reward signal and an interaction track to obtain a reward result. And collaborative evolution of strategy generation and environmental response is realized. According to the method, by introducing dynamic environment information and domain constraints, compliance correction and optimization updating of the generative strategy are realized, and the stability and practicability of the model in a complex environment are improved.

The invention discloses a multi-satellite task scheduling method based on a genetic algorithm, and belongs to the technical field of satellite task scheduling. Establishing a scheduling model based on satellite and ground station visible window data, and integrating a scheduling period constraint, a task uniqueness constraint, an equipment protection time constraint and a frequency band and orbit type matching constraint by taking task income maximization as a target; a simulated annealing local search mechanism is introduced into the genetic algorithm, local optimum is jumped out through a probability acceptance inferior solution, and a self-adaptive strategy of mutation probability is dynamically adjusted according to population fitness change; setting an early stop mechanism to terminate iteration in advance when the fitness is continuously not improved; and finally generating a scheduling scheme. According to the method, the problems that a traditional genetic algorithm is prone to falling into local optimum and constraint processing is rigid are solved.

The invention discloses a data quality treatment method and system based on an AI Agent, and belongs to the technical field of artificial intelligence and data treatment. An initial data semantic distribution map is constructed, cross-dimension correlation feature factors are extracted, a multi-scale quality anomaly sensitive factor matrix is constructed, time sequence evolution weights are embedded, and a dynamic feature evolution trajectory is formed; performing perception modeling on the evolution trajectory by using an AI Agent, generating a multi-level quality risk thermodynamic diagram, extracting a deviation dense region and constructing an anomaly propagation path set; in combination with upstream and downstream data links and task flow information, calculating a potential impact factor weight, constructing a causal traceability map, and injecting a correction strategy label for a key field node; the AI Agent autonomously selects an adaptive strategy combination according to the target data segment, and performs online intervention on the target data segment; according to the method, closed-loop treatment of the data quality problem from perception and judgment to intervention and feedback is realized, and the method has the advantages of self-adaption, high interpretability and the like.

The invention relates to the technical field of agent path planning, in particular to a multi-agent collision-free path planning method based on a fusion DQN algorithm. The multi-agent collision-free path planning method comprises the following steps: firstly, constructing a two-dimensional grid map as an environment, and carrying out feature extraction by utilizing a CNN (Convolutional Neural Network); then, behavior clone learning is carried out through the expert model to obtain a BC model; the core innovation lies in that a BC model and a CNNDQN model are fused, an adaptive strategy learning framework is constructed, and intelligent dynamic combination of expert experience and reinforcement learning exploration is realized by adopting uncertainty estimation, antagonistic knowledge distillation and performance perception sampling technologies; and finally, further processing an initial path output by the fusion model by a CBS algorithm, and completing multi-agent collision-free path planning. According to the method, the accuracy and efficiency of path planning are optimized through a mixed learning strategy.

The invention relates to artificial intelligence generation content AIGC, in particular to an AIGC test case self-adaptive generation system based on risk feedback, which comprises a semantic variation generation module for selecting a variation operator from a variation operator library according to a self-adaptive strategy, performing variation processing on a seed case provided by a basic test case library by using the variation operator, and generating a semantic variation result; generating a new test case; the AIGC model takes the test case as model input and performs model output to the risk assessment and quantification module through an AIGC model interface; the risk assessment and quantification module is used for performing multi-dimensional risk assessment on a model output result by utilizing the risk detection model library, calculating a risk level and generating a test report; the risk feedback processing module is used for storing the high-risk test case, performing feature extraction and converting a risk assessment result into an operable risk feedback signal according to extracted features; according to the technical scheme provided by the invention, the defects of low efficiency and difficulty in automatically adapting to environment change can be overcome.

The invention discloses a logistics management system based on artificial intelligence, and the system is used for improving the efficiency and intelligent level of logistics scheduling and resource configuration, and comprises the following steps: building a unified data set through collecting and positioning, sensor and operation log and other types of logistics data; a modeling method is used for fusion to form a data structure containing time-space and transportation state information, the system adopts a causal analysis model to identify key influence relations between logistics events, a cross-modal feature extraction technology is combined to generate representation vectors such as transportation tracks, loads and states, and the logistics events are subjected to self-adaptive analysis through a self-adaptive strategy. And the causal information and the multi-modal features are fused to generate a decision basis, and the fusion proportion is dynamically adjusted according to actual operation feedback, so that transportation path optimization, distribution task planning and reasonable resource allocation are realized. The system has a continuous learning capability, and can keep the accuracy and flexibility of scheduling decisions in a continuously updated data environment.

The invention belongs to the technical field of image processing. According to the target detection method and system in the dark dusty environment, adaptive improvement processing is carried out on an original model for a dark dusty scene, a target in the dark dusty environment can be rapidly identified and detected through an environment adaptive strategy and an attention mechanism, and the target detection efficiency is improved. The accuracy of target detection is greatly improved; according to the method, a multi-layer detection network is constructed, a target detection network is divided into a plurality of adaptation layers, namely a low-illumination adaptation layer, a dust adaptation layer, a comprehensive adaptation layer and a standard processing layer according to different environment brightness and dust concentration, adaptive processing can be performed according to different environment conditions, and the image processing efficiency and precision are greatly improved.

The invention provides a traffic control protection method, device and equipment based on congestion prediction, and aims to accurately identify a congestion rule and predict a propagation path by constructing a feature vector set based on historical data and a congestion propagation path prediction mechanism so as to realize timely release of congestion early warning information. Risk assessment grading protection and optimal intervention opportunity dynamic optimization are adopted, the protection opportunity is accurately mastered, and corresponding protection intensity is matched; a regional risk redistribution strategy and a protection effect gradient analysis technology are introduced, risk transfer control among multiple regions is coordinated, and disordered diffusion of congestion risks is avoided; in combination with cross-space-time effect tracing analysis and a strategy conflict detection mechanism, the actual contribution degree of each protection element is evaluated, and a weight adaptive strategy set and a protection capability matrix are constructed; finally, multiple links such as prediction, evaluation and coordination are fused to form a self-adaptive comprehensive protection system, and active prevention and intelligent management and control of urban traffic congestion are realized.

The invention relates to the technical field of power system automation, discloses a remote operation and maintenance method, system and equipment for wide-area substation automation equipment, and aims to solve the problems of data islanding, inaccurate diagnosis, strategy stiffness, low safety cooperation efficiency and the like in remote operation and maintenance of the wide-area substation automation equipment. According to the method and the system, a comprehensive platform is constructed, and multi-source heterogeneous data acquisition, intelligent fusion analysis, fault diagnosis and prediction, adaptive strategy generation, security remote control and block chain evidence storage and collaborative operation and maintenance are integrated. The method can solve the problem of data islands, improve the diagnosis and prediction precision and timeliness, optimize the operation and maintenance strategy, prolong the service life of equipment, and enhance the remote operation and maintenance safety and cooperation efficiency.

The invention belongs to the field of robot and computer vision, and particularly relates to a visual SLAM (Simultaneous Localization and Mapping) back-end optimization technology based on a data self-adaptive iterative reweighted robust kernel. Comprising the following steps: S1, a front-end data reading and processing module: a robot carries a camera to collect an image, performs feature extraction and feature matching on the image, and generates an initial camera pose and map points; s2, a rear-end optimization module: according to the initial camera pose and map point generated at the front end, calculating a median absolute deviation in an adaptive strategy, calculating a robustness parameter, then substituting the median absolute deviation and the robustness parameter into a target function, simplifying the target function by using an iterative reweighting strategy, and finally solving the optimal camera pose and map point by using an L-M algorithm; s3, the loopback detection module detects loopback according to whether the robot reaches a previous position or not, and transmits information to a rear end for processing; and S4, constructing a map module: constructing an environment map according to the optimized camera pose and map points.

A man-machine collaborative decision-making planning system and method based on environment adaptive trajectory optimization comprises a man-machine interaction module, a rule-based motion planning module and a learning-based parameter generation module, the man-machine interaction module performs data processing such as environment perception and state estimation according to data collected by an onboard sensor and a remote controller, and the parameter generation module generates parameter parameters according to the data. Input information required by the motion planning and parameter generation module is obtained; the motion planning module performs topological path search, visibility detection and trajectory optimization processing according to the local map and the user instruction information to obtain a planned trajectory of unmanned aerial vehicle tracking control; and the parameter generation module performs environment adaptive strategy processing according to the distribution information of the trajectory in the environment to obtain a speed constraint parameter for adjusting motion planning. According to the method, semantic information and depth information in environmental perception are considered, the speed parameters of the planner are automatically and dynamically adjusted, and the adaptability of the planned trajectory to different scenes is effectively improved, so that the cognitive and operation burden of a pilot in a complex environment is reduced, and the navigation efficiency and the system safety are improved.

The invention discloses a method and system for predicting juvenile depression based on intestinal flora, and relates to the technical field of bioinformatics and artificial intelligence, and the method comprises the steps: firstly, obtaining an original sequence of a microbiome, carrying out the preprocessing of the original sequence of the microbiome, and obtaining a feature matrix; and screening core flora characteristics with stable trans-folding by adopting characteristic importance evaluation and interpretability analysis based on a gradient boosting decision tree. A mixed weighted graph is constructed based on Spearman correlation and a proximity relationship, and an absolute value of a correlation coefficient is taken as an edge weight and an edge density is adjusted through a threshold adaptive strategy. And finally, through an improved graph attention neural network, based on edge weight attention, layer normalization and random inactivation, enhancing robustness, and adopting adaptive optimization to complete parameter learning. And determining a dynamic classification threshold according to the AUC of the target patient, and outputting a sample discrimination result and confidence. According to the method, the accuracy, stability and biological interpretability of juvenile depression recognition are remarkably improved.

The invention relates to the technical field of electric power data analysis, in particular to an electric power data analysis method and device based on graph calculation, and the method comprises the steps: constructing an electric power network graph, carrying out the deep learning processing of multi-dimensional and heterogeneous data, and analyzing a dynamic change prediction trend; power dispatching is optimized in combination with reinforcement learning, an optimal power dispatching decision is matched with minimization of energy loss as a target, the technical problems that high-dimensional data and heterogeneous data are complex to process, the change trend of the power system cannot be accurately predicted and an effective dispatching decision is lacked in the operation process of the power system are solved, and the power system can be effectively dispatched according to the real-time power grid operation state and the prediction trend. According to the method, the aim of self-adaptive strategy optimization learning is achieved, power dispatching flexibility is greatly improved, energy loss minimization is taken as the aim, an optimal power dispatching decision is accurately matched through continuous iterative optimization, power system data analysis efficiency is improved, and strategies are adjusted to cope with continuously changing supply and demand conditions and power grid operation conditions; and the overall stability and reliability of the power grid are improved to the greatest extent.

The invention relates to the technical field of thermal management optimization, and discloses an active cooling and power optimization system of a plateau turbine generator, which comprises a multi-physics field modeling module, a reinforcement learning decision module, a strategy optimization training module and an edge execution control module, and is used for resolving a temperature field / flow field / stress field coupling state in real time, providing environment dynamic characterization, and optimizing the power of the plateau turbine generator. According to the method, multi-source sensing data is processed, a cooling liquid flow rate and cooling fin angle control instruction is generated, strategy iteration optimization is carried out in a digital twin environment, the physical rationality and energy efficiency optimality of a control algorithm are ensured, a lightweight strategy network is deployed, and sub-second-level real-time response and fault safety protection are achieved. Through intelligent control and multi-physical field cooperation, the temperature adjusting precision and the system response speed are improved, efficient power output of the turbine is kept under different altitudes and climate conditions through a self-adaptive strategy, and a reliable clean energy solution is provided for high-altitude areas in combination with multiple safety protection mechanisms.

The invention discloses a thermal power generating unit decommissioning path planning method and device based on meteorological and hydrological risks, equipment, a medium and a product, and relates to the field of power system decision optimization. The method comprises the steps of obtaining meteorological and hydrological risk data and a decommissioning index of a thermal power generating unit; according to the decommissioning index and the meteorological and hydrological risk data under the climate change, a progressive hierarchical screening logic method is adopted for screening and constraint progressive correction processing; performing mapping and nonlinear correlation interaction analysis by adopting Monte Carlo simulation and a gradient boosting decision tree and combining multiple groups of parameter combination data, such as water temperature and runoff volume in meteorological and hydrological variables; and performing collaborative optimization processing based on the adaptive strategy model and a set analysis function to obtain an optimal strategy, and then performing mapping and analysis processing in combination with an analysis result to obtain targeted strategy list information for planning the decommissioning path of the thermal power generating unit. The invention aims to realize the planning of the decommissioning path of the thermal power generating unit.

The invention discloses an intelligent cloud computing resource scheduling method and system based on a digital technology, and relates to the technical field of cloud computing, and the method comprises the steps: collecting and preprocessing multi-dimensional resource use data, constructing a resource demand prediction model to carry out resource scheduling prediction, extracting feature factors based on predicted resource demands, and carrying out resource scheduling prediction. And calculating a comprehensive priority score, designing a scheduling strategy, optimizing a resource scheduling strategy by adopting an adaptive strategy combining cat group optimization and ant colony algorithm and simulated annealing, and implementing resource scheduling. Through combination of cat group optimization, ant colony optimization and simulated annealing algorithms, efficient combination of global search capability and local optimization capability is realized in a scheduling process, an efficient adaptive task scheduling mechanism is formed, the problems that priority division is unreasonable, a scheduling path is easy to fall into local optimum and the like in a traditional method are avoided, and the scheduling efficiency is improved. Therefore, the scheduling efficiency and the global resource utilization are optimized.

The invention discloses a strategy dynamic optimization system and method based on big data analysis and prediction, and relates to the technical field of data analysis and processing, and the system comprises a data collection layer which is used for building a data source; the data processing layer is used for performing data feature extraction after performing data cleaning on the data source; the extreme scene simulation layer is used for executing coping strategy training in an adversarial environment; the competition optimization layer is used for configuring a competition agent and updating an adaptive strategy set after adaptive evaluation of strategy quality, performing game optimization and generating a game optimization result; and the strategy dynamic optimization layer is used for receiving the game optimization result and establishing a dynamic response strategy. The technical problems that in the prior art, analysis deviation is caused by poor data quality, a static strategy is difficult to adapt to a complex and changeable market environment and cannot effectively cope with dynamic games of different competitive roles, and the strategy adaptability and flexibility are insufficient are solved, and the technical effect of improving the decision-making flexibility and adaptability of the fluctuating market environment is achieved.

The invention discloses a task scheduling method, electronic equipment and a storage medium. The method is applied to a configuration center side and comprises the following steps: acquiring a configuration instruction subjected to permission verification; acquiring multi-dimensional service scene data based on the configuration instruction, and inputting the multi-dimensional service scene data into a preset adaptive strategy model to generate a corresponding target adaptive strategy; the preset adaptive strategy model is obtained based on rule learning and transfer learning training; and performing visual simulation verification on the target adaptive strategy, and issuing the target adaptive strategy passing the verification to a server side. According to the scheme, the target self-adaptive strategy adaptive to the current service scene is generated through the multi-dimensional service scene data collected in real time and the preset self-adaptive strategy model, and the target self-adaptive strategy passing the visual simulation verification is issued to the server side for scheduling execution, so that the optimal task scheduling in the dynamic service scene can be realized, and the task scheduling efficiency is improved. And the autonomy and maintainability of the system are improved.

The invention discloses a road disease real-time identification embedded method and system based on a lightweight CNN and an attention mechanism. According to the method, a lightweight feature extraction module integrating reconfigurable convolution and channel attention is constructed, a multi-scale feature fusion network of grouped convolution and cross-stage connection is adopted, and a three-level prediction head focusing on diseases of different sizes is configured, so that the feature sensing and positioning capability on tiny diseases is remarkably improved. A composite loss function fusing classification, regression and attention perception is adopted in model training, and optimization is carried out in combination with self-adaptive strategies such as course learning and difficult case mining. Finally, the model scale is compressed through network pruning, quantification and knowledge distillation technologies, the model is finally deployed on edge computing equipment, and high-precision and low-delay real-time detection and response of road diseases are achieved through a reasoning acceleration engine. According to the invention, the problems of low precision and slow speed of road small target disease detection on an embedded terminal in the prior art are effectively solved.

The invention relates to the technical field of tourism resource information processing, in particular to a tourism hotel dynamic scoring and recommending method and system based on multi-factor fusion, and the method comprises the steps: constructing a dynamic game portrait containing a user space-time behavior mark and hotel multi-modal attributes, and carrying out the dynamic scoring and recommending of a hotel based on the dynamic game portrait; the method comprises the following steps: analyzing a recessive preference vector of a user and an adaptive strategy vector of a hotel by simulating an interactive game process of the user and hotel attributes, performing dynamic equalization matching on the recessive preference vector and the adaptive strategy vector to generate a hotel situation score, generating an anti-consensus recommendation sequence according to the hotel situation score, and performing recommendation on the hotel situation according to the anti-consensus recommendation sequence. The anti-consensus recommendation sequence is subjected to interpretable packaging, and a final recommendation result with a decision guide clue is output to the user, so that the user can obtain a personalized hotel recommendation list and can also understand logic and diversity values behind recommendation, and the user experience is improved. Therefore, the transparency, the user trust degree and the exploration satisfaction degree of the recommendation system are improved.

The invention relates to the technical field of software engineering and artificial intelligence, in particular to a meteorological field-oriented adaptive instruction generation driven large model fine tuning method, which comprises the following steps of: constructing a traffic meteorological seed data set; based on a preset adaptive instruction fine tuning strategy, performing data augmentation on the traffic meteorological seed data set to obtain an instruction data set; and based on a preset ChatGLM3-6B model and the instruction data set, performing fine tuning on the pre-training language model by using a preset low-rank adaptive strategy until the pre-training language model after fine tuning reaches a preset performance condition, obtaining a target language model, and performing meteorological prediction and / or providing meteorological information meeting a preset customization requirement through the target language model. Therefore, the problems of difficulty in creating a large-scale instruction data set, high resource consumption and the like in related technologies are solved, the large model can be finely adjusted by using the automatically generated meteorological field instruction data, and the performance of the model on the meteorological field task is improved.

The invention relates to the technical field of multi-modal feature fusion, in particular to a training method of a multi-cause regulation AI large model and an intelligent decision making system. Through data preprocessing, feature expression optimization, class center construction of support data, efficient distance calculation, dynamic parameter adjustment and a self-adaptive hyper-parameter and field self-adaptive strategy, the multi-cause regulation AI large model is obtained; according to the technical scheme, the robustness and accuracy of model training are remarkably improved. Meanwhile, the modularized hardware implementation scheme ensures the high efficiency and stability of the overall operation of the system, and can adapt to the requirements of the data size and the field change, thereby providing an efficient, stable and intelligent neural network model training platform for an intelligent decision-making system.