1016 results about "Intelligent computing" patented technology

The invention relates to the technical field of construction engineering cost management, and discloses an intelligent engineering cost calculation system based on multi-source heterogeneous data fusion, and the system comprises a multi-source data collection module which is used for collecting structured data and unstructured data from a design file, a market database, a construction monitoring system, a contract document, and a historical project library; and the heterogeneous data fusion module is connected with the multi-source data acquisition module and analyzes the risk terms in the contract text by adopting a natural language processing technology. According to the invention, the multi-source data acquisition module is used for widely collecting data in multiple aspects of design, market, construction, contract and the like, the problems of data splitting and information isolated island in traditional cost management are solved, integration of multi-source heterogeneous data is realized, and the heterogeneous data fusion module utilizes advanced technologies of natural language processing, image recognition and the like, so that the cost management efficiency is improved. Contract texts and design drawings can be efficiently analyzed, the processing capacity of unstructured data is improved, and the error rate and omission rate of manual interpretation are reduced.

The invention discloses an attention state recognition neural network modeling and reasoning method based on an electroencephalogram sequence. The core is to construct and train a deep neural network model suitable for electroencephalogram signals so as to realize intelligent recognition and classification. Firstly, wavelet transformation and time-frequency feature extraction are carried out on electroencephalogram time sequence signals, and a multi-dimensional input tensor is generated in combination with channel position information; and inputting the feature into a deep network fusing spatial convolution, gating circulation and a residual connection structure, and extracting spatio-temporal joint features. A cross-time-step attention mechanism and a dynamic loss adjustment strategy are introduced in a training stage, so that the discrimination capability of the model on an alertness state transition region is improved. The final model can conduct reasoning on electroencephalogram data of any length, and a classification label and a confidence score are output and used for measuring classification reliability. The method focuses on construction and optimization of a specific calculation model, reflects application characteristics of an intelligent algorithm in cognitive state recognition, and belongs to an intelligent calculation method with a neural network as a core.

The invention discloses an intelligent robot control method based on multi-sensor cooperation, and relates to the field of intelligent robots, and the method comprises the following operation steps: S1, carrying out the architecture deployment of sensors; s2, establishing communication; s3, edge intelligent calculation enhancement; s4, a dynamic adaptive fusion strategy; s5, carrying out group intelligent decision making; and S6, carrying out sensor role dynamic conversion. According to the intelligent robot control method based on multi-sensor cooperation, the sensor data credibility can be adjusted in real time according to the environment condition, the fusion weight can be dynamically allocated, the data accuracy is ensured through the cross check of the adjacent sensor data, and meanwhile, the intelligent robot control method based on multi-sensor cooperation can be realized through the ant colony cooperation algorithm in combination with the self-adaptive pheromone volatilization regulation and control. The path exploration strategy can be flexibly adjusted according to task complexity and environment dynamic change, the search efficiency can be effectively improved, various environments can be more quickly adapted, and the optimal action path can be quickly found.

The invention relates to the technical field of power management, in particular to an intelligent computing resource energy-saving scheduling system based on load prediction, which comprises a cache access monitoring module, a burst load judgment module, a frequency fluctuation analysis module, a load classification and identification module and an intelligent regulation and control strategy module. According to the method, the access frequency fluctuation is analyzed by monitoring the cache access data, the access mode change is identified, the load judgment precision is improved, the burst load identification is enhanced in combination with the hit rate reduction range, the resource scheduling is more accurate, the CPU / GPU operation frequency and the fluctuation rate are monitored, the abnormal fluctuation is identified, the calculation resource allocation is optimized, and the energy consumption waste of frequency adjustment is reduced; load classification, resource matching degree optimization, dynamic adjustment of CPU / GPU computing power and memory allocation, coordination of computing and data transmission, optimization of energy consumption management, guarantee of task stability and reduction of idle and unbalanced resource problems are carried out based on a time proportion and task characteristics.

The invention discloses an intelligent enterprise financial data analysis system and method based on artificial intelligence, and belongs to the technical field of enterprise informatization management and intelligent finance. The objective of the invention is to solve the technical problems of low processing efficiency, insufficient data insight, lagging risk identification, weak intelligent decision-making assistance capability and the like in the existing enterprise financial data analysis process. The system generally integrates a data automatic acquisition and integration module, a financial data deep analysis and modeling engine based on artificial intelligence, a key financial index and operation performance intelligent calculation module, a multi-dimensional financial risk intelligent identification and dynamic early warning module and an interactive visual analysis report and decision support module. The core of the method is that internal and external multi-source heterogeneous financial and operation data of an enterprise are automatically collected and fused, deep processing and intelligent analysis are performed on the data by applying artificial intelligence technologies such as machine learning and natural language processing, and accurate portrait of the financial condition of the enterprise, dynamic evaluation of operation performance and real-time monitoring and prediction of financial risks are realized. And analysis reports and optimization suggestions with insight are generated, so that enterprise managers can make efficient and scientific operation decisions. According to the method, the automation and intelligence level of enterprise financial data analysis can be remarkably improved, the financial information quality and decision support efficiency are improved, the risk prevention capability of enterprises is enhanced, and the enterprises are assisted to realize refined operation and value creation.

The invention discloses a dynamic load balancing optimization method for AI intelligent computing system server equipment, which relates to the technical field of load balancing optimization, and comprises the following steps: acquiring equipment performance data, and preprocessing the acquired equipment performance data; based on the preprocessed equipment performance data, extracting equipment state features through time sequence analysis and periodic fluctuation detection; predicting the future load change of the equipment based on the state characteristics of the equipment, dividing the priority of the equipment according to the prediction result, and analyzing the resource demand; generating a resource allocation strategy by applying a resource matching algorithm according to the device priority and the resource demand; and executing the resource allocation strategy, receiving feedback information in real time in the execution process, and optimizing the resource allocation strategy according to the feedback information. According to the method, the response speed and the service quality of the server equipment are remarkably improved, and the resource utilization efficiency is maximized.

The invention discloses an earth and rockfill dam seepage-deformation early warning method and system based on time-space combined anomaly, and belongs to the field of dam body safety data research. The method comprises the following steps: constructing a spatio-temporal topological graph based on an engineering coordinate system, integrating multi-dimensional data by nodes, and constructing a dynamic adjacency matrix according to spatial distance and seepage relevance; extracting features by using a space-time diagram convolutional network, a self-loop mechanism and cross-layer attention; and executing dual-drive early warning through standard threshold preliminary screening, multi-scale LSTM prediction and a time decay evidence theory. The system comprises a sensor network and an intelligent computing module, and the intelligent computing module has adaptive modeling and visualization functions. According to the scheme, seepage-deformation space-time correlation quantitative analysis is achieved, the hysteresis effect is captured, the threshold value is dynamically corrected, multi-source evidences are fused, the early warning timeliness and accuracy are improved, and the risk of false alarm and missing alarm is reduced.

The invention discloses a large-model-driven intelligent calculation method and system for a water conservancy mechanism model. According to the method, natural language input, structured conversion and intelligent optimization calculation of a scheduling target are realized by integrating a field-enhanced large language model and a water conservancy professional mechanism model. The method comprises the steps of receiving a calculation target expressed by a user in a natural language, and analyzing and converting the calculation target into a constraint condition and a target function which can be recognized by a water conservancy mechanism model; a hydrological model, a hydraulic model, a hydrodynamic model and other models are called based on a workflow engine, and reverse calculation is carried out by adopting a hybrid optimization strategy of'coarse adjustment-fine adjustment-verification '; synchronously and visually displaying the parameter change and the result convergence state in the calculation process; and outputting a calculation result including parameter adjustment logic, standard conformity analysis and multi-scheme comparison. The system comprises a natural language interaction module, a target conversion module, an intelligent calculation engine module, a visualization module and a result generation module, and supports multiple application scenes such as multi-target scheduling, emergency decision making and ecological guarantee. Compared with a traditional scheme, the method has the advantages that the model use threshold is lowered, the dispatching efficiency and calculation transparency are improved, and the method is suitable for complex hydraulic engineering calculation tasks such as reservoir dispatching, cross-basin water transfer and flood control emergency.

The invention provides a computing card group scheduling method and device based on task perception and generation length prediction, and the method comprises the steps: sending a plurality of reasoning requests with user prompt words to a computing card group, packaging each reasoning request as a request object, and enabling the computing card group to have a plurality of intelligent computing cards, the video memories of all the intelligent computing cards form a shared video memory pool; adding the request object into a user request pool; performing generation length prediction on the user prompt word of the request object in the user request pool by adopting a predictor module to obtain a predicted generation length; creating a corresponding reasoning task and metadata for each request object in the user request pool, applying for a video memory block as required from the shared video memory pool according to the predicted generation length, obtaining a video memory block handle with a video memory allocation address and size, and writing the metadata into the video memory block handle; generating a task scheduling sequence according to the metadata of the reasoning task and the running state of each computing card; and reasoning tasks are selected from the task scheduling sequence in sequence.

The invention relates to an intelligent computing resource allocation method based on reinforcement learning. According to the technical scheme, the method comprises the steps that monitoring data from different sources including hardware monitoring, software logs and network bandwidths are fused, and a multi-dimensional time sequence state space is formed; carrying out dimensionality reduction and de-noising processing on the multi-modal data by using a depth auto-encoder; multi-task learning MTL is introduced into time sequence modeling, and resource requirements and state evolution of multiple tasks are predicted at the same time; generating a plurality of predictive resource scheduling strategies by using a GAN (Generative Adversarial Network), and dynamically selecting a strategy scheme when a load changes; each strategy is realized by an independent sub-network, and part of core knowledge is shared; through a strategy evolution mechanism, according to a historical feedback optimization strategy combination including task completion time and resource consumption, in a heterogeneous resource environment including a plurality of cloud computing platforms and edge computing nodes, based on difference of resource types, fine-grained scheduling of strategies is carried out; and scheduling the decision by using a distributed Q-learning mechanism in the reinforcement learning model.

The invention discloses a multi-dimensional computing power dynamic perception routing decision-making method and system based on SRv6 driving. The method comprises the following steps: firstly, screening an optimal target computing power node based on a service demand and a node real-time load; and then, based on the dynamic network topological graph fusing the service sensitivity, calculating an optimal network path reaching the computing power node by using an improved SPFA algorithm, and realizing a global collaborative decision of selecting the most suitable computing point and searching the most efficient connection path. Meanwhile, based on an SRv6 driving service chain dynamic generation and closed loop execution method, a target computing power node and necessary network functions are abstracted into a programmable SID, an SID sequence (service chain) is dynamically constructed according to a double-stage decision result and is packaged in an SRH head, second-level path issuing and state monitoring are achieved through a programmable controller, and a real-time monitoring result is obtained. According to the method, the problems of poor real-time performance, single dimension, weak cooperative capability and the like in the prior art are solved, and efficient, stable and intelligent development of a future-oriented distributed intelligent computing network system can be promoted.

The invention discloses an intelligent computing power integration service management method and platform based on cloud side-end cooperation, and relates to the technical field of computing power integration management, and the method comprises the steps: sensing the computing power resource state of a side-end cooperation port and terminal equipment in real time at a cloud controller, and building a resource topological graph; a dynamic task demand is introduced, and a computing power scheduling vector is generated; carrying out lightweight segmentation on the cloud training model, and determining adjacent edge nodes under hierarchical limitation to form a regional elastic cluster; and deploying a digital twin simulation engine rehearsal computing power distribution scene, establishing a heterogeneous resource pooling mechanism, and carrying out computing power integration service management. The technical problems of low management efficiency and insufficient utilization rate of heterogeneous computing power resources in the prior art are solved, and the technical effects of realizing efficient management of intelligent computing power integration services and improving the utilization rate of the heterogeneous computing power resources are achieved.

The invention discloses a distributed model training-oriented routing reconstruction method in a hybrid photoelectric network, and belongs to the field of network communication optimization. The method comprises the following steps: constructing a collaborative routing model based on a current network flow state and topological information, and introducing an adaptive weight mechanism to dynamically adjust the priority of optical channels and electric packet forwarding, thereby realizing efficient collaborative utilization of multi-layer forwarding resources. According to the method, path reconstruction can be rapidly completed according to task characteristics and a network dynamic state, so that the system resource utilization rate is improved, and the maximum task completion time delay is reduced. Compared with a traditional fixed routing and semi-dynamic routing method, the method has the advantages that the average completion time delay and the maximum time delay are improved by more than 45% in the large-flow environment generated by large-scale model training, and the method has good expandability and real-time performance and is particularly suitable for burst flow intensive intelligent computing cluster scenes.

The invention belongs to the field of rehabilitation medical treatment and intelligent computing, and particularly discloses a cerebral apoplexy rehabilitation personalized treatment recommendation system based on artificial intelligence. Comprising a multi-modal data acquisition module, a feature extraction and clustering modeling module, a personalized path making module and a recommended path generation and feedback optimization module, and can realize dynamic monitoring and intelligent analysis of rehabilitation data of a stroke patient; a multi-modal aggregation neural network model is constructed by fusing multi-dimensional data such as images, biomarkers, clinical behaviors and cognitive emotions, a rehabilitation path is formulated according to a medical data clustering result, and a recommended path is dynamically adjusted and optimized in combination with treatment feedback of a patient; compared with a traditional method, the method has the advantages that scientificity and adaptability of personalized treatment are improved, the method has the advantages of real-time updating of rehabilitation paths, closed-loop management of a system, high model robustness and the like, rehabilitation intervention efficiency and accuracy can be remarkably improved, and technical support is provided for intelligent development of rehabilitation medicine.

Systems and methods for managing a computing task are provided. The system includes a controller configured to receive the computing task, present to a client a set of factors defining desired conditions related to executing the computing task, receive a demand computing strategy associated with the computing task indicating one or more factors selected by the client and a constraint and a weight related to each factor from the one or more selected factors, receive a plurality of supply computing strategies from a plurality of datacenters, each supply computing strategy corresponding to a datacenter from the plurality of datacenters, calculate a task scheduling strategy based at least on the demand computing strategy and a supply computing strategy, select a candidate datacenter from the plurality of datacenters according to a predetermined rule, and schedule the computing task for execution on the candidate datacenter according to the task scheduling strategy.

The invention relates to a mineral reserve dynamic monitoring method and system based on satellite remote sensing and a block chain. The method comprises the following steps: acquiring original remote sensing data and metadata of a mining area, executing hash operation on the data by utilizing a satellite-borne trusted execution environment to generate integrity proof, and performing digital signature to construct a trusted data packet; transmitting the data packet to a ground station to verify the validity of the signature, extracting verified data, and submitting the verified data to a block chain for evidence storage; based on the on-chain evidence storage data, using a deep learning model to extract earth surface features and calculating reserves changes; and finally, a reserve change result is returned to the block chain, and dynamic supervision and early warning are realized through an intelligent contract, so that a technical system from data source credible guarantee, transmission process tampering prevention, reserve intelligent calculation to supervision automatic execution is constructed. The problems that in traditional mineral product monitoring, the authenticity of a data source is lack, the processing process is not traceable, and supervision response is lagged are solved, and the reliability, auditing performance and supervision timeliness of a mineral product reserve dynamic monitoring result are improved.

The invention relates to the technical field of remote monitoring, and discloses a coal unloader remote monitoring system based on AI learning and a cloud platform, and the system comprises an impact force monitoring module, a vibration mode analysis module, an operation state evaluation module, a health diagnosis module, and an intelligent early warning module. According to the method, through calculation of the magnitude, the direction, the stress time and the stress frequency of the impact force, fine-grained recognition of coal flow impact force distribution and fusion of vibration data and impact force data are enhanced, comprehensiveness and multi-dimensional data cross analysis of fault recognition of the coal unloader are improved, the abnormal misjudgment rate is reduced, intelligent calculation based on AI learning is achieved, and the fault recognition accuracy of the coal unloader is improved. According to the invention, a health assessment result can be dynamically adjusted, misjudgment caused by single-point abnormity is avoided, a cloud data processing mode enables remote management to be converted from data collection to intelligent early warning, the prediction capability of equipment operation abnormity is enabled to be more timely, remote monitoring is combined with intelligent learning, multi-level assessment is formed, and the accuracy and decision efficiency of remote management are improved.

The invention provides an intelligent computing fusion network routing system and a routing method thereof. The system comprises an entrance gateway, a routing forwarding node, a computing node and a computing network controller, the computing network controller is used for carrying out path selection and computing resource allocation in combination with a real-time network state and load information of the computing node, dynamically generating and updating a computing network routing table through a computing network fusion routing algorithm, and issuing the computing network routing table to the entry gateway and the routing forwarding node; and the entry gateway receives and stores the computing network routing table issued by the computing network controller as a data stream access point, receives a computing power data stream which is input from the outside and carries a computing power service request, executes computing power routing table matching according to the computing power service request, and forwards the computing power data stream to a proper routing forwarding node according to a matching result. According to the invention, by optimizing the joint scheduling of the computing power service paths and the computing nodes, a plurality of paths of the anycast computing power routing and the weights of the paths are generated, the service flow level load balancing is realized, and the end-to-end time delay is reduced.

The invention relates to the technical field of network security monitoring scheme design based on computing power, in particular to a network security monitoring method and system based on computing power and electronic equipment. The method comprises the following steps: collecting network traffic, system logs and user behavior data; dynamically distributing CPU / GPU computing power according to network flow, attack frequency and the like; performing cleaning and format conversion on the data; intrusion detection, vulnerability scanning and traffic anomaly analysis are executed through multi-task parallel processing; a potential attack mode is mined in combination with deep learning model and rule engine association analysis; and judging threats according to a preset rule base and triggering early warning. According to the method, the resource utilization rate is improved through intelligent computing power scheduling, the threat recognition capability is enhanced in combination with deep learning and rule double engines, the method can adapt to a complex network environment, the threat detection accuracy is 98.7% according to actual measurement display, the resource occupation is reduced by 30%, and the method is suitable for high-concurrency scenes such as a cloud platform and the Internet of Things.

The invention provides a super-node system which comprises an intelligent computing resource pool and a general computing resource pool, the intelligent computing resource pool comprises a first in-pool switching module, a first inter-pool switching module and a plurality of GPUs, the general computing resource pool comprises a second in-pool switching module, a second inter-pool switching module and a plurality of CPUs, and decoupling of heterogeneous computing resources is achieved. The GPUs in the same intelligent computing resource pool and the different intelligent computing resource pools can communicate through the first in-pool switching module, the CPUs in the same general computing resource pool and the different general computing resource pools can communicate through the second in-pool switching module, and the system can provide intelligent computing and general computing resources at the same time. And the intelligent computing resource pool and the general computing resource pool can be expanded respectively, so that elastic matching of resources is realized, and the utilization rate is improved. Intelligent computing resources and general computing resources are pooled and deployed in different cabinets, deployment decoupling of heterogeneous resources is achieved, the single-cabinet GPU density of the intelligent computing resource cabinets is improved, the influence range of single-point faults is reduced, and maintenance is easy.

The invention relates to a cross-modal document information extraction method based on space-semantic alignment, and belongs to the field of artificial intelligence, computer vision and natural language processing. According to the method, the spatial feature and semantic information bidirectional alignment model is designed, by constructing the spatial feature and semantic feature bidirectional alignment model, the document layout information can dynamically adjust attention distribution of text semantic features, meanwhile, semantic information reversely optimizes the spatial features, collaborative modeling of spatial layout and semantic information is achieved, and the document layout efficiency is improved. Therefore, the accuracy and robustness of complex document information extraction are improved. According to the method, a hierarchical cross-modal information extraction model is designed, through the hierarchical cross-modal information extraction model, the overall structure of a document is recognized on the global level, local key content is focused on the regional level, fine modeling is conducted on fine-grained texts and visual elements on the entity level, and accurate recognition of a cross-modal entity and the semantic relation of the cross-modal entity is achieved; and the generalization ability and applicability of information extraction are enhanced.

The invention discloses an intelligent computing center automatic operation and maintenance management method based on computing power resource allocation, and relates to the technical field of resource management, and the method comprises the steps: obtaining real-time computing power load data and a to-be-processed task queue of an intelligent computing center; constructing a resource state matrix according to the real-time computing power load data, performing historical data backtracking analysis on the resource state matrix by using a sliding window algorithm, and calculating a resource load predicted value and a resource availability score of each computing node; based on the resource availability score, an improved genetic algorithm is adopted to carry out optimal allocation solution on the constructed task-resource matching matrix, and an optimal task allocation scheme is generated; and converting the optimal task allocation scheme into a resource allocation instruction set, sending the resource allocation instruction set to each target computing node to execute task scheduling, and updating the resource occupation state of the corresponding node in the task-resource matching matrix. According to the invention, the technical jump from passive response type management to active prediction type management is realized, and the resource configuration efficiency of an intelligent computing center is improved.

The invention discloses a multi-data center multi-computing power collaborative optimization method and system based on computing power and refrigeration system comprehensive energy consumption cost, and a storage medium. The method comprises the following steps: S1, uniformly dividing a total conventional computing power resource and a total intelligent computing power resource which need to be scheduled into a plurality of sub-computing power resources; s2, sequentially allocating and starting a data center for each conventional sub-computing power resource; S2.1, calculating the cost of each data center after the conventional sub-computing power resource is started, and selecting to start the data center with the minimum cost; s2.2, repeating the step S2.1 until the total conventional computing power resource needing to be scheduled is reached; s3, sequentially allocating and starting a data center for each intelligent sub-computing power resource: S3.1, solving an optimal operation period deployment scheme of the intelligent sub-computing power resource in each data center through a genetic algorithm, calculating the cost after starting the intelligent sub-computing power resource according to the optimal operation period deployment scheme, and selecting to start the intelligent sub-computing power resource with the minimum cost; s3.2, repeating the step S3.1 until the total intelligent computing power resource needing to be scheduled is reached; according to the method, the computing power resource operation cost can be minimized.

A full-automatic movie and television video production system based on digital actors belongs to the field of artificial intelligence, computer graphics and movie and television production and comprises a creative engine unit, a production unit, an intelligent shooting unit, a post-processing unit and a quality inspection and rendering unit. The method has the advantages that 1, the manufacturing efficiency is greatly improved, the traditional film and television manufacturing period is usually 6-12 months, the manufacturing period can be shortened to 4 hours by the method, and the efficiency is improved by about 1000 times; 2, the manufacturing cost is obviously reduced: the manufacturing cost of traditional movies and other movies is usually 5 million to 20 million dollars, but the method only needs the computing power cost, and the cost is less than one percent of that of the traditional movies; 3, the style consistency is guaranteed, traditional film and television production depends on personal ability, the style consistency is difficult to guarantee, and the style consistency is strictly guaranteed through an algorithm and is improved by 100%.

The invention discloses an electric power analysis decision calculation method and system based on a diffusion process graph neural network, and belongs to the technical field of electric power system analysis, and the method specifically comprises the steps: carrying out the subproblem construction of a calculation architecture, and obtaining an electric power graph intelligent calculation framework based on diffusion process induction; constructing a graph structure of the power system according to the physical topological structure and the real-time operation data of the power system; aiming at the obtained parameter change characteristics of the graph structure of the power system in iteration, designing a node embedding method fused with physical constraints to obtain a complete node embedding representation; gNN iteration training is carried out on the obtained complete node embedding representation, node state diffusion and convergence are realized through multiple times of iteration calling, and a convergent reinforcement learning model is obtained; performing offline training on the obtained reinforcement learning model; and inputting real-time data into the trained reinforcement learning model, predicting to obtain a future state of the power system, realizing power analysis decision calculation, and generating a control decision based on the future state.

The invention relates to the technical field of intelligent charging, in particular to an edge intelligent calculation enabling type peak-valley energy storage charging system which comprises a power grid frequency judgment module, an energy storage dispatching output module, a load trend recognition module, a multi-station matching sorting module and a load dynamic distribution module. According to the invention, through the continuous data analysis of the power grid frequency, the energy storage unit power and residual energy, and the load current, the system can accurately identify the linkage relationship of various parameters during the frequency abnormal response period, refine the condition judgment of peak-valley energy storage triggering and scheduling, optimize the energy storage output decision path, and improve the reliability of the system. Through dynamic screening of load trends and changing stations, the real-time performance and flexibility of energy allocation among multiple stations are enhanced, the adaptive capacity of the system to complex supply and demand changes is further improved in station matching and priority ranking, energy storage resources and loads are dynamically allocated, the electric energy utilization rate and the cooperative regulation and control level of the energy storage system are effectively improved, and the energy utilization rate and the cooperative regulation and control level of the energy storage system are improved. And intelligent upgrading of energy management is promoted.

The present invention discloses a method for constructing an intelligent computation engine of an artificial intelligence cross-platform model based on knowledge self-evolution. The method comprises: determining source and target moments; dividing a discrete manufacturing system data set; initializing a dynamic discrete manufacturing system model; preprocessing data and constructing a task pool; constructing a meta learning framework; migrating a trained neural network to new tasks; iterating until convergence and storing model parameters; and testing in a new environment. This invention shortens the convergence time of model parameters, significantly benefiting the training of dynamic discrete manufacturing models subject to temporal disturbances in actual production.

The application discloses a student intelligent psychological state evaluation system, and relates to the technical field of data communication. First, the quality of student published data is checked and automatically corrected to ensure the reliability of subsequent processing input. Then, data sequence clusters are divided through clustering analysis, and resource scheduling strategies are dynamically adjusted based on the similarity confidence of each cluster to realize the optimization of analysis frequency and sampling rate of different clusters. Meanwhile, the resource scheduling strategy in the cluster is dynamically adjusted, and the optimal task allocation plan is intelligently calculated and issued according to the scheduling exposure index and resource proportion index of each sequence cluster to realize the fine matching of computing power and manpower. Finally, the real-time monitoring of key performance indicators improves the real-time performance and accuracy of student psychological state evaluation, and significantly optimizes the resource utilization efficiency and stability of the system in a high-concurrency scenario.

The invention discloses a heterogeneous intelligent computing power optimization management scheduling system for accelerating a large model reasoning task, and relates to the technical field of computing power optimization management scheduling. The video memory fragmentation problem in a long sequence scene is converted into a controllable block migration task, and the performance bottleneck of a traditional video memory exchange mechanism is broken through; based on an operator-level scheduling strategy of a hardware capability fingerprint database, position coding and other compute-intensive tasks are accurately matched with vector instruction set hardware, and resource mismatch loss caused by black-box scheduling is eliminated; an expert selection process is reconstructed by an integer routing and counting sorting algorithm, near-lossless reasoning is realized at a limited node of an instruction set, and the potential value of an old computing power pool is activated.

The invention discloses a reinforcement learning calculation simulation method and device, electronic equipment and a storage medium, and relates to the technical field of artificial intelligence computing, and the method comprises the steps: inputting the determined current model parameter configuration, current hardware configuration and current working load into a target simulation system to obtain a plurality of parallel grouping combinations, determining a target simulation system according to the current hardware configuration, determining an effective parallel packet combination from the plurality of parallel packet combinations based on a preset Monte Carlo method, inputting the effective parallel packet combination into a simulator of a preset neural network model, and performing delay time calculation according to the effective parallel packet combination through the simulator to obtain a delay time sequence; and the combination corresponding to the shortest delay time is used as a target parallel grouping combination, so that the technical problems of mismatching of simulation scenes, insufficient precision and lack of effective support for heterogeneous clusters are solved, reliable performance prediction and optimal parallel strategy suggestions are provided through high-precision performance modeling and automatic exploration, and the method is suitable for large-scale popularization and application. Therefore, the resource consumption of large-scale GRPO training is reduced.