12 results about "Security industry" patented technology

The invention relates to the technical field of safety information intelligent processing, in particular to a robot question and answer service method and system based on safety industry information source data. The mixed information source collection strategy comprises collection of public internet data through an intelligent crawler, butt joint of closed information source data through an authorization interface and capture of regional real-time data through a localized edge node, and collection frequency is dynamically adjusted based on an information source type and event emergency degree; and carrying out real-time processing on the acquired data by adopting a stream processing framework, wherein entity extraction and format standardization are carried out on unstructured data. According to the method and the device, the global security domain data is acquired by adopting a mixed information source acquisition strategy, and the acquisition frequency is dynamically adjusted, so that the problem that information blind areas or lagging exist in question and answer responses due to the fact that traditional robot question and answer services mostly adopt single information source or fixed frequency acquisition is solved.

The invention relates to a novel security industry data resource analysis method, and belongs to the technical field of security, efficient processing and intelligent decision making of security data are realized through an end-side cloud cooperation architecture, real-time target detection is performed through deploying a lightweight YOLOv8 model through an edge end, and preliminary anomaly recognition is completed; an LSTM neural network is trained on the cloud platform based on historical alarm data, and high-risk area prediction in the next 24 hours is realized; a DBSCAN clustering algorithm is combined to carry out spatial correlation analysis on the personnel trajectory, and a suspicious gang is automatically identified; and finally, an entity relation network is constructed through the knowledge graph, and cross-system event reasoning is supported. According to the technical combination, the system has the capabilities of full space-time coverage, full factor fusion and full process intelligence, a closed-loop safety prevention and control system is formed, and the automation level and the actual combat capability of the security and protection system are effectively improved.

The invention discloses an AI native intelligent central operating system based on reinforcement learning, and relates to the technical field of reinforcement learning, and the AI native intelligent central operating system comprises the steps: determining a corresponding regulation action based on a preset business target and state representation; according to the invention, the plurality of sub-regulation items are determined based on the recognition of the regulation action, and the quantitative execution is carried out along each sub-regulation item, so that the change information of the regulated key performance indicator is predicted, and the accuracy of the change information of the regulated key performance indicator is improved. Determining a corresponding reward value according to the identification of the change information of the key performance indicator; determining to-be-optimized content according to reinforcement learning of the reward value; according to the method, each optimized sub-regulation item is determined according to the matching of the to-be-optimized content and each sub-regulation item, and the final operation content of the security industry AI native intelligent center is constructed in combination with the action constraint relationship of the security database, so that the accuracy of the final operation content of the security industry AI native intelligent center is improved.

The invention discloses a dynamic capacity expansion scheduling method, device and system for a security industry large model, and belongs to the technical field of security industry and large model capacity expansion, and the method specifically comprises the steps that a large model management platform sends a state calculation request carrying a copy resource demand to a node; the node calculates a time sensitive state score based on the real-time time sensitive parameter and replies the time sensitive state score; and the large model management platform calculates and ranks the final state scores of the nodes in combination with the scores replied by the nodes and the non-time sensitive parameters, and selects the optimal node for scheduling. The method effectively solves the problems of insufficient real-time performance of node parameters, low scheduling decision accuracy and high scheduling delay of a central platform in the prior art, and is particularly suitable for dynamic resource scheduling of large model applications such as real-time research report generation, intelligent customer service and abnormal transaction monitoring in the security industry.

The application discloses a securities industry large model factor mining and back testing method based on a Xinyuan executable environment, relates to the technical field of big data, and utilizes a large language model to identify a transaction factor from a financial research report and extract description information of the transaction factor; utilizes code generation to generate a large model to generate factor calculation code for the identified transaction factor; utilizes the factor calculation code and historical data in a historical market database to back test the transaction factor to obtain a factor value, utilizes a preset strategy testing method to evaluate the transaction factor based on the factor value to obtain an evaluation index for evaluating effectiveness of the transaction factor, and displays the transaction factor and the evaluation index through a visual interface on a client. The powerful natural language understanding capability and code generation capability of the large model realize automatic mining and back testing of the transaction factor in the research report, improve efficiency and accuracy of factor positioning, and automatically generate the factor calculation code, so that the back testing process is automated.

The invention belongs to the technical field of data processing, and particularly discloses a method for constructing a public safety industry knowledge base based on a large model, and the method comprises the following steps: firstly, obtaining data through an internal system and an external public channel of a public safety organ; cleaning and preliminarily integrating the collected data; then extracting entities, entity relationships and entity attributes from the cleaned and integrated data, and outputting a knowledge triple table; the extracted knowledge is fused; and finally, storing the fused clean knowledge into an internal table, and synchronizing the knowledge in the internal table to the GDB to form the public safety industry knowledge base. By means of data collection, data preprocessing, knowledge extraction and knowledge fusion and conflict resolution, public safety industry data storage and knowledge base construction are achieved, and the problems that due to the fact that public safety system data are dispersed in different service systems and are different in format, the data are difficult to achieve efficient integration and utilization, and the system reliability is poor are solved. Therefore, the working efficiency is influenced.

The invention relates to the technical field of safety protection, in particular to a security integrated command and dispatch platform system. Various data such as monitoring videos and real-time personnel positioning are integrated, data sharing and cooperative processing are realized through a software architecture, and a data island is broken; by applying artificial intelligence and big data analysis technologies, various security events are predicted and analyzed, and a scientific basis is provided for command decision making; various events in the security industry are monitored in real time, early warning rules are set, problems are found in time, and early warning is triggered, so that disposal measures are taken quickly; unified commanding is carried out on the security unit and the resident unit, efficient resource allocation and cooperative combat are achieved, and the emergency response speed is increased; various types of data are presented in a visual mode, such as electronic map labeling and tree diagram display, so that a command decision maker is assisted to intuitively understand the situation and make a decision. The system has the advantages of being easy to install and maintain, stable, reliable and capable of guaranteeing stable operation.

The invention discloses a network convergence-oriented AI-driven heterogeneous cooperative system and a data processing method. The network convergence-oriented AI-driven heterogeneous cooperative system comprises a routing control board card, an AI computing board card, a security board card, an internal and external network synchronous protection module, a software system and a high-speed interconnection bus. Through deep integration of hardware and software, intelligent, secure and highly reliable control of a data transmission process is realized. The system can complete intelligent link management, dynamic encryption protection, service priority scheduling and real-time data synchronization through multi-computing-power cooperation and multi-network fusion in a diversified environment so as to support high-security industry communication scenes of future railways / low altitude / satellites and the like.

The present application relates to the technical field of terminal equipment management and control, in particular to a terminal equipment management and control method based on public key certificate, the present application unifies a security model, is based on domestic and foreign general technical standards of security industry, is irrelevant with business rules, and the system is more efficient, unauthorized terminal equipment is rejected in the network protocol kernel level service, in addition, the safety factor is higher, and terminal management and control platform equipment certificate acquisition needs to pass the only path of self-research trusted authentication SDK of the office, and trusted authentication SDK acquisition needs to go through the relevant authorization approval process. The security protection of the kernel level can effectively prevent system infection risk, and finally supports domestic signal creation, which is based on large number decomposition inverse module operation or elliptic curve encryption algorithm, and the present application is not limited, and the signal creation environment can be flexibly selected and customized according to actual business requirements.

The application discloses a WeChat group public opinion on electricity rapid screening, marking and closed-loop treatment method and system based on RPA and AI image recognition, and belongs to the field of information technology and industrial automation. The method comprises the following steps: automatically collecting multimedia messages of WeChat groups through RPA robots; performing image character recognition and conversion by using a locally deployed OCR engine; performing rapid screening and false alarm elimination based on a pre-set domain sensitive word library and an associated non-sensitive word library; performing risk level determination and visual marking on the screening result; finally, generating a standardized work order and automatically pushing the work order to a business system to form a closed loop. Correspondingly, the system comprises a message collection module, a local recognition module, an intelligent screening engine, a hierarchical marking module and a work order pushing module. The core innovation of the application lies in that an AI recognition and processing architecture is locally deployed offline, so that data is not externally overflowed in the whole process, and through intelligent screening, structured preprocessing and business coding management, a full-process, automatic and intelligent closed loop from public opinion discovery to work order treatment is realized, and the application is particularly suitable for social group public opinion monitoring and service improvement in high-security industries such as electric power.

The invention relates to a security industry scheduling planning method based on a plan scheduling optimization algorithm, and belongs to the technical field of intelligent manufacturing and production planning. The method comprises the following steps: firstly, acquiring order, resource and production data through an industrial sensor and a manufacturing execution system (MES); and then, an improved genetic algorithm model embedded with exclusive constraints of security and protection production is constructed and trained, the model adopts an order-process-resource three-dimensional coding system, a scheduling scheme is ensured to be compliant through a dynamic punishment mechanism, and an optimization result including a basic scheduling, an emergency plan and a cost suggestion is output. According to the invention, the adaptability and flexibility of production scheduling in the security industry are effectively realized, and the production efficiency, the resource utilization rate and the system reliability are remarkably improved.