30 results about "Security metric" patented technology

The invention belongs to the technical field of AI large model evaluation, and particularly relates to an AI large model performance and risk automatic evaluation platform and method fusing multiple normal forms. According to the AI large model performance and risk automatic evaluation method fusing multiple normal forms, multi-dimensional evaluation and accurate repair of the AI large model in a complex scene are achieved, the performance index of the model is considered, the safety index and the reliability index are introduced, and therefore the actual situation of the model can be reflected more comprehensively, and the risk of the AI large model can be evaluated more accurately. By implementing a progressive pressure test and monitoring a performance attenuation curve and a failure critical point of the model in real time, potential problems of the model can be found in time, major faults in practical application are avoided, meanwhile, fragile links of the target model can be accurately positioned through backtracking analysis and iterative repair, iterative repair is performed according to a repair priority, and the fault detection accuracy of the target model is improved. Therefore, the stability and reliability of the model are effectively improved.

Methods, systems, and computer storage media for providing brand-focused security administration in an item listing system are described. Using a brand-focused machine learning model, the brand-focused security administrator engine supports generating different comprehensive reports and analytics on brand-related security metrics, including brand compliance rates, infringement incidents, enforcement actions taken, and overall item listing system integrity. The brand-focused security administrator engine also supports visualizing the brand-focused security analytics results data in a manner that enhances usability and decision-making by presenting complex brand-focused security analytics results data in a clear, actionable format. In operation, brand-focused security data is accessed. The brand-focused security data is analyzed using a brand-focused machine learning model that is trained based on a multi-dimensional authenticity analysis dataset. Brand-focused security analytics results data is generated. The brand-focused security analytics results data is communicated to cause display of one or more visualizations based on the brand-focused security analytics results data.

Systems and methods for migrating users and modifying workspace definitions of persona groups are described. In some embodiments, a memory storage device may have program instructions stored thereon that, upon execution by a workspace orchestration service, cause the service to: distribute a first workspace definition based upon a first persona group associated with a user to a client device operated by the user, where the user IHS is configured to instantiate a first workspace based upon the first workspace definition; and in response to a change in a productivity or security metric of the first workspace, determine whether to: (a) migrate the user to an existing second persona group, (b) create a third persona group and migrate the user to the third persona group, (c) modify the first persona group and migrate the user to the modified first persona group, or (d) identify a potential security threat.

There is provided an apparatus suitable for use for facilitating trustworthiness assessment. The apparatus can include a first module and a second module. The first module can be configured to receive at least one input signal which can be processed by the second module to generate one or more output signals. The input signal(s) can be associated with at least one user requirement and / or at least one security objective. The input signal(s) can be processed by manner of identifying at least one security metric associated with at least one security requirement. The security requirement(s) can be based on at least one user requirement and / or at least one security objective, determining a positive determination, a negative determination or an indeterminate determination concerning fulfillment of a security metric in respect of at least one security objective and / or at least one user requirement and deriving at least one trustworthiness score accordingly.

A method of quantifying the satisfaction of security requirements is provided via characterizing a security feature; matching the security feature to a security metric; computing a quantification score that indicates the exploitability of a system to which the security feature is applied; and outputting the quantification score to a security analyst.

It is provided a method for determining a security metric of isolation between different domains (11). The method comprising: determining (40) a scope (11, 12, 13) for a security metric; determining (42) a security control set comprising a plurality of security controls; evaluating (44), for each security control in the security control set, whether the security control is implemented in the scope (11, 12, 13), wherein an implemented security control in the scope (11, 12, 13) defines coverage of the security control; determining (46), for at least each covered security control in the security control set, a correctness value indicating how well the security control is implemented; and calculating (48) the security metric based on both the coverage of the security controls and the correctness values.

The invention provides a comprehensive security index evaluation method based on multi-task learning, and belongs to the technical field of network security management, and the method comprises the steps: synchronously collecting original data from a plurality of heterogeneous security data sources; a multi-task learning neural network model is constructed, and the model comprises a shared feature coding network for outputting shared features; the at least two heterogeneous task decoding networks are used for respectively mapping the shared features into preliminary security assessment results with different attributes; the fusion network is used for carrying out weighted fusion on the preliminary security assessment result to generate a comprehensive security index; using the multi-source security data set to train the multi-task learning neural network model; and inputting real-time security data of a to-be-evaluated system into the trained model, and outputting a comprehensive security index. The method has the advantages that the shared feature coding network can extract high-level feature representations with commonality from different data sources, and feature engineering dependence for a single data source or a single task is avoided.

The present invention provides a software heterogeneous security evaluation method and device. The method includes: performing heterogeneous processing on the original software by using multiple software heterogeneous technologies with different granularities to correspondingly obtain multiple different software variants; respectively extracting gadget feature information affected by the software heterogeneous technologies from the original software and the multiple software variants, and measuring the differences in the difficulty of constructing a gadget attack chain, the potentially available computing power of the attacker, and the cost of the attacker searching for gadgets in the original software and the software variants according to the gadget feature information; obtaining a security metric value of the software heterogeneity according to the differences in the difficulty of constructing a gadget attack chain, the differences in the potentially available computing power of the attacker, and the differences in the cost of the attacker searching for gadgets.

This application discloses a method for full life-cycle damage detection and safety assessment of tunnel structures, belonging to the field of damage detection and safety assessment technology. The method includes: installing and configuring a mobile detection platform along the tunnel's axial direction; generating the mobile detection platform's trajectory and detection point distribution scheme; controlling the mobile detection platform to move sequentially to each detection section position according to the detection path, simultaneously activating each detection module for continuous data acquisition during the movement; generating a three-dimensional mathematical model of the tunnel, a lining thickness distribution map, a thermal radiation anomaly distribution map, and a crack damage distribution map; dividing the tunnel's three-dimensional model into N detection segments at preset intervals, and statistically analyzing the average lining thickness, the number of thermal radiation anomaly points, the crack length density, and the damage level of each segment as safety assessment indicators; performing statistical analysis and trend judgment on the safety indicators of each segment, identifying abnormal segments exceeding safety thresholds, and determining their risk levels.

The application provides a real-time security evaluation method and device. The method comprises the following steps: acquiring a monitoring data stream, and extracting security level information corresponding to each time point of all data blocks from the monitoring data stream; updating the weight change of the security index based on the security level information through a KernelSHAP method; incrementally updating an evaluation model by using the updated security index and weight, and predicting a security evaluation result by using the updated evaluation model; and determining a security evaluation result according to a preset decision threshold and the security evaluation result. By using the KernelSHAP method, the importance of the newly collected monitoring data stream is determined based on a wide learning system structure, the negative influence caused by the approximation error is reduced, the drift detection is combined with the adaptive and generated explanation ranking preference, and therefore the result of the real-time security evaluation task can be better predicted.

The invention belongs to the technical field of intelligent transportation, and relates to a test case generation method for single-intersection multi-subject cooperative passage. Aiming at the problems that an existing test case generation method is insufficient in authenticity in a complex interaction scene and coupling parameters are difficult to model effectively, the method comprises the following steps: firstly, analyzing real natural driving data, and extracting deep coupling time-varying parameter characteristics in a left turn-straight conflict scene of the intersection; further providing a safety index-expected collision distance ACD for quantifying the conflict risk; on the basis, a Markov method is combined to carry out sampling estimation on probability distribution of scene parameters, a Markov chain is utilized to construct a state transition probability matrix, and dynamic trajectory generalization generation of coupling parameters is realized; for a complex scene, an independent parameter combination strategy is further fused to generate a complete test case set, and a reliable test benchmark is provided for an IVCPS system. According to the method, the generation efficiency and authenticity of the test case in a complex interaction scene are remarkably improved, and the technical blank in the field is filled.

Apparatus receives configuration information. The configuration information is for monitoring at least one security event and transmitting a security report including at least one security log for the at least one security event. The configuration information indicates at least one security event to be monitored and a criterion to transmit the security log report (SLR). The configuration information may be received from a policy control function (PCF). The security log report may be transmitted to a continuous security monitoring and analytics function or an edge computing server. The configuration may also be used for computing security metrics. The configuration may indicate a parameter, a layer, an interface, a level of detail or a format of the security log report which are to be monitored. The format of the report may comprise a plurality of key value pairs. The criterion to transmit the security log report may include a threshold, a periodicity, or a geographic area. The apparatus may be a user equipment (UE) which may be of the smart phone, internet of things (IoT), extended reality (XR), unmanned aerial vehicle (UAV) type. Also disclosed, receiving security policies. Translating the security policy into configuration information, and transmitting the configuration information.

The invention relates to the technical field of safety assessment, and particularly discloses a smart campus safety assessment method based on the Internet of Things, and the method comprises the steps: data collection and processing, safety index selection, evaluation matrix construction, index weight calculation, index comprehensive evaluation and result decision. According to the scheme, the random forest is constructed, the feature set is screened out through split bagging ensemble learning, the new random forest model is constructed based on the feature set for training, and it is ensured that the finally selected features have high importance in all data; the information entropy is calculated, the dispersion degree of each safety index is measured, the scientificity of weight evaluation is improved, the weight of each safety index is quantified, the redundancy is calculated, and the effectiveness of the indexes is better understood; the weighted decision matrix is calculated to enhance the accuracy of security assessment, so that more important indexes play a greater role in assessment, the relative advantages and disadvantages of each index are quantified, subsequent sorting and decision making are facilitated, and a reliable basis is provided for security improvement.

The invention relates to a shield tunneling control method based on safe adjustable space and Euler spiral planning. The shield tunneling control method comprises the following steps that S1, current shield attitude information and target attitude information are obtained; s2, based on the current attitude information, the target attitude information and a mapping relation between a preset shield attitude variation and a plurality of safety constraint indexes, analyzing an attitude variation adjustable range when the shield completes current ring tunneling under all safety constraint conditions; according to the method, by establishing the mapping relation between the attitude variation and the security index, the attitude adjustable space conforming to the security constraints of all equipment is analyzed, and the physical security of the decision basis is guaranteed; and furthermore, performing optimization decision in the safe space, generating a smooth path with continuous curvature by using an Euler spiral, and finally realizing accurate tracking through closed-loop control. According to the method, the transformation from artificial experience decision-making to model-driven automatic optimization is realized, and the safe and efficient optimal tunneling attitude can be automatically found under multiple constraints.

Techniques for optimizing routing decisions based on security metrics within a network environment are described herein. In some cases, by using various security metrics, such as encryption indicators, attestation indicators, secureness metrics, and reliability metrics, an exemplary system can assess the security level and reliability of network paths. These metrics may provide valuable insights into the trustworthiness and integrity of participating nodes and links and enable informed decision-making regarding path selection.

The present invention enhances the security visibility of private network resources within a telecommunications network operated by a mobile network operator (MNO). A private network resource is provided to an entity and network security metrics are determined for multiple hops / links within the private network resource. Device security metrics are also determined for multiple wireless devices connected to the private network resource. The network security metrics indicate the security posture of the multiple hops / links, including at least one intermediate hop between the telecommunications network and an endpoint where the entity receives wireless communication data. The device security metrics indicate the security posture of the multiple wireless devices. The network security metrics and device security metrics are output on an individual basis to the entity via a dashboard on a user interface. This allows the entity to view respective security metrics associated with each hop / link and each wireless device.

In some implementations, a system is provided for asynchronously performing and monitoring cybersecurity investigations. A scheduler selects event detection jobs to be performed during a specified time interval, and provides the selected event detection jobs to an orchestrator. For each event detection job, the orchestrator executes a corresponding event detection job worker instance. Performing the event detection job includes querying at least one event data log for event data that represents events that have occurred over a specified time range, processing the event data to generate a detected security metric, and in response to the detected security metric meeting a threshold value, publishing a cybersecurity alert to a cybersecurity investigation result repository. A response system subscribes to the cybersecurity investigation result repository to identify cybersecurity alerts that have been published to the repository, and in response to identifying a cybersecurity alert, performs a security response.

The invention discloses a path planning system for dike breach rescue equipment, and the system comprises an environment sensing module which is used for obtaining environment data and position information, and constructing a rescue environment digital model; the path planning model is used for determining path planning based on the rescue environment digital model; the path planning model is formed by coupling a reward function of a strategy network and a security prediction network, fusing a reliability security index to form an agent decision-making model, and performing training construction based on a PPO algorithm; wherein the policy network is used for acquiring shared features, outputting an action mean value to the security prediction network, and determining a path trajectory of an equipment target according to a total reward; and the interaction module is used for outputting the real-time path planning result, the risk probability and the security classification. According to the technical scheme, dynamic coupling of geological bearing capacity, water flow power, pontoon bridge stability and equipment dynamics is comprehensively considered, and global safety constraints are formed; the transportation safety of the equipment in a complex environment is guaranteed, and the rescue efficiency is greatly improved.

The invention belongs to the technical field of artificial intelligence, and relates to a security policy arrangement method and system based on a large model, and the method comprises the steps: S1, constructing a security component library which is provided with a plurality of security components; s2, constructing an environment state vector; s3, setting different priorities for each security index and forming a user preference vector; s4, constructing a multi-target reward function; s5, training a security policy arrangement model based on reinforcement learning; and S6, when a current user request is received, utilizing the trained security policy arrangement model to dynamically generate a security component calling sequence according to the current environment state vector and the user preference vector, and gradually examining and processing the current user request and candidate replies based on the security component calling sequence. According to the method, normal form transition from a static rule link to dynamic intelligent arrangement can be realized, so that a large model has high safety, high availability, low cost and self-adaptive capability.

The disclosure relates to a computer implemented method for assisting a user managing the data-security of backup copies of a computer system having a plurality of nodes, the method comprising: receiving status data for backup copies associated with a plurality of nodes, wherein, for each node, the status data provides a status of one or more backup copies associated with the node with respect to a plurality of data-security criteria; determining a backup security metric for each of the plurality of nodes based on the status data; and providing the security metrics for the user to demonstrate the relative level of backup data-security of the plurality of nodes.

A PROCESSOR-IMPLEMENTED RUNTIME LAW GOVERNS AUTONOMOUS AND SEMI-AUTONOMOUS SYSTEMS BY ENFORCING PROOF-BEFORE-ACTION. EACH <=100 MS CYCLE, A DETERMINISTIC PROCESSOR OR SECURE-ELEMENT MODULE COMPUTES SAFETY METRICS INCLUDING COHERENCE, INTEGRITY, ROBUSTNESS, STABILITY, AND TIMING; FORMS A NON-COMPENSATORY RESIDUAL (GAMMA); AND PERMITS ACTUATION ONLY WHEN GAMMA=0 AND A SIGNED, METRICS-ONLY EVIDENCE RECORD COMMITS IN THE SAME CYCLE. A FEDERATED HIGH COMMISSION AGGREGATES WINDOW SUMMARIES, COMPUTES FLEET CONSISTENCY AND COHERENCE, AND ISSUES OR REVOKES SHORT-LIVED TOKENS UNDER A STRICT-AND PERMISSION RULE. THE ARCHITECTURE YIELDS MACHINE-ENFORCEABLE PROOF-COUPLED SAFETY, BOUNDED REVOCATION, PRIVACY-PRESERVING AUDIT, AND DETERMINISTIC REPLAY, CONVERTING POLICY INTO EXECUTABLE LAW FOR AUTONOMOUS AND FEDERATED AI GOVERNANCE.

The invention provides a risk scene test method and system based on information uncertainty and a program product, and relates to the technical field of intelligent driving. According to the method, a standardized front vehicle braking basic scene is constructed and an automatic test engine is combined to systematically combine and accurately inject leak detection occurrence time and leak detection duration for sensing a leak detection event, so that comprehensive excitation and closed-loop test of behaviors of the intelligent driving decision-making system under information uncertainty disturbance are realized. According to the method, a large number of test cases covering a multi-dimensional disturbance space can be efficiently and repeatedly executed, key safety indexes of the system under different leak detection conditions are accurately captured, quantitative analysis of system risk response characteristics and clear definition of safety boundaries are achieved based on the mapping relation between the indexes and the input parameters, and the system risk detection accuracy is improved. And the detection capability of unknown unsafe scenes and the engineering level of SOTIF verification are improved.

The application discloses an intelligent body running state security deviation real-time detection and self-trap stop method and system, and belongs to the technical field of intelligent body security monitoring. The application carries out sliding window statistical analysis on the intelligent body running time sequence behavior log, parallelly feeds normalized multidimensional security indexes into at least two independent detectors to capture sudden abnormality and persistent slight deviation, determines the security deviation grade through deterministic fusion, and finally executes corresponding control actions through a hierarchical self-trap stop state machine containing an irreversible absorption state. The application determines the security baseline through offline learning period, does not depend on probability distribution hypothesis, random variable or iterative optimization throughout the whole process, always produces the same output for the same input, and provides an intelligent body with a deterministic, auditable and non-bypassable running state security monitoring and self-trap stop scheme.

The invention discloses an intelligent dispatching method and device, a storage medium and a program product. The intelligent dispatching method comprises the steps of data preprocessing, wherein gas volume data preprocessing is conducted through a data cleaning and data integration method; constructing a characteristic index system; building a gas flow prediction model based on an XGBoost model; building a gas volume automatic checking model based on the tree model; constructing an intelligent work order dispatching model; according to the invention, research is carried out based on gas big data, the cost of manual troubleshooting is reduced, and the recognition efficiency is improved; the method comprises the following steps of: constructing a gas utilization safety index by using data such as historical state monitoring data, historical data and potential safety hazard occurrence times of gas utilization equipment, reconstructing electric power historical big data, and performing deep mining on the historical data; a gas utilization safety hidden danger identification model is constructed in a mode of combining supervised learning and unsupervised learning, and automatic output of a gas utilization safety state of a user can be realized; and an intelligent dispatching platform is researched and developed, so that work order dispatching is quickly responded and properly completed.

Techniques for optimizing routing decisions based on security metrics within a network environment are described herein. In some cases, by using various security metrics, such as encryption indicators, attestation indicators, secureness metrics, and reliability metrics, an exemplary system can assess the security level and reliability of network paths. These metrics may provide valuable insights into the trustworthiness and integrity of participating nodes and links and enable informed decision-making regarding path selection.

This invention provides an event-driven method and system for evaluating power system frequency security indicators, applicable to modern power grids containing a high proportion of renewable energy. The method first parameterizes the system's dynamic frequency trajectory using an open-loop decoupling strategy, then constructs an event-driven piecewise analytical model. This model automatically handles high-order dynamics and strictly nonlinear elements within frequency regulation units. By combining the parameterized frequency trajectory with an iterative solver, it accurately calculates key frequency security indicators after a fault, including the time of the frequency minimum, the rate of frequency change, and the quasi-steady-state frequency deviation. This method approximates detailed closed-loop time-domain simulation in terms of evaluation accuracy, while simultaneously improving computational efficiency by tens of times, providing a highly efficient "white-box" analysis tool for online dynamic security assessment and primary frequency regulation reserve optimization.

Methods, systems, and computer storage media for providing brand oriented security management in an item listing system are described. By using a brand-oriented machine learning model, a brand-oriented security administrator engine supports the generation of different integrated reports and analyses of brand-related security metrics, including brand compliance, infringement events, law enforcement actions taken, and overall item list system integrity. The brand-oriented security administrator engine also supports visualization of brand-oriented security analysis result data in a manner that enhances availability and decision making by presenting complex brand-oriented security analysis result data in a clear, operable format. In operation, brand oriented security data is accessed. Brand-oriented security data is analyzed using a brand-oriented machine learning model trained based on a multi-dimensional authenticity analysis dataset. And generating brand-oriented security analysis result data. The brand-oriented security analysis result data is transmitted such that one or more visualizations are displayed based on the brand-oriented security analysis result data.

The invention belongs to the technical field of password target range safety operation and maintenance, and particularly discloses and provides a password target range health degree prediction and fault self-diagnosis method based on big data, which comprises the following steps: constructing a high-fidelity digital twin fusing a multi-level architecture and an operation log, defining a collaborative optimization model containing endogenous safety indexes, and performing collaborative optimization on the collaborative optimization model; the intelligent decision agent is trained through reinforcement learning, a verification execution strategy is deployed in a physical target range, and strategy increment retraining is achieved through actual operation and new attack and defense data backflow. According to the method, the fitting capability of the digital twinborn body to a real network environment is enhanced, the time sequence consistency of the digital twinborn body is ensured, fine management of security risks is reflected, closed-loop feedback between a physical target range and the digital twinborn body is realized, and an intelligent decision agent can continuously learn and adapt to a new scene.

The invention discloses an artificial universal intelligent security assessment method and system based on an AHP and a genetic algorithm. According to the scheme, firstly, a data acquisition module extracts data features related to security from an AGI model; then the index hierarchy construction module establishes a safety index system based on an analytic hierarchy process; the weight calculation module calculates an initial weight and performs consistency verification; if the consistency ratio (CR) does not meet the threshold value, the genetic algorithm module executes optimization and correction; the optimized weight is input into a grey clustering module to calculate a clustering coefficient and a comprehensive safety score; and finally, the result output module generates a safety assessment report and a risk level. According to the scheme of the invention, a hierarchical structure system of AGI security risks is established through AHP, and consistency correction and global optimization are carried out by using a genetic algorithm, so that evaluation weight calculation is more objective and stable; and meanwhile, a grey clustering model and a whitening weight function are combined to realize quantitative classification of different risk types, so that a comprehensive safety score with dynamic adjustability and interpretability can be output.

In some implementations, a system is provided for asynchronously performing and monitoring cybersecurity investigations. A scheduler selects event detection jobs to be performed during a specified time interval, and provides the selected event detection jobs to an orchestrator. For each event detection job, the orchestrator executes a corresponding event detection job worker instance. Performing the event detection job includes querying at least one event data log for event data that represents events that have occurred over a specified time range, processing the event data to generate a detected security metric, and in response to the detected security metric meeting a threshold value, publishing a cybersecurity alert to a cybersecurity investigation result repository. A response system subscribes to the cybersecurity investigation result repository to identify cybersecurity alerts that have been published to the repository, and in response to identifying a cybersecurity alert, performs a security response.