29 results about "Risk behavior" patented technology

The invention discloses a financial fraud behavior intelligent tracking and early warning system, and the system comprises an obtaining module which is used for obtaining a risk behavior in a financial interaction event, and a plurality of interaction objects and a plurality of interaction behavior signals corresponding to the risk behavior; the construction module is used for constructing a risk data chain based on the plurality of interaction objects and the plurality of interaction behavior signals; the first determination module is used for determining risk information of each risk node in the risk data chain; the judgment module is used for judging whether financial fraud behaviors exist or not according to the risk information of each risk node; and the tracking and early warning module is used for carrying out intelligent tracking and early warning on the financial fraudulent behavior according to the risk data chain when the financial fraudulent behavior is determined to exist. The risk information of the risk node is locally determined based on the risk data link, whether financial fraud behaviors exist or not is judged from the whole risk data link, and accurate tracking and early warning are achieved.

The invention discloses an internal control risk dynamic identification method and system based on a multi-modal model, and relates to the technical field of risk dynamic identification methods. According to the method, through an unsupervised state transition analysis mode, the hidden abnormal jump paragraph in the behavior mode can be dynamically identified, manual rule setting is not needed, and the self-adaptive capability of risk identification is enhanced; when a time lag problem exists in a multi-modal log, through operation chain backtracking and a virtual calibration mechanism, the time consistency and causal integrity of a cross-modal behavior chain are ensured, and the misjudgment probability caused by record inconsistency is remarkably reduced; a state sequence is mapped into a process topological graph by adopting a sliding matching method, free fragments which cannot be matched structurally are effectively identified, and starting and ending points of risk behaviors can be accurately positioned.

The embodiment of the invention provides a risk behavior recognition method and device, equipment, a storage medium and a program product, and relates to the field of big data. The method comprises the following steps: receiving encrypted entity relationship data uploaded by a plurality of mechanisms; using a federated learning algorithm to generate fusion features of a plurality of pairs of transaction entities according to the encrypted data uploaded by the plurality of mechanisms; constructing a dynamic graph according to the fusion features of the plurality of pairs of transaction entities; dividing the dynamic graph into a plurality of sub-graphs; and identifying risk behaviors existing in each sub-graph. According to the method provided by the invention, by using the federated learning algorithm, the encrypted entity relationship data of the multiple mechanisms are fused, so that cross-mechanism risk behavior identification is effectively realized on the basis of guaranteeing the privacy of the user.

This invention relates to the field of artificial intelligence technology, specifically to an AI-based medical information consultation method and apparatus. The method includes: acquiring multimodal medical consultation input from a user; preprocessing and extracting features from the multimodal input, and fusing them to generate a unified multimodal feature representation; based on the feature representation, performing topic indexing in a pre-constructed medical knowledge graph, and extracting a set of highly relevant medical knowledge using unsupervised clustering analysis and a dynamic threshold filtering mechanism; inputting the filtered medical knowledge set and multimodal features into a medical large-scale language model for logical reasoning, and running a multi-turn dialogue intent tracking algorithm during the reasoning process to complete missing entities; intercepting high-risk behaviors and correcting compliance issues in the preliminary response generated by the large-scale language model, and outputting a safe and professional medical information consultation response. This invention significantly improves the cross-modal perception capability, logical reasoning accuracy, and clinical application legitimacy of medical AI.

The invention discloses a behavior early warning method, device and equipment, a medium and a product, and relates to the technical field of intelligent supervision. The method comprises the following steps: acquiring multi-modal behavior data and scene feature information of a supervised target, wherein the multi-modal behavior data comprises positioning track data, physiological data and supervision behavior data; according to a pre-training weight model and the scene feature information, determining a dynamic weight matrix of a supervision scene to which the supervised target belongs; and according to the dynamic weight matrix, the multi-modal behavior data and a pre-training behavior analysis model, determining a risk behavior assessment result of the supervised target and performing early warning. Scene feature information is converted into quantifiable weight parameters, and then multi-modal behavior data fusing positioning, physiology and supervision behavior data are combined to predict and evaluate risk behaviors, so that an evaluation result is obtained. The risk focuses of different supervised targets can be finely distinguished, and evaluation is carried out through multi-modal data, so that the evaluation result is more accurate and reliable.

This invention discloses a method, system, and electronic device for personnel operation assessment based on video temporal joint analysis, relating to the field of power safety technology. This application involves acquiring video from the work site and extracting frames; extracting the sequence of key human points from each frame; dynamically defining risk areas for the head, torso, and hands based on these key points; calculating hand movement speed, distance from the risk area, and movement consistency characteristics; triggering refined behavior recognition when hand movement characteristics indicate a tendency to enter the risk area; extracting multi-scale temporal features before and after the trigger; and judging specific risk behaviors using a pre-trained action recognition model. This invention, through a two-level mechanism of "trend triggering + refined recognition," achieves early warning and accurate identification of risk behaviors such as removing helmets and gloves, effectively improving the real-time performance and preventative capabilities of work safety monitoring. This application realizes a shift from post-event identification to pre-event warning, resulting in better safety monitoring performance.

The invention relates to the technical field of deep learning, and discloses a behavior monitoring and risk early warning system for obstetrical patients, which comprises a data sensing acquisition module, a data analysis processing module, a human body area modeling module, a behavior early warning module, an edge calculation module and a comprehensive judgment module. According to the system, pixel-level segmentation and key point regression are performed on patient image data by constructing a mask key point-continuous time attention fusion network, a continuous time attention mechanism inspired by a neuron loop is introduced, dynamic modeling and weight distribution are performed on continuous frame attitude features, and a continuous frame attitude fusion model is constructed. Therefore, the continuous time evolution trend of the posture change of the obstetrical patient is described. And in combination with human body area modeling and behavior early warning strategies, intelligent identification and early warning of risk behaviors such as postpartum dysphoria, abnormal turning over and falling down are realized. The accuracy and stability of behavior monitoring are improved, and the method is suitable for intelligent safety management in an obstetrical monitoring scene.

The invention provides a model training method, a risk behavior identification method, a system, equipment and a medium. The training method comprises the following steps: acquiring a training data set; the training data set comprises image data and text data; based on the training data set, utilizing a multi-stage distillation training strategy to train a risk behavior identification model, and obtaining a trained risk behavior identification model; the multi-stage distillation training strategy comprises the step of carrying out distillation training on semantic representation, attention distribution and / or risk label stages of the risk behavior recognition model. According to the method, the risk behavior recognition model is trained by using the multi-stage distillation training strategy, so that the memory and computing power consumption is remarkably reduced while the multi-task performance of a small model is close to the level of a large model, and the problems of large parameter quantity and difficulty in edge deployment of a traditional view language model for recognizing risk behaviors in the prior art are solved.

The application discloses an Internet of Things-based safety tool intelligent management full-life-cycle tracing system, and relates to the technical field of warehouse management.The system comprises an Internet of Things data acquisition module, which is used for acquiring environmental parameters and tool use behavior data; a behavior feature extraction module, which is used for feature extraction to obtain a behavior feature sequence; a risk behavior mining module, which is used for statistically obtaining risk behavior features of a fault tool based on the behavior feature sequence, defining a risk behavior graph, and learning embedding vectors and attention weights of the risk behavior by using a graph convolution network; a risk level prediction module, which is used for encoding real-time risk behavior features by using the embedding vectors and the attention weights, and predicting risk levels of the encoded real-time risk behavior features by using an LSTM model; and a risk tracing update module, which is used for generating an event chain and an early warning log when the risk level is greater than or equal to a risk level threshold, and updating the risk behavior graph and the LSTM model, so that potential tool use risks can be found in advance.

The invention discloses a drug rehabilitation place management and control method based on a security and protection Internet of Things sensing platform, and relates to the technical field of intelligent management and control. Deducing the behavior trajectory of the drug addict by using a neural differential prediction model based on the video stream to obtain a behavior trajectory evolution path coordinate sequence, establishing a high-risk behavior pattern library based on historical relapse action feature rules, and marking feature fragments in combination with the behavior trajectory evolution path coordinate sequence; establishing a symbol rule base based on the historical behavior characteristic data, and generating a risk report according to the characteristic fragment, the behavior track evolution path coordinate sequence and the symbol rule base; matching an emergency disposal scheme according to the risk report, positioning the position of a high-risk person, and scheduling the nearest security resource to execute a hierarchical intervention action; according to the method, the spatial position rule, the motion mode rule and the physiological response rule are integrated through the symbol rule base, so that collaborative analysis of multi-source heterogeneous data and accurate matching of risk characteristics are realized.

The invention discloses a video monitoring personnel risk behavior identification method and system based on deep learning, and belongs to the technical field of risk behavior identification, and the method comprises the steps: obtaining a plurality of monitoring videos, and carrying out the preprocessing of the monitoring videos, and obtaining a target data set; constructing a risk behavior recognition network, and training and testing the risk behavior recognition network according to the target data set; obtaining a target video frame set, and processing the target video frame set according to the risk behavior recognition network to obtain a risk behavior recognition result corresponding to the target video frame set; the multi-modal features and the spatial-temporal features are integrated through the fusion of the enhanced fusion module and the spatial-temporal feature extraction module, and the recognition capability of subtle actions and complex scenes can be improved, so that the recognition precision of the risk behavior recognition network is enhanced.

The invention discloses a user risk behavior early warning method and device for an operating system, a medium and a product, and relates to the technical field of computers, and the method comprises the steps: monitoring a target user behavior of a target operating system, and inputting the obtained information into a preset risk early warning network, matching the corresponding target monitoring information by sequentially utilizing each locally pre-configured target knowledge base through a preset risk early warning network, and if each piece of target monitoring information is successfully matched in the corresponding target knowledge base, determining a preset weight and a target risk coefficient corresponding to each piece of target monitoring information by utilizing the target knowledge base, and finally, determining a risk entropy value corresponding to the target user behavior according to a preset weight value and the target risk coefficient, and determining whether risk behavior early warning is performed or not based on the risk entropy value. Therefore, risk behavior early warning can be performed according to various operation information of the user, so that the damage of the risk behavior of the user to the operation system is effectively avoided.

The application provides a model training method, a risk behavior identification method, a system, a device and a medium. The training method comprises: acquiring a training data set; the training data set comprises image data and text data; a risk behavior identification model is trained based on the training data set by using a multi-stage distillation training strategy, and a trained risk behavior identification model is acquired; the multi-stage distillation training strategy comprises distillation training on a semantic representation, an attention distribution and / or a risk label stage of the risk behavior identification model. The application trains the risk behavior identification model by using the multi-stage distillation training strategy, so that the small model has a multi-task performance close to the level of a large model while significantly reducing memory and computing power consumption, thereby solving the problems of a large number of parameters of a conventional visual language model for identifying risk behaviors and difficulty in edge deployment in the prior art.

The invention provides a vehicle risk intervention method and system based on real-time multi-source data and a knowledge graph, and the method comprises the steps: collecting multi-source heterogeneous data and environment data in real time, and generating a real situation feature according to the environment data; and performing space-time alignment on the multi-source heterogeneous data to generate a standardized event stream. Constructing a dynamic knowledge graph based on the ontology definition; and carrying out risk behavior identification based on a heterogeneous graph neural network and time sequence convolutional network joint model, outputting a risk category and a risk probability, and outputting the risk category and the risk probability. And matching a corresponding intervention action in a pre-constructed intervention knowledge graph according to the risk category and the real situation feature, and issuing and executing the intervention action. And acquiring intervened data, evaluating an intervention effect through a causal inference algorithm, and feeding back and updating the joint model and the intervention knowledge graph. According to the invention, vehicle risks can be rapidly identified, graded early warning based on situations and adaptive accurate intervention can be realized, and secondary accidents and false alarm rates can be reduced.

The invention discloses a risk behavior detection and alarm method and system based on log analysis, and belongs to the technical field of big data security analysis and big models.The method comprises the steps that log collection is conducted, specifically, a multi-source log is captured in real time through a distributed agent and standardized; feature engineering: extracting time, behavior, statistics and session features and constructing feature vectors; calculating an operation interval and a non-working period mark according to the time characteristic; according to the behavior characteristics, the difference between high-frequency operation and the geographic position is counted; analyzing session duration and an operation sequence mode according to the session characteristics; reasoning by adopting a logistic regression model: outputting a risk score based on the feature vector by adopting the logistic regression model; and graded alarming: triggering alarms of different grades according to the risk score. According to the invention, accurate identification and rapid disposal of system risk behaviors can be realized.

The application discloses a multi-factor fusion dangerous behavior early warning method and system, and particularly relates to the technical field of dangerous behavior early warning, which comprises the following steps: acquiring multi-source abnormal data for a target object or a target area in the same early warning period, extracting an abnormal time, an abnormal position, an abnormal category and a change direction, and generating an abnormal event set; reading the abnormal event set, and sequentially comparing the abnormal position, the abnormal category and the change direction of each abnormal event according to the abnormal time. The abnormal event is further written into a trigger event, a support event, an exclusion event, an escalation event and a release event, and the entering qualifications of different abnormal signals are distinguished according to the events, so that weak evidence, conflicting evidence and escalation evidence are no longer directly triggered in the same position. Therefore, the application helps to realize the orderly competition, progressive release and credible escalation of multi-theme early warning in the same object or area, and relatively alleviates the problems of early warning sequence disorder and insufficient result credibility.

The invention relates to the technical field of public safety risk pre-judgment, in particular to a public safety risk pre-judgment method and system for crowd emotion recognition. The method comprises the steps of performing real-time feature monitoring on an undetermined area, collecting feature parameter values, establishing a risk prediction model, and calculating risk scores at different time stages; and performing tropism parameter marking on the specific region, and dynamically adjusting the risk prediction model according to a marking result. According to the method, real-time feature monitoring is carried out on an undetermined area, feature parameter values are collected, a risk prediction model is established, risk scores in different time stages are calculated, tropism parameter marking is carried out on a specific area, and specific area feature recognition is carried out on the specific area. And the risk prediction model is dynamically adjusted according to the marking result, the risk behavior and the non-risk behavior are specifically divided according to different scenes, and the overall pre-judgment accuracy is improved during specific risk pre-judgment.

The application discloses a target object risk behavior early warning method and related equipment. In the specific method, emotion analysis is first performed based on a preset emotion analysis model and a plurality of historical information published by a target object, and then it is determined whether the target object is emotionally abnormal. Then, a preset risk behavior analysis model is used to determine the risk behavior type that the target object is most likely to have. Finally, information mining is performed on the historical information corresponding to the risk behavior type, so that various types of key information related to the risk behavior type are clearly presented. Finally, risk behavior early warning of the target object is realized based on the key information related to the risk behavior, which can perform deep risk behavior information mining on the emotionally abnormal object and perform early warning based on the mined risk behavior information, thereby improving the accuracy of risk behavior early warning of the target object.

The invention provides a risk service identification method and device, equipment, a storage medium and a program product, and relates to the technical field of network security, a risk user terminal is identified through a pre-constructed risk application library, and risk behaviors are identified through a pre-constructed risk behavior library, so that automatic identification of risk service data is realized, and the user experience is improved. And the identification efficiency of the risk business data is improved. The risk behavior library is constructed according to the key features of the historical risk services, and the efficiency of identifying the risk behaviors is improved. And the risk application library is constructed according to the risk tool of the historical risk business, so that the efficiency of identifying the risk user terminal is improved. According to the method and the device, the risk business data can be identified only by extracting the user terminal and the user behavior of the business data, the privacy information of the user does not need to be acquired, and the privacy of the user is protected while the risk business is identified.

The embodiments of the present specification describe a training method of a behavior prediction model, a risk behavior prediction method and device. According to the method of the embodiments, the type identification of a sample behavior event and the time information of the sample behavior event can be obtained when training the behavior prediction model. Then the sample behavior event is characterized in the continuous time domain, and then the behavior prediction model is trained according to the behavior event characterized in the continuous time domain, so as to optimize the prediction value of the type identification and the time output by the behavior prediction model. By characterizing the sample behavior event in the continuous time domain, the association between the behavior event and the time when it occurs is realized, so that the model can fully learn the regularity and periodicity characteristics presented by the behavior event and the time when it occurs, thereby improving the accuracy of risk behavior prediction.

The application provides a risk behavior recognition method and device, relates to the technical field of artificial intelligence, and can be applied to the technical field of finance or other technical fields. The risk behavior recognition method comprises the following steps: acquiring a video image, extracting a skeleton sequence from the video image; extracting skeleton key point information features and skeleton connection relationship features from the skeleton sequence; inputting the skeleton key point information features and the skeleton connection relationship features into a first behavior prediction model and a second behavior prediction model respectively, and obtaining a first behavior and a second behavior respectively; and obtaining a risk behavior recognition result according to the first behavior and the second behavior.

An ICU patient abnormal behavior recognition method and system based on computer vision, relating to the technical field of computer vision, comprising: acquiring a continuous image sequence of a monitored area by a camera device; extracting motion parameters of a patient in the monitored area according to the continuous image sequence; according to the motion parameters, identifying corresponding behavior nodes according to a preset rule; according to a preset risk behavior logic rule, logically associating the identified behavior nodes in time sequence, matching a risk behavior causal chain representing a preset risk behavior; when the risk behavior causal chain is successfully matched, outputting an alarm information. By constructing the causal chain of the risk behavior, even if a key action cannot be directly observed due to obstruction, the complete risk event can still be inferred through other observable behavior nodes before and after it, such as sitting up and moving the limbs to the dangerous area, effectively reducing the false negative rate.

The invention discloses a risk behavior identification method and apparatus, an electronic device and a storage medium. The method comprises the steps of determining a to-be-processed fence region; determining a first risk behavior recognition result based on the to-be-processed object in the to-be-processed fence area through a supervised first behavior recognition model, and / or determining a second risk behavior recognition result based on the to-be-processed object in the to-be-processed fence area through an unsupervised second behavior recognition model; and based on the first risk behavior identification result and / or the second risk behavior identification result, determining whether a risk behavior occurs in the fence area to be processed. According to the method, the sensitivity, the accuracy and the adaptability of risk behavior identification can be improved, the false alarm rate is reduced, and the method has a good application effect especially for a complex dynamic scene.

The invention discloses a multi-factor fusion dangerous behavior early warning method and system, and particularly relates to the technical field of dangerous behavior early warning, and the method comprises the steps: obtaining multi-source abnormal data for a target object or a target region in the same early warning period, extracting an abnormal moment, an abnormal position, an abnormal category and a change direction, and generating an abnormal event set; and reading the abnormal event set, and comparing the abnormal position, the abnormal category and the change direction of each abnormal event in sequence according to the abnormal time. Abnormal events are further written into a trigger event, a support event, an exclusion event, an upgrade event and a release event, and entry qualifications of different abnormal signals are distinguished according to the events, so that weak evidence, conflict evidence and upgrade evidence do not directly trigger an alarm at the same position any more; therefore, orderly competition, progressive release and credible upgrading of multi-theme early warning can be realized in the same object or area, and the problems of disordered early warning sequence and insufficient result credibility are relatively relieved.

The invention discloses a construction site risk perception and dynamic management and control system integrating BIM and an intelligent safety belt, and relates to the technical field of building construction safety monitoring. Comprising a data acquisition and behavior identification module, an AI risk behavior evaluation and scoring module, a BIM model fusion and space mapping module, a three-dimensional risk thermodynamic diagram generation module, an AI behavior learning and trend prediction module and a management linkage module. According to the method, through the intelligent safety belt and the AI behavior recognition module, dangerous behaviors such as high-place operation are monitored and judged in real time, and the timeliness and initiative of safety management are greatly improved. According to the method, precise spatial mapping is realized in combination with the BIM model. According to the method, the dynamic three-dimensional risk thermodynamic diagram is generated, the three-dimensional thermodynamic diagram can be dynamically updated according to the risk behavior frequency and grade of different areas and different components, the potential safety hazard distribution and change trend are visually displayed, and a manager can visually perceive and make a judgment conveniently.

The invention discloses a risk number identification method based on adaptive learning. The method comprises the following steps: acquiring encrypted call behavior data of a suspected risk behavior number; extracting call behavior characteristics of the suspected risk behavior number from the call behavior data, expanding data statistical characteristics of the suspected risk behavior number through a statistical method, and performing data cleaning and filling preprocessing on the expanded data to obtain a data characteristic set; constructing a soft threshold objective function as a risk number behavior recognition model, calculating a target value y by taking the preprocessed data feature set data as a feature X of the objective function, and performing iterative updating and tuning on parameters of the risk number behavior recognition model in combination with a cross entropy loss function to obtain a final risk number behavior recognition model; and identifying the data feature set by using the final risk number behavior identification model, generating a suspected risk number list, and performing early warning on suspected risk numbers. According to the method, the identification effect of the risk number is effectively improved through adaptive learning of parameters and multiple iterations.

The invention discloses a risk identification method for elderly people living alone in a community based on multi-task collaborative learning. The method comprises the following steps: collecting a multi-source heterogeneous data set of elderly people living alone in the community and preprocessing the multi-source heterogeneous data set; performing data feature extraction on the preprocessed data; taking the extracted data features as the input of a pre-trained multi-task collaborative expert network; selecting different experts and distributing expert weights according to task characteristics by using a customized multi-task gating unit network based on a time attention mechanism; performing anomaly judgment on the input data through the selected expert network; fusing the output of each expert network to obtain a fused feature; judging the risk score of the corresponding specific task through the task tower according to the fusion feature; and comprehensively balancing the risk assessment result through a fuzzy comprehensive evaluation method to obtain a final comprehensive assessment value. According to the method, the coordination work of multi-source data is realized, the problems of missed judgment in a high-risk period and misjudgment in a low-risk period are solved, and the risk behavior prediction accuracy is improved.

The invention provides a low-voltage operation capability evaluation and optimization method based on fusion operation behavior data, and the method comprises the following steps: carrying out the multi-dimensional risk feature labeling of operation behavior data, and recognizing and classifying an electric shock contact behavior and a key operation response delay behavior as a high-risk abnormal behavior; an enhanced sample is generated by adopting an expansion sampling method, and the proportion of high-risk abnormal behaviors in a data set is increased; setting risk sensitivity according to the data source type; adjusting an abnormal identification threshold and a screening priority, and setting the high-risk abnormal behavior as the highest retention level; retaining data screened as abnormal data and marking the data as high-risk key samples; inputting a high-risk key sample and a normal sample into an evaluation model, and adjusting the influence on model training and reasoning; and comparing an original label according to a model output result to obtain a self-adaptive optimization result of the recognition precision and the evaluation effect. According to the invention, the recognition precision and the safety decision reliability of the operation capability evaluation system on high-risk behaviors can be improved.

The present invention provides an image-based patient risk behavior monitoring system for, by using a patient analysis model, identifying a patient and objects, which are subjects to be monitored, in an image acquired through a camera, detecting the location and posture of the patient and an object interacting with the patient on the basis of the image, determining whether the behavior of the patient fits into the category of risk behavior on the basis of the location and posture of the patient and the interacting object, and if the behavior of the patient comes under the category of risk behavior, generating a risk behavior event and storing the type, start time, and end time of the risk behavior in a risk behavior log.