23 results about "Real world data" patented technology

The invention relates to the technical field of automobile intelligent driving, in particular to an intelligent network connection automobile digital twinning evaluation method and system for extreme scenes. The evaluation method comprises the following steps: constructing a knowledge graph; generating an extreme scene; the automatic driving system is tested in a virtual environment generated by the digital twin test platform based on an extreme scene; predicting a safety risk score in real time, and if the safety risk score exceeds a safety threshold, sending an intervention instruction to the digital twin test platform; performing anomaly detection on the full-quantity test log, deriving a scene and injecting the scene into a scene library if a performance inflection point is found; and calculating a root cause based on the weight vector, the system fault cause and effect graph and the test data, and generating a diagnosis result. According to the method, extreme scenes can be covered, the system can be scored from multiple dimensions, the performance bottleneck can be accurately positioned, and the closed loop of the test scene can be realized. According to the system, two-way closed loop and common evolution of physical test data and a virtual test environment are realized through a cloud side end architecture, and continuous synchronization and two-way interaction with real world data are realized.

The application discloses a prognosis risk prediction model construction system and method based on chronic disease real world data, comprising a multi-source real world data acquisition and access module, a data standardization and quality control module, a multi-dimensional feature engineering and screening module, a comorbidity feature fusion and data set construction module, a prognosis risk prediction model training and optimization module, a model multi-dimensional verification and evaluation module, a model deployment and risk stratification output module; a special multi-source real world data acquisition and quality control system for chronic liver disease, diabetes and hypertension is constructed, data standardization processing is realized in combination with a chronic disease clinical diagnosis and treatment guideline, through multi-dimensional missing value, abnormal value processing and data deduplication integration, the quality and availability of the real world data are effectively improved, the core problem of disorderly and low-quality real world data in the prior art is solved, and a high-quality data source basis is provided for model construction.

Examples including a system described herein implement a novel liquid time constant neural network (LTC-NN) based architecture to recover an underlying model of physical dynamics from real world data. The automatic differentiation property of LTC-NN nodes overcomes problems associated with low sampling rate, the input dependent time constant in the forward pass of the hidden layer of LTC-NN nodes creates a massive search space of implicit physical dynamics, the physics model solver based data reconstruction loss guides the search for the correct set of implicit dynamics, and drop out in dense layer ensures extraction of the sparsest model. Further, to account for perturbation timing error, the LTC-NN based architecture of the system utilizes dense layer nodes to search through input shifts that results in the lowest reconstruction loss.

The present disclosure relates to a method and apparatus for providing an off-label indication in an off-label indication provision system, and the method for providing an off-label indication may include collecting real world data from health care big data, collecting drug-related information, converting the drug-related information and the real world data to international standard terms, generating relationship information by analyzing an indication relevant to an active ingredient based on the drug-related information and the real world data that are converted to the international standard terms, filtering an indication with low relevance to the active ingredient from the relationship information, and providing an off-label indication based on the filtered relationship information.

The application discloses a follow-up content processing method, system, storage medium and electronic equipment. The hospital data of a patient is obtained through a preset follow-up mode, the hospital data and the hospital data obtained in advance are integrated through a preset integration mode, multi-dimensional patient health data is obtained, the multi-dimensional patient health data is used for representing multi-dimensional patient health data in a real world environment in a complete follow-up cycle, and a corresponding scientific research report is generated according to the multi-dimensional patient health data. Based on the above, the follow-up mode of online consultation dialogue, intelligent follow-up, telephone follow-up, incentive follow-up and the like is supported in the follow-up process of the patient who leaves the hospital, the doctor and the patient are facilitated to communicate, the electronic medical records, scientific research follow-up records, home equipment data and the like of the patient outside the hospital are integrated to form complete real world data and generate a scientific research report, the doctor is provided with scientific research and clinical decision support, the clinical data management efficiency is improved, and the scientific research implementation cost is reduced.

To support efficient use of RWD (RealWorldData).SOLUTION: The computer system is accessibly connected to a first database that stores clinical data acquired at a clinical site and including information on a treated patient as an item, and a second database that stores non-clinical data acquired for research purposes and including information on a treated patient group as an item. The computer system selects a predetermined number of pieces of clinical data from the first database, generates comparison data by aggregating the selected pieces of clinical data, executes statistical analysis for analyzing a difference between items using the non-clinical data and the comparison data having the same treatment content, and records a result of the statistical analysis.SELECTED DRAWING: Figure 9

The application discloses a cold start user recommendation method, system and device based on heterogeneous transfer learning, which comprises the following steps: using feature enhancement technology to combine source domain and target domain data into an enhanced matrix, then using an autoencoder to map the enhanced matrix to a latent feature space, performing feature matching and distribution alignment processing in the space, and finally reconstructing user features. Then, a solving model is established according to the reconstructed user features, that is, the solving model can be solved through matrix feature decomposition, thereby improving the recommendation accuracy for cold start users. In addition, the application also extends the model to a neural network, and the effectiveness of the model is verified through experiments on multiple real world data sets, and the performance index is excellent. The application can effectively utilize auxiliary source domain information to assist the recommendation task of the target domain, thereby providing more accurate and personalized recommendations for cold start users.

The application discloses a point of interest check-in sequence generation method based on a diffusion model, and belongs to the technical field of spatiotemporal data mining and deep learning. Data preprocessing is performed to clean and smooth a real world data set, so as to ensure data quality and consistency. Spatiotemporal lossless coding is performed to convert check-in sequences of different lengths into check-in vectors of equal lengths, and the original sequence information is retained by coding through a spatial frequency vector and a time bucket vector. A diffusion model is established to construct a spatial diffusion module and a time diffusion module, and spatiotemporal features are captured by using a forward diffusion process and a backward reconstruction process. A conditional U-shaped network is introduced into the diffusion module, and a denoising network is proposed to capture complex spatiotemporal correlations by modeling through a self-attention mechanism. A contrast learning strategy is used to further capture the spatiotemporal correlations of the check-in sequence by using ternary contrast learning, and the connection between the time and space diffusion modules is strengthened.

The invention provides a method and a system for constructing a clinical knowledge graph based on real world data. The method comprises the following steps: acquiring electronic health archives, medical insurance data, medical literatures, medical equipment data, patient report outcome and public health data, and carrying out cleaning, desensitization, medical term code mapping and structured conversion on the data; schema adaptive to the clinical field is designed, and entities and core relations of patients, diseases, drugs, symptoms, inspection and treatment schemes and the like are defined; entity, relation and attribute extraction is carried out by adopting a rule method in combination with machine learning, and knowledge fusion is completed through entity alignment, confidence fusion and conflict resolution; and the fused knowledge is stored in a knowledge base, and hidden knowledge is reasoned and mined by adopting rules, embedding or a graph neural network, is finally used for clinical decision support, drug alert and disease research, and is continuously updated along with new data and user feedback, so that the quality and practicability of the graph are improved.

The invention discloses a personalized robust recommendation defense method based on a dynamic reward mechanism, and the method comprises the steps: obtaining a real world data set related to a recommendation system, carrying out the cleaning of the data set, filtering invalid interaction data, retaining an effective user-article interaction pair, carrying out the normalization of scoring data, and dividing the scoring data into a training set, a verification set and a test set; constructing a robustness and fairness dual-objective influence evaluation mechanism, and quantifying an influence value of each user-article interaction sample on the model; the disturbance resisting amplitude is dynamically adjusted according to different user behavior characteristics; designing a reward mechanism integrating embedded interaction, behavior similarity and optimal batch, and guiding the model to dynamically balance defense and recommendation quality; and based on the filtered data set, training a model in combination with personalized adversarial disturbance and a multi-dimensional reward function. According to the method, pollution of poisoning samples to the model is effectively reduced, the adaptability of the model to heterogeneous user behaviors is improved, the attack resistance of the system is remarkably enhanced, and the method has wide applicability and practicability.

The application discloses a kind of chronic disease auxiliary drug curative effect evaluation and personalized recommendation method, comprising the following steps: step one: real world data acquisition, step two: data standardization and annotation, step three: curative effect evaluation model construction, step four: patient population stratification, step five: personalized drug recommendation, step six: data update and model iteration;The application is aimed at the clinical characteristics of three kinds of chronic diseases, liver disease, diabetes and hypertension, and constructs a multidimensional real world database and a targeted curative effect evaluation model, realizes the personalized and accurate recommendation of the auxiliary drug of the three kinds of chronic diseases, breaks through the limitation of single disease drug recommendation, effectively improves the clinical curative effect of auxiliary drug, reduces the risk of adverse drug reactions, and provides scientific support for the clinical rational use of drugs for the three kinds of chronic diseases.

The invention provides an electric power data virtual acquisition method based on a stochastic differential equation, and the method constructs a diffusion model network structure with a continuous time distribution modeling capability aiming at the problem that data missing is easy to occur in a new energy load data acquisition process. The method comprises the following steps: firstly, masking an input sequence through a random mask strategy in a training process so as to simulate a potential data missing scene, and carrying out supervised training on observed data; and then accurate virtual acquisition of missing values is realized through learning conditional probability distribution. On the basis, a stochastic differential equation (SDE) is adopted, noise is slowly injected, complex data distribution is smoothly converted into known prior distribution, and reverse time SDE of a time correlation gradient field which only depends on disturbance data distribution is adopted, and the prior distribution is converted into data distribution by slowly removing noise. Experimental verification on two real world data sets shows that the method is superior to an existing method in the aspect of virtual acquisition accuracy, and the superiority of the method is proved.

Methods and systems are provided for digitally replicating a real-world infrastructure to simulate the behavior of a real-world object, the methods and systems supporting the ability to extend a wireless network by providing services implemented in the network, and supporting optimization of network usage and digital representation accuracy. In one embodiment, the method includes instantiating one or more digital representation (D-Rep) instances corresponding to a digital replica of the real world object included in the real world infrastructure to form a digital infrastructure (D-Inf), and generating one or more digital representation (D-Rep) instances corresponding to a digital replica of the real world object included in the real world infrastructure. The method further includes providing a communication plane to provide a service to the D-Inf. The method further includes connecting to the real world object using the communication plane to collect real world data defining a state of the real world object, and associating the real world data with the D-Rep instance; the behavior of the real-world object is simulated by the D-Rep instance using a simulator that can be accessed by the D-Inf instance using the communication plane.

The invention belongs to the technical field of clinical drug treatment monitoring, and discloses a risk prediction model for predicting the limit exceeding of the steady-state blood concentration of linezolid, and the model is a column diagram model constructed based on multi-factor Logistic regression analysis. The risk probability calculation module is used for calculating the risk probability that the blood concentration is greater than or equal to 8 mg / L according to clinical indexes of a patient; the input variables of the model comprise age, albumin level and serum creatinine level; the albumin level and the serum creatinine level are the albumin level and the serum creatinine level detected last time before blood concentration monitoring of the patient. The method is constructed based on real world data and verified by an independent queue, the model shows good distinguishing ability and calibration degree, and the prediction accuracy is good. The model only contains three clinically conventional easily available indexes of age, albumin and serum creatinine, and is presented in the form of a column diagram without complex calculation, so that clinical doctors can quickly complete risk assessment at the bedside.

The invention discloses a real world data-based personalized follow-up visit strategy generation and optimization method for patients with type 2 diabetes mellitus, which comprises the following steps of: calculating a patient state transition probability by utilizing a Markov decision process model through integrating demographic characteristics, a baseline health state, a laboratory examination result and life style data; and further, an individual follow-up visit frequency and a follow-up visit mode are dynamically generated. When a patient's clinical event or health status changes, the method may adjust the follow-up plan in real time. Compared with an existing standardized follow-up visit method, the method has remarkable advantages in the aspects of reducing the occurrence rate of complications, prolonging the health life and reducing the medical cost, is particularly suitable for basic medical institutions and resource-limited environments, and has universality and popularization value.

This paper discloses a method for detecting and correcting AI / ML model drift in Robotic Process Automation (RPA). It analyzes information related to the input data of the AI / ML model to determine if data drift has occurred, analyzes information related to the results of the AI / ML model's execution to determine if model drift has occurred, or both. Based on the analysis of this information, the AI / ML model is retrained when changes in conditions are detected, when a change threshold is met or exceeded, or both. The retrained AI / ML model can then be deployed to provide better predictions for real-world data.

The application discloses a kind of based on quantum heuristic algorithm's mobile edge computing task scheduling method, belong to network communication technical field. Including the following steps: construction scene;System architecture;System parameter update;Problem construction;Solution method.The application is used in the task scheduling optimization in mobile edge computing (MEC) environment, by integrating the concept of quantum mechanics into double deep Q network (DDQN), successfully improves the adaptability and robustness of task scheduling.QI-DRL algorithm is demonstrated in theory The advancement of, also by extensive experimental verification with real-world dataset Its actual application effectiveness.The application also provides a new perspective and technical path for handling resource management and task scheduling problems in MEC environment, so that in the case of high demand change and resource limitation, the system can more effectively manage and schedule tasks, significantly optimize the operation efficiency and response speed in MEC environment.

The application discloses a kind of semi-supervised semantic segmentation method, equipment and its storage medium.Establish picture data set by obtaining picture data in real world data, build depth model and semi-supervised semantic segmentation model, predict to obtain deep information and semantic segmentation information, preset threshold, filter out pixel area in any two categories in single picture by preset threshold, obtain depth set and logarithm difference set, obtain intra-class logarithm difference loss according to depth set and logarithm difference set, calculate the regularization loss of logarithm difference;Obtain unlabelled data, establish strong enhanced view set and weak enhanced view set, obtain robust supervision signal;Exponential normalization is used to suppress large abnormal fluctuations for intra-class depth difference, and weight is adaptively assigned based on entropy, so as to promote robust feature learning and learn rich discriminative information.The application promotes prediction consistency to maximize the use of unlabelled data and further improve model performance.

Methods and systems for predictive signal occlusion detection using a digital replica of a communication network are provided. The method includes receiving a request for predictive occlusion detection (PBD) within a fragile region covered by the wireless network, the predictive occlusion detection indicating future signal degradation within the fragile region. The method further includes retrieving entity data associated with one or more digital representations (D-RIPs) corresponding to a digital replica of a real-world entity, the entity data including real-world data collected from the real-world entity. The method further includes performing the predictive occlusion detection based on the retrieved entity data and sending an occlusion alert when an occlusion is predicted. The method provides interoperability, efficient access and utilization of mass data, and the ability to derive timely and accurate predictions and schemes.

Disclosed is a computer-implemented system (110) for providing a virtual reality platform for a virtual environment that has characteristics of a real-world. The system (110) comprises a processor(s) (210) that obtains, from one or more user devices, user profile information and real-world data corresponding to a real-world user and thereby generates at least a virtual clone of the real-world user. The processor (210) further integrates the virtual clone with the virtual environment and monitors aging of the virtual clone in the virtual environment. Thereafter, the processor (210) tracks, with respect to a real-world time, one or more activities of the generated virtual clone based on a plurality of fundamental clone factors corresponding to the real-world user. Additionally, the processor (210) modifies an appearance of the generated virtual clone based on the monitoring of the aging and the tracking of the one or more activities.

A system and associated methods extend neural architectures such as liquid time constant neural network (LTC-NN) or continuous time recurrent neural networks (CT-RNN) or neural ordinary differential equations (NODE) to obtain advanced neural structures (LTC-NN-MR, CT-RNN-MR, NODE-MR) that can recover model coefficients of a dynamical system under low sampling rate conditions. The forward pass of these advanced neural structures has the same form as bilinear approximations of nonlinear dynamics. Measurements of real data can be used to convert the set of non-linear dynamics to an over-determined system of equations that are linear in terms of the model coefficients.

The invention relates to a safety risk early warning method and system after traditional Chinese medicine listing. The method comprises the following steps: constructing'patient-traditional Chinese medicine-adverse event 'original data based on multi-source heterogeneous real world data; the method comprises the following steps: preprocessing patient-traditional Chinese medicine-adverse event original data to obtain a structured data set; extracting various characteristic values of the structured data set and constructing a characteristic matrix; training a learning model by using the feature matrix to obtain a traditional Chinese medicine monitoring model; and completing anomaly detection of the target traditional Chinese medicine by using the traditional Chinese medicine monitoring model. Through advanced technologies of natural language processing, deep learning, multi-source heterogeneous data fusion and the like, accurate mining and dynamic evaluation of potential adverse reaction signals in real-world complex scenes of traditional Chinese medicines are realized, and the accuracy and timeliness of signal identification and the adaptability of traditional Chinese medicines can be remarkably improved; meanwhile, the active monitoring capability and robustness of the system are enhanced, and the complex conditions of multiple traditional Chinese medicine components, drug combination, crowd heterogeneity and the like can be flexibly handled.

The invention discloses a semi-supervised semantic segmentation method and device and a storage medium thereof. The method comprises the following steps: acquiring picture data in real world data to establish a picture data set, building a depth model and a semi-supervised semantic segmentation model, predicting to obtain deep information and semantic segmentation information, presetting a threshold value, screening out pixel regions for any two categories in a single picture through the preset threshold value, obtaining a depth set and a logarithmic difference set, and then obtaining a depth set and a logarithmic difference set; obtaining intra-class logarithmic difference loss according to the depth set and the logarithmic difference set, and calculating regularization loss of logarithmic difference; obtaining label-free data, establishing a strong enhancement view set and a weak enhancement view set, and obtaining a robust supervision signal; the intra-class depth difference is subjected to exponential normalization to suppress large abnormal fluctuation, and weights are adaptively allocated based on entropy, so that robust feature learning is promoted, and rich discrimination information is learned. The present invention promotes prediction consistency to maximize the utilization of label-free data and further improve model performance.