151 results about "Native client" patented technology

The present invention provides a traffic flow prediction method and apparatus, a medium, and a device. The method comprises: collecting historical traffic flow data in a target area; constructing initial prediction models in a central server; performing distributed collaborative training on the initial prediction models at local clients to obtain trained initial prediction models, and inputting the trained initial prediction models into the central server to perform quality screening and aggregation; performing credibility calculation on each local client at which the corresponding initial prediction model satisfying preset quality requirements is trained, and performing weight parameter customization on an aggregated initial prediction model to obtain a traffic flow prediction model; and inputting the historical traffic flow data into the traffic flow prediction model for traffic flow prediction to obtain a traffic flow prediction result in the target area. The traffic flow prediction model can be trained without sharing data of all parties and invading privacy, and the traffic flow in the target area is predicted by means of the traffic flow prediction model having relatively high performance, thereby improving the prediction precision of the traffic flow.

The invention belongs to the technical field of federated learning privacy protection, and particularly relates to a heterogeneous federated learning adaptive privacy protection method and device based on momentum optimization. The method comprises the steps that feature extractor parameters broadcasted by a central server are received, the local gradient of a current iteration round is calculated based on a local data set, and a momentum mechanism is introduced into local gradient updating for a classifier part; a gradient descent algorithm is applied, and local model parameters are updated through the local gradient obtained through calculation; for the feature extractor part, calculating a Fisher information matrix; performing normalization processing on the Fisher information of each layer, and screening strategies; and then differential privacy processing is carried out, the global model parameters are uploaded to a central server for parameter aggregation, and updated global model parameters are broadcasted back to each local client for next round of training. According to the method, the problems of negative influence of a differential privacy mechanism on model training, slow model convergence and poor privacy protection effect are solved.

The invention relates to a federal location recommendation method based on heterogeneous graph semantics, which comprises the following steps: constructing a corresponding local heterogeneous graph by using a local private data set, learning POI node features in the local heterogeneous graph by using a heterogeneous neural network, and learning meta-path semantic features through an attention mechanism. Fusing the track sequence features and the corresponding meta-path semantic features to train a local model; and fusing the parameters of the local model obtained after the training into global information through a knowledge distillation method. And encrypting and uploading the local model parameter, the local data volume parameter and the training result after the knowledge distillation of this round to a server. And after the server receives the uploads of all the clients, global aggregation is carried out, and global model parameters are updated by using the aggregated parameters, so that one-time training is completed. And for a new user, inputting the historical trajectory data of the new user into the optimal local model on the corresponding local client, and outputting the historical trajectory data as the next point of interest recommended for the new user.

A WTRU may receive a first message comprising one or more criteria associated with an application server. If one or more local application servers satisfy the one or more criteria, the WTRU may send a second message to the local client device. The second message may comprise information for accessing the one or more local application servers in the local network. If no local application servers in the local network satisfy the one or more criteria, the WTRU may send a third message comprising the one or more criteria to a central application management server. The WTRU may receive a fourth message from the central application management server indicating one or more non-local application servers that satisfy the one or more criteria. The WTRU may send a fifth message to the local client device with information for accessing the one or more non-local application servers.

The invention provides an efficient heterogeneous client federated learning method based on knowledge distillation. The efficient heterogeneous client federated learning method comprises the following steps: selecting a chairman model, namely an RC model, of a training round from client models participating in federated learning by using a heuristic chairman selection strategy; uploading the selected chairman model to a server; sending the chairman model to each local client; learning knowledge in the local model through knowledge distillation; the knowledge is obtained from a client model through local data training learning; uploading the trained chairman model to the server again; the server collects and aggregates the chairman models uploaded by the client to obtain the aggregated chairman model; redistributing the aggregated chairman model to each client; the chairman model guides the client model to update, and the client model is trained; and judging whether 200 rounds of training are reached or not, if not, returning, and if so, ending.

The invention discloses a federal learning method and system based on global prototype guidance. The method comprises the following steps: a server sends a global model and a global prototype to a local client; the local client performs rebalance comparison learning by using the global prototype, and updates the local model; the local client generates a balanced confrontation positive prototype instance set and trains a classifier; the local client side calculates a local prototype and uploads the local prototype and the updated local model to the server side; and the service aggregates the updated local model and local prototype, and updates the global model and global prototype. The system comprises a server and a local client. According to the method, the Non-IID problem in federated learning is solved, and the accuracy and generalization of the model are improved. The method can be widely applied to the technical field of distributed machine learning.

Provided in the present application are a large-model training method, apparatus and device, and a storage medium. A large model comprises an adaptor header, a model main body and an adaptor tail, wherein the adaptor header comprises a first cloud part and a first local part, and the adaptor tail comprises a second cloud part and a second local part; and during the process of training the large model, the first cloud part, the model main body and the second cloud part are deployed on the cloud, and the first local part and the second local part are deployed on a local client. In the method, only the adaptor header and the adaptor tail are distributed to the cloud and the client, encrypted client data is used to perform model training for an adaptor part, the majority of the structure of the model remains on the cloud, and cloud computing power is used to perform training and inference, thereby ensuring the computing power performance; in addition, an adaptor fine-tuning mechanism is effectively used, thereby realizing effective training of the large model while securely and fully utilizing the computing power.

The invention discloses a remote intelligent operation and maintenance system and method for coal mine equipment, and relates to the technical field of fault diagnosis, and the system comprises a local client which is used for obtaining fault sample data of the coal mine equipment and training a local model; the cloud server is used for aggregating the parameters of the local model to obtain a diagnosis model; and the controller is used for acquiring the real-time state data and the real-time operation data in the operation process of the coal mine equipment, and obtaining an early warning result of the fault of the coal mine equipment based on the diagnosis model. According to the method, the diagnosis model is constructed in a federated learning mode, and the diagnosis model is written into the controller of the coal mine equipment, so that fault early warning can be carried out in the operation process of the coal mine equipment, the downtime of the equipment is greatly shortened, and the benefits of coal mine production are improved.

A system, method, and apparatus for estimating life-cycle impact while preserving data privacy uses federated learning to train machine learning models across multiple local clients without exposing private data. Local clients train local models on private datasets. Privatized update data from these local models may be transmitted to a central server which aggregates this data to enhance a global model. The global model is redistributed to clients, thereby improving its accuracy through successive federated learning rounds, without accessing any raw private data. Local clients can query their local models to obtain impact scores for products or processes based on the aggregated learning. A range of privacy-preserving computation protocols limit exposure of sensitive data. This cooperative framework enables collective model refinement reflective of broad data access without compromising confidentiality. The system provides accurate impact values alongside data security assurances for clients.

The invention discloses a large model security reasoning method and system fusing a trusted execution environment and differential privacy, and belongs to the technical field of artificial intelligence. The method comprises the steps that a large model is divided into an embedded layer and a core processing layer, the embedded layer is deployed in a local client SGX2 trusted environment, and the core processing layer is deployed in a local client SGX2 trusted environment; the core processing layer is deployed on a remote server equipped with multiple GPU acceleration cards; based on a local differential privacy technology, performing text confusion on original input data of a user by adopting an output noise maximum value algorithm; performing vector mapping on the confused text data through an embedded function; and the embedded vector is encrypted and transmitted to a server GPU, the embedded vector is processed through a self-attention mechanism and a feedforward neural network, and a prediction result is generated and returned to a client. According to the method, the problem of contradiction between data privacy protection and calculation efficiency in existing large model reasoning can be solved.

The invention provides a method and a system for performing information association review based on RPA and AI. The method comprises the following steps: an RPA client automatically controls a browser to log in an enterprise query website; the RPA client reads the supplier file, inputs supplier information into an enterprise query website for query, and stores basic information of each supplier to the supplier file; the RPA client calls an enterprise association relationship query function of an enterprise query website, and stores a supplier association relationship picture to the local; the RPA client calls the multi-modal model through an interface and identifies relation picture information; the RPA client sends the correctly identified information and the supplier basic information to a large language model, checks whether a preset association rule is triggered or not, and outputs a result; and the RPA client writes the review result into a local file and sends the review result to a designated person for review. According to the invention, the examination efficiency, accuracy and data integration capability of information association examination can be improved.

The invention relates to the technical field of privacy protection, and particularly discloses a privacy protection method and system for data cross-domain sharing and computer equipment, and the method comprises the steps: constructing a global GNN model based on federated learning, distributing the constructed global GNN model to local clients, constructing sub-graphs of the local clients, and distributing the sub-graphs of the local clients to the local clients; configuring a local GNN model of the local client based on the global GNN model; training local client data based on a local sub-graph and a local GNN model to obtain a local model and generate node embedding, uploading the node embedding of the local model to a central server by each local client, performing weighted aggregation operation on the node embedding based on node similarity, and generating global node embedding; according to the method, parameter weights based on a federated learning framework are updated, and the privacy problem of data elements in cross-domain circulation and sharing in a multi-source heterogeneous service fusion scene is solved by introducing a sub-graph, a graph neural network and a federated learning technology.

The invention relates to the field of federated learning and molecular discovery, in particular to a multi-party cooperation federated learning model method for molecular discovery. The invention discloses a multi-party cooperation federation learning model method for molecular discovery. A heterogeneous difference privacy technology is introduced into each local client, and it is ensured that molecular local model parameters are protected in the training process. Next, the protected local model parameter information is uploaded to a central server, projection transformation and weighted averaging are carried out through a Lanczos algorithm, and finally new global model parameters are generated and released to a client for multiple times of communication updating training; according to the method, the privacy protection problem of different molecular data sources is effectively solved by utilizing heterogeneous differential privacy, so that the privacy of the molecular data is guaranteed; and meanwhile, a Lanczos algorithm is adopted to carry out projection transformation to optimize a molecular model fusion mode in a multi-party cooperation process, new global model parameters are generated, and the prediction accuracy of the molecular model can be improved.

The embodiment of the invention discloses a traffic flow prediction and model training method based on federal space-time diagram learning. The specific implementation mode of the method comprises the following steps: generating a node feature matrix of a space-time diagram based on historical traffic flow stored by a local client; determining a space-time dependency relationship based on the node feature matrix, a pre-trained node embedding matrix and a pre-trained time embedding matrix; the time-space dependency relationship is sent to the server, and an aggregation time-space dependency relationship returned by the server is received, and the aggregation time-space dependency relationship is obtained by aggregating the time-space dependency relationships from different clients by the server; and determining future traffic flow through a pre-trained prediction model based on the aggregation space-time dependency relationship. According to the embodiment, the space-time dependency relationship between the clients and inside the clients can be effectively captured, so that the accuracy of traffic flow prediction is improved.

The invention relates to a singular value decomposition-based personalized federated learning method, which is used for solving the problem of uncertainty of a system caused by equipment isomerism and the problem that a client falls behind due to extra burden caused by a storage and transmission mechanism in the prior art. Firstly, a customer selection strategy based on resource awareness is determined, and a server side establishes a quantitative scoring system by comprehensively analyzing computing resources and communication conditions of clients. And screening a plurality of clients with optimal comprehensive performance from all online clients to participate in local training. During local training of the client, historical weight parameters are reserved by adopting self-knowledge distillation (SKDSVD) combined with a singular value decomposition method, residual errors after decomposition are personalized model parameters and are stored in the local client, and universal parameters are transmitted to the server for aggregation. The SKDSVD algorithm reduces the training time delay and the fall-behind rate of the local client while optimizing the global model of the server to obtain the optimal convergence precision.

The invention relates to an intelligent mapping and two-way synchronization system and method for a cloud disk and a local file system, and belongs to the field of IT and software development.The method comprises the steps that the intelligent mapping relation between the cloud disk and the local file system is established, and when a cloud disk mounting request initiated by a user through a local client side is received, the cloud disk mounting request is sent to the local client side; the intelligent mapping module obtains a cloud disk root directory structure by calling an interface of a cloud disk interface module, and converts the cloud disk root directory structure into a local virtual directory tree; the expansion attribute of the local file system is adopted to store the specific attribute of the cloud disk; when it is monitored that the files in the cloud disk or the local file system change, the mapping relation is refreshed; the change of the cloud disk file and the local file system is monitored in real time, and the bidirectional synchronization of the cloud disk file and the local file is realized. According to the system and the method, cloud disk attribute visualization, cross-platform compatibility and access efficiency optimization are realized.

The invention relates to the technical field of risk assessment, and discloses a risk assessment method, a local client, a federated server and a system based on multimodal transport privacy computing, the method is applied to the local client in the risk assessment system, the risk assessment system further comprises the federated server, the local client is connected with the federated server for use, and the federated server is connected with the local client. The method comprises the following steps: acquiring historical local data, and training a local federated topology model based on the historical local data; determining a parameter update quantity based on the trained local federated topology model, and sending the parameter update quantity to a federated server, so that the federated server performs aggregation based on the parameter update quantity to obtain global model parameters; under the condition that the global model parameters sent by the federated server are received, the local federated topology model is updated according to the global model parameters; and performing risk assessment based on the updated local federated topology model. On the premise of protecting data privacy of each participant, collaborative risk assessment of multi-party data is realized.

The present application relates to geothermal energy development and utilization technical field, specifically to a kind of energy underground subway station design method based on federal learning, through parameter collection, local client local model update, calculation result is uploaded to central server, central server carries out update calculation, issue to local client local model, update iteration until global model robustness, the design result of each module is calculated, and heat exchange efficiency and thermal stress are predicted.The energy underground subway station design method proposed in the present application can fully consider the coordination between indoor user end, unit equipment selection and three kinds of energy station structure pipe end in station, and the design scheme obtained can greatly meet the energy-saving and emission-reducing design concept, with low calculation cost, the influence of heat exchange efficiency and heat exchange process on station structure can be accurately evaluated, and compared with traditional test method and field monitoring method, the cost of manpower and material resources can be greatly reduced.

Requests for artifact data for a containerized application cluster are proxied from a first region of a cloud service to a second region of the cloud service, or another cloud service. A request for the artifact data is received at a first container registry at the first region. The first container registry determines that the artifact data is locally unavailable and in response thereto identifies a peer container registry at the second region that has a copy of the artifact data. A request is sent to the peer container registry for the copy of the artifact data. If the request originated from a local client, then the artifact data is forwarded to the local client via an application programming interface, in some examples.

The invention provides a maintenance progress tracking method and device for an electric power facility problem and electronic equipment, and relates to the technical field of electric power. The method is applied to a local client, and comprises the following steps: based on the local client, receiving first multi-modal data of an electric power facility problem sent by a target mobile terminal and corresponding first label information, and according to the first multi-modal data and the first label information, creating a problem maintenance schedule of the electric power facility problem. And after starting to maintain the power facility problem, re-receiving the second multi-modal data and the corresponding second label information sent by the target mobile terminal every preset time interval. Meanwhile, target multi-modal data associated with the second multi-modal data and a corresponding target problem maintenance progress list are determined from the first multi-modal data. And further updating the target problem maintenance progress table based on the second multi-modal data and the second label information. According to the invention, the tracking efficiency and tracking accuracy of the maintenance progress of the electric power facility problem can be improved.

The invention relates to the technical field of federated learning algorithms, in particular to a fair federated learning method based on multi-agent reinforcement learning, which comprises the following steps: respectively initializing network parameters of a server and a local client, and calculating an initial state vector of a local environment by the local client according to the initialized network parameters; the server calculates the initial state of the global environment, and a local client selects and executes an action; the server updates the classification network and broadcasts the classification network to all local clients, and the local clients update local states and rewards; the server updates the global environment state and stores the data in an experience batch; and the server updates the value network, and the local client updates the strategy network. The strategy network is trained based on centralized training and distributed execution, and the convergence speed of the strategy network can be greatly improved.

The invention discloses a remote key security management method and device for a cloud virtualization cryptographic module, and aims to solve the problems of cloud key trust root, transmission security and the like. The core of the method is to keep the master key control right of a user locally, and through cooperative work of a local client, a cloud management platform and a distributed vHSM instance, through the four steps of system initialization and key fragmentation distribution, remote certification and secure channel establishment, distributed cooperative operation and local result synthesis, and by utilizing IBE, Shamir secret sharing, a threshold cryptographic algorithm and TEE technologies, the user master key control right and the distributed vHSM instance are subjected to remote certification and secure channel establishment, distributed cooperative operation and local result synthesis. And the key is available and invisible. The device comprises three modules, namely a local client, a cloud management platform and a vHSM cluster. The method guarantees secret key confidentiality and completeness, supports a national secret algorithm, meets compliance requirements, adapts to cloud native elasticity requirements, and is suitable for cloud secret key management in industries such as electric power and the like.

The application belongs to the field of artificial intelligence and relates to a federated double-regularization multi-error fusion time series anomaly detection method, which comprises the following steps: a central server initializes a global model, obtains global model parameters, and distributes the global model parameters to multiple local clients; the multiple local clients respectively receive the global model parameters and perform local training by using local private data to obtain local model parameters; the local clients respectively perform multi-error fusion detection based on the corresponding local model parameters on local servers to obtain dynamic fusion anomaly scores; the central server receives the local model parameters returned by the multiple local clients, aggregates the multiple local model parameters, and obtains new global model parameters; based on the new global model parameters, multiple new local model parameters are obtained until the global model converges, and optimal global model parameters are obtained; and the optimal global model parameters are distributed to the multiple local clients; and the detection precision and robustness are improved.

Methods, systems, and computer storage media for providing connection management using a connection management engine of a cloud access management system are described. The connection management engine operates based on local client pre-connect connections, and remote client pre-graphics operations and graphics operations to reduce connection times for remote client sessions. In operation, a request to launch a remote client is received, the request associated with a local client. A determination is made whether the local client is associated with a connected state, an express connection state, or a disconnected state. The express connection state is identified from the following: a pre-connect connection state; and cloned network context state. A determination is made that the local client is associated with the express connection state. Based on determining that the local client is associated the express connection state, a remote client session is established between the remote client and the local client.

The present disclosure provides an image processing method and device, electronic equipment and medium, relates to the technical field of computers, and particularly relates to the technical field of electronic maps, cloud computing and cloud services. The specific implementation scheme is as follows: obtaining map update information of a target trajectory point from a server; the map update information comprises a target trajectory point identifier of the target trajectory point; determining a cache state of target index information in a local client according to the map update information; the target index information comprises the target trajectory point identifier; according to the cache state of the target index information, the target index information is obtained, and a target trajectory image of the target trajectory point is determined according to the obtained target index information. The present disclosure realizes the effect of reducing the data amount of data issued by the server under the premise that the target trajectory image can be determined, reduces the pressure of the server, and avoids the occupation of additional bandwidth of the server.

The invention discloses a target detection method based on homomorphic encryption, and the method comprises the steps: generating a pair of encrypted public and private keys at a local client through a homomorphic encryption algorithm, encrypting original image data through the public key to obtain an encrypted image, and transmitting the encrypted image to a server; the server preprocesses the encrypted image, calculates the preprocessed encrypted image data through a pre-trained convolutional neural network model, extracts an image feature value in an encryption domain, locates a target area, and generates an encrypted target detection result; the server returns the encrypted target detection result to the local client; and the local client decrypts the encrypted target detection result returned by the server by using the private key to obtain the final target position and category information. When the convolutional neural network is used for target detection, the data is encrypted during transmission and processing, so that the security of the data is improved, privacy information is protected, and meanwhile, the model performance is not influenced.

The invention provides a communication method and device, an electronic control unit and a readable storage medium, the communication method is applied to a client electronic control unit ECU, and comprises the following steps: when a plurality of processes in the client ECU request a first service on a server ECU, sending the first service to the server ECU; a local client identifier in the service request of each process is converted into a remote client identifier of the client ECU to obtain a plurality of request messages, the request messages comprise session identifier fields, and session identifiers of different request messages are different; establishing a mapping relationship between the local client identifier of each process and the session identifier in the corresponding request message; and sending the plurality of request messages to the server ECU.

This invention provides a network attack detection method and system for vehicle-to-grid (V2G) interaction. It includes: a local client capturing raw traffic data and extracting key features; then receiving the latest global model parameters, detecting the raw traffic data and encoding the attack detection results into metadata tags, while simultaneously using the key features to train the global model parameters to calculate the quantization gradient; extracting feature vectors from the raw traffic data and combining them with the metadata tags to perform attack prediction and obtain the attack prediction probability; the local client executing defensive actions based on reinforcement proxy learning on the attack prediction probability; the local client uploading the quantization gradient and metadata tags to the server via a secure sparse aggregation protocol; the server reconstructing the global gradient to update the global model parameters via a key exchange protocol and updating the attention weights of the global model using the metadata tags; the latest global model outputting the attack probability of V2G interaction, determining whether to issue an alarm and execute defensive actions.

Methods, systems and computer program products are provided for intelligent secure access to private resources. A security service (e.g., SASE ZTNA) may maintain the same or similar security posture for users who work remotely and / or locally by providing authentication, authorization, and / or ongoing conditional access via a security service (e.g., private or public SASE) while intelligently routing remote client traffic to private resources through the security service and routing local client traffic to private resources locally. A traffic routing determination may be made by a security client and / or security server. A traffic routing determination may be based on the location of a client computing device, such as a trusted network detection for a private / trusted network. Traffic routing determinations may be based on conditions alternative or in addition to location, such as the type of private resource or information being accessed by a client computing device.

The invention provides a deployment mode switching method and device, equipment, a medium and a program product, relates to the technical field of computers, and can solve the problem of insufficient flexibility of an overall deployment architecture. According to the technical scheme, when a service application is started, metadata information of all services in a configuration file is analyzed, and all the services are registered to a service directory; receiving a service request of a client through the data operation service; judging a deployment mode identifier of the service application; if the mode is a distributed deployment mode, acquiring an address of a customer transaction service corresponding to the unique identifier of the target service from a service directory through a data operation service, and calling a remote customer transaction service based on the address; if the mode is the single deployment mode, corresponding service metadata information is acquired locally through the data operation service according to the unique identifier of the target service, and the business process of the local customer transaction service is executed based on the service metadata information. And a process for deploying and calling services.