18 results about "Online machine learning" patented technology

The invention relates to an electric power marketing business abnormity real-time detection method and system based on stream-oriented computation, and belongs to the technical field of electric power system optimizing.The method comprises the steps that data snapshots are extracted from an electric power marketing business system, difference comparison is conducted on the data snapshots and historical snapshots of an intermediate library, and standardized increment events are generated and stored; capturing an incremental event in real time through a data change capturing tool and pushing the incremental event to a message queue; a streaming computation engine consumes the event stream, sequentially performs data cleaning, association with a static dimension table and sliding window statistical feature calculation, and constructs a feature vector; and performing parallel analysis and weighted fusion on the feature vectors based on a business rule base and an online machine learning model to generate a comprehensive risk score, and outputting an abnormal event when the score exceeds a threshold value. According to the method, the problems of exception identification lagging and complex work order process in a traditional batch processing mode are solved, the crossing of the business risk from hour-level detection to minute-level real-time perception is realized, and the timeliness and accuracy of power marketing risk management and control are improved.

In some implementations, the EMV uses a calibration to inform autonomous control over the EMV. To calibrate an EMV, the system first selects a calibration action comprising a control signal for actuating a control surface of the EMV. Then, using a calibration model comprising a machine learning model trained based on one or more previous calibration actions taken by the EMV, the system predicts a response of the control surface to the control signal of the calibration action. After the EMV executes the control signal to perform the calibration action, the EMV system monitors the actual response of the control signal and uses that to update the calibration model based on a comparison between the predicted and monitored states of the control surface.

There is a need for solutions that classification solutions in hierarchical prediction domains. This need can be addressed by, for example, performing one or more online machine learning, co-occurrence analysis machine learning, structured fusion machine learning, and unstructured fusion machine learning. In one example, structured predictions inputs are processed in accordance with an online machine learning analysis to generate structurally hierarchical predictions and in accordance with a co-occurrence analysis machine learning analysis to generate structurally non-hierarchical predictions. Then, the structurally hierarchical predictions and the structurally non-hierarchical predictions in accordance with processed by a structured fusion model to generate structure-based predictions. Afterward, the structure-based predictions and non-structure-based predictions are processed in accordance with an unstructured fusion model to generate one or more unstructured-fused predictions.

The invention discloses an advertisement putting effect real-time optimization method and system based on machine learning, and particularly relates to the technical field of internet advertisements. Comprising a multi-source real-time data acquisition module, a dynamic feature processing module, a multi-dimensional effect evaluation module, an online machine learning model module, a real-time delivery decision and execution module and a visualization and operation management module. According to the invention, through cooperative work of the multi-task online learning model and the reinforcement learning decision model, advertisement putting parameters are optimized in real time; a multi-dimensional evaluation system containing short-term and long-term indexes is adopted, the index association relation is analyzed, and the short-term effect and the long-term value are balanced; the feature calculation frequency is dynamically adjusted by using the self-adaptive time window feature, and the feature accuracy is improved; and continuous optimization of the putting strategy is realized through closed-loop feedback and a strategy iteration mechanism. According to the invention, real-time optimization of the advertisement putting effect is realized, and the rate of return on investment of advertisement putting is obviously improved.

The invention provides an all-digital networked fault-tolerant calculation control method and system, and relates to the technical field of power station operation control, and the method comprises the steps: independently collecting reactor data in real time through a main control unit and a plurality of standby control units, and carrying out the synchronization of corresponding output data messages through a high-speed redundant communication bus; verifying the data message, performing automatic correction when a correctable error is detected, and immediately generating and reporting a fault positioning code when an uncorrectable error is detected; identifying a fault component through the fault positioning code, taking the function daughter board as a minimum granularity isolation fault source, and starting a recovery mechanism; operation parameters of each control unit and a field sensor are collected in real time through a safety data bus, and an online machine learning model is utilized to predict potential faults and take prevention measures. According to the invention, the technical problems of low reliability, safety and intelligence of the power station control system are solved.

A distributed, online machine learning system is presented. Contemplated systems include many private data servers, each having local private data. Researchers can request that relevant private data servers train implementations of machine learning algorithms on their local private data without requiring de-identification of the private data or without exposing the private data to unauthorized computing systems. The private data servers also generate synthetic or proxy data according to the data distributions of the actual data. The servers then use the proxy data to train proxy models. When the proxy models are sufficiently similar to the trained actual models, the proxy data, proxy model parameters, or other learned knowledge can be transmitted to one or more non-private computing devices. The learned knowledge from many private data servers can then be aggregated into one or more trained global models without exposing private data.

A big data job planner uses online machine learning to recommend algorithms for each job in a sequence to enhance various aspects of job execution. A job algorithm prediction server processes a job's pre-runtime metadata to create predictions of relevant big data job behavior characteristics based on previously learned models of the corresponding big data system's behavior and provides a recommended algorithm to the client big data system. After the big data job is executed, the job metadata coordinator persists pre-runtime metadata and corresponding post-runtime big data job behavioral metadata actually observed. Repeatedly through time, online machine learning is performed on this sequence of job behavior metadata so that machine learning models will learn difficult-to- recognize phenomena, as well as, changes in system behavior through time. Unlike statically coded rules, the system models changing patterns and behaviors, enabling the algorithm predictions to remain useful, thereby enhancing big data system quality.

The invention relates to the technical field of big data stream processing, in particular to a big data stream processing method and system based on machine learning, and the method comprises the steps: collecting the real-time feature data of each state in a stream processing task, the real-time feature data comprising the state access frequency, access time interval, state size and survival age; predicting the access probability of each state in a future time window through an online machine learning model; distributing an importance factor for each state, and reflecting the key degree of the state in correct operation and rapid recovery of the system through the importance factor; calculating a comprehensive priority score of each state based on the access probability and the importance factor; and performing state management of the data flow and hot update of the online machine learning model according to the comprehensive priority score.

The application discloses a privacy protection method and system based on a centerless flow federated learning and a terminal, wherein online learning is performed on local real-time data flow by an edge node, then a communication interaction opportunity between nodes is adaptively determined based on changes of local model parameters, and a privacy protection based on a Laplace mechanism is performed on the model parameters during the communication interaction, and then the model parameters are broadcasted and shared with adjacent nodes, and no parameter transmission is performed during a non-communication interaction time, so as to reduce communication overhead and a privacy budget. Finally, dynamic model training and updating of global data flow are cooperatively performed by the edge node under the premise of privacy protection. The application has good application effects in a privacy protection scene of actual large-scale distributed node cooperative online machine learning, and can be applied to data privacy protection scenes in application scenes such as vehicle networking driving intelligence, mobile socialization and online recommendation.

Various aspects of the present disclosure relate to an apparatus for wireless communication. The apparatus may be configured to, capable of, or operable to: receive a request for a trained machine learning model; select at least one further apparatus to provide feedback on a machine learning model based on data available for at least one feature of the machine learning model to the at least one further apparatus; request feedback for the at least one feature of the machine learning model from the selected at least one further apparatus, wherein the request includes a data indication; receive feedback information that includes at least one predicted label of at least one feature; and update at least one trained label of at least one feature of the machine learning model based on the received feedback information.

The invention discloses a vehicle cooperative anti-skid control method and device. The method comprises the steps of obtaining real-time operation data of a vehicle and sending the real-time operation data to a cloud platform; on the cloud platform, key dynamic parameters of the cloud digital twin model are calibrated in real time based on the real-time operation data by using an online machine learning model, and real-time calibration parameters are obtained; issuing the real-time calibration parameters to a vehicle end controller of the vehicle; in the vehicle-end controller, the real-time calibration parameters are utilized to update the vehicle-end digital twinborn model; and based on the updated vehicle end digital twinborn model, adopting a model prediction control algorithm to predict the driving state of the vehicle in a future prediction time domain, solving to obtain an optimal control instruction, and executing the optimal control instruction. According to the method, traditional passive anti-skid control is converted into active predictive control, dynamic changes of key parameters such as road adhesion coefficients can be adapted in real time, and the driving safety and stability of the vehicle under extreme road conditions such as ice and snow are remarkably improved.

The invention discloses a direct-driven wind session hyper-synchronous oscillation rapid online detection and timing method, electronic equipment and a storage medium, and relates to the field of power systems and automation thereof. The method comprises the following steps: acquiring operation data of a power system, generating a sliding window sample through sliding window movement, and extracting time-frequency characteristics of the sliding window sample by adopting a time-frequency analysis method; inputting the time-frequency characteristics into a pre-trained oscillation detection model, and judging whether the sliding window sample oscillates or not; if the oscillation detection model judges that oscillation occurs, inputting the time-frequency characteristics into a pre-trained oscillation timing model, and determining oscillation occurrence time in the sliding window sample; and outputting the oscillation judgment result of the oscillation detection model and the oscillation occurrence time of the oscillation timing model. According to the method, whether oscillation occurs or not can be rapidly and accurately judged, the specific oscillation occurrence time can be obtained, and judgment conditions are provided for machine learning oscillation identification algorithm and subsequent suppression measure investment in the online process.

The application provides a two-section nonlinear radiation surface space atmospheric density sensor heating cover, the top end of which is provided with an opening for the entry and exit of space atmospheric molecules; the radiation surface is divided into two sections, and the generatrix of each section of the radiation surface is a curve, the shape of which follows different nonlinear functions; the bottom end of the heating cover is provided with a mounting seat for fixing with a spacecraft body. The generatrix of each section of the radiation surface of the heating cover is designed according to a self-guided online machine learning method, the undetermined coefficients of the function analytical expressions followed by the generatrix of each section of the radiation surface are selected as factor variables, and the heating time is taken as a target variable, and the best variable design value is obtained by combining a heuristic algorithm with a deep neural network. The two-section nonlinear radiation surface space atmospheric density sensor heating cover provided by the application can enhance the radiation heat transfer characteristics between the sensor front end quartz crystal balance and the sensor, reduce the heating time required for the desorption of space atmospheric molecules, and thus reduce the power consumption of the spacecraft in orbit flight.

The application provides a two-section linear radiation surface space atmospheric density sensor heating cover, the top middle part of which is provided with a circular opening for the entry and exit of space atmospheric molecules; the inner radiation surface is divided into two sections, the generatrix of each section of the radiation surface is a straight line with a certain slope, and the slope of the generatrix of the first section of the radiation surface is smaller than that of the second section; the bottom of the heating cover is provided with a mounting seat in the circumferential direction for being fixed with a spacecraft body. The generatrix of each section of the radiation surface of the heating cover is designed according to a self-guided online machine learning method, the horizontal span and height of each section of the generatrix are selected as factor variables, and the heating time is selected as a target variable, and a heuristic algorithm is combined with a deep neural network to obtain the optimal variable design value. The two-section linear radiation surface space atmospheric density sensor heating cover provided by the application can enhance the radiation heat transfer characteristics between the sensor and the front quartz crystal vibration balance, reduce the heating time required for the desorption of space atmospheric molecules, and thus reduce the power consumption of the spacecraft in orbit flight.

The application discloses a method for detecting clogging of a stopper in a continuous casting process, comprising the following steps: S1, in the early stage of pouring, online machine learning modeling is performed on real-time high-frequency data to obtain a decision coefficient R 2 value for evaluating the accuracy of a model; S2, every m window, the performance of the evaluation model is evaluated by using the high-frequency data in the window to obtain the R 2 value of the evaluation model in the m window; S3, the R 2 value of the window is compared with the decision coefficient R 2 value to perform early warning or alarm; and S4, the steps S2 and S3 are repeated until the pouring of the current pouring is finished. The application can perform early warning of clogging of the stopper in the continuous casting process, remind an operator to remove the clogging or replace the stopper, and ensure the sequence in the production process and the quality of the continuous casting slab.

Method for automatically placing data in a storage system (40) having different types of storage media according to at least one data category, the method comprising at least the step of classifying the data according to their category, the classification being performed by an online machine learning technique using information relating to said data.

A safety interlock recommendation system includes at least one process data source, an edge device, wherein the process data source is configured for providing IOS device stream data to the edge device; wherein the edge device comprises an operational technology edge application unit, OT edge application unit, and a stream analysis unit; wherein the OT edge application unit is configured for providing operation technology stream data, OT stream data; wherein the stream analysis unit comprises an online machine learning model, being configured for determining online analysis data using the provided process stream data and the provided OT stream data; wherein the OT edge application unit is configured for determining a short-term recommendation using the online analysis data.

A dynamic system controller includes a first error detection unit that determines a first error between a joint state and a joint state target. The joint state describes an actual state of a set of joints in a dynamic body. The set of joints has one or more degrees of freedom that are controlled by a set of actuators. An impedance controller generates an impedance torque or force based on the first error. A first neural network accepts the joint state target and outputs an estimated actuation force. A second error detection unit determines a second error between an actual actuation force applied to the set of joints by the set of actuators at the joint state and a combination of the impedance torque or force and the estimated actuation force. A force controller generates a set of actuation commands based on the second error.