1291 results about "Learning unit" patented technology

The invention discloses a UE4-based digital twin model construction system and method, and relates to the technical field of computer vision, and the system comprises a geometric modeling unit which generates three-dimensional model data based on geometric modeling data of a target physical entity, and achieves dynamic LOD rendering level switching through virtualization processing; the real-time data interaction unit is used for analyzing the real-time sensor data flow and generating equipment state data in combination with a transmission operation and maintenance instruction; the simulation engine unit is used for executing rigid body motion and collision simulation based on the equipment state data and outputting a collision event sequence and state transition data by utilizing an event trigger; the machine learning unit is used for training an LSTM (Long Short Term Memory) model based on the collision event sequence and the state transition data, and optimizing state covariance parameters through adaptive Kalman filtering; and the visual interaction unit is used for realizing data synchronization of the command large screen and the mobile terminal through multi-screen collaboration. According to the method, the digital model capable of accurately reflecting physical entity behaviors and characteristics is constructed.

The embodiment of the invention provides an intelligent learning path recommendation method and system based on a dynamic state. The method is applied to the technical field of intelligent learning recommendation, and comprises the following steps: calculating a skill adaptation weight, a load weight and an emergency degree weight in real time based on a dynamic state of a target employee, including a current skill improvement condition, a workload and a task emergency degree; combining the skill matching degree, the learning strength and the task association degree of each learning unit in the candidate learning path, and performing comprehensive evaluation by using a multi-weight scoring function; and sorting the paths according to the comprehensive score, generating a personalized recommended learning path set and sending the personalized recommended learning path set to an employee terminal, thereby realizing intelligent learning path recommendation with dynamic adaptation and accurate matching. According to the scheme, personalized learning path pushing aiming at actual post requirements and working states of the employees can be realized, correlation, urgency and acceptability of learning contents are improved, learning efficiency and task adaptability are remarkably enhanced, and the employees are helped to quickly compete with post targets.

It aims to appropriately manage the bias of the signal processing amount. 【Solution means】 The abnormality management device 1 uses, as teacher data, normal data indicating the number of normal packets distributed to each device for each time zone, and learns, by maximum likelihood estimation, the parameters of a probability model that outputs the posterior probability that the number of packets distributed to each device for each time zone is normal. A learning unit 11, the posterior probability for each device for each time zone estimated by the learned probability model, the probability distribution of the normal data for each device estimated based on the normal data, and the prior probability of being normal. A derivation unit 12 that derives the probability distribution of abnormal data indicating the number of abnormal packets distributed to each device in the time zone, and the probability distribution of the normal data and the probability distribution of the abnormal data for each device for each time zone are equal. And a setting unit 13 that sets, as a threshold value for abnormality determination of the number of packets distributed to each device, the number of packets corresponding to the value of the probability distribution that becomes the value.

To increase the amount of information to be utilized.SOLUTION: An information processing system includes an information processing apparatus, a first terminal device that holds first user characteristic information, and a second terminal device that holds second user characteristic information. The information processing apparatus includes: an information collection unit which collects, from the first terminal device, first anonymized information generated by anonymizing the first user characteristic information, and collects, from the second terminal device, second anonymized information generated by anonymizing the second user characteristic information; a learning unit which generates a model configured to learn, by machine learning, a relationship between the first user characteristic information and the second user characteristic information, using the first anonymized information and the second anonymized information, as learning data, and output, on receipt of the first user characteristic information, estimated user characteristic information estimated from the relationship between the first user characteristic information and the second user characteristic information; and a model output unit which outputs the model to the first terminal device.SELECTED DRAWING: Figure 3

A heavy traffic abnormal event detection and state discrimination system belongs to the field of intelligent traffic and comprises a data acquisition and labeling module, an image processing module, a traffic state analysis module, a state determination and abnormal event identification module and a transfer learning module. The data acquisition and labeling module comprises a data acquisition unit and a data labeling unit; the image processing module comprises an image enhancement unit and an image compression and reconstruction unit; the traffic state analysis module comprises a traffic bottleneck identification unit and a traffic flow wave analysis unit; the state judgment and abnormal event recognition module comprises a vehicle statistical unit, a congestion state recognition unit, a state transition analysis unit and a compressed sensing model establishment unit; and the transfer learning module comprises a transfer learning unit. According to the method, the false detection rate of the traffic flow abnormal events in bad weather can be reduced, so that the traffic abnormal events can be detected all day long, and the occurrence probability of the traffic events in bad weather conditions can be reduced.

The invention discloses a new energy medium-voltage switch intelligent control method based on multi-data interaction, and particularly relates to the field of switch intelligent control. According to the method, electrical, mechanical and temperature multi-dimensional operation data are collected in real time through a multi-source sensor, fault diagnosis is performed by adopting a hybrid reasoning architecture, features are extracted in combination with Hilbert-Huang transform and Mel cepstrum coefficient algorithms, and a three-level reasoning mechanism is constructed to realize accurate classification; based on a diagnosis result, a system dynamically optimizes a control strategy, the control strategy optimizes a switch-on and switch-off time sequence based on phase-space reconstruction and LSTM prediction, plasma impedance closed-loop adjustment and a quantum genetic algorithm are combined to realize arc suppression and mechanical compensation, and a six-dimensional compensation vector is introduced to correct mechanical operation deviation; the self-learning unit is based on a depth deterministic strategy gradient and an online knowledge distillation technology, verifies the validity of parameter update by using digital twin simulation, and forms a fault feature-control strategy-maintenance scheme triple knowledge base.

It aims to more reliably manage abnormal communication. 【Solution means】 The abnormality management device 1 is configured such that a first learning unit 12 treats each observed value of a normal presence time series as discrete values independent of each other, and estimates probability parameters of a probability model representing the normal presence time after quantization based on the appearance frequency of each observed value; and a second learning unit 13 configured to fix the generator parameters of a generator 131 that generates pseudo-normal data sufficiently deviated from the distribution of true normal data, with each observed value of the normal presence time series being true normal data, and update only the discriminator parameters of a discriminator 132 that discriminates between true normal data and pseudo-normal data in a direction to maximize the discrimination accuracy, based on the probability parameters of the normal presence time estimated by the first learning unit 12.

The invention relates to the technical field of intelligent education, in particular to a history self-learning system based on virtual intelligent tutoring, which comprises a screening classification unit, a recommendation learning module, a learning unit, a deepening module, a feedback module and a secondary deepening module. Wherein the screening and classifying unit is used for classifying learners into a function and benefit oriented type and an emotional resonance type by adopting a learner classifying algorithm based on POS (Point of Sale) annotation and a random forest; the recommendation learning module is used for recommending different historical learning contents to the learner according to the classification result and a self-learning mechanism; the learning unit is used for providing various learning modes for the learner; the deepening module is used for deepening the learning of the learner in an intelligent tutoring mode; the feedback module is used for feeding back a learning condition to the learner and optimizing a self-learning mechanism according to a feedback result; and the secondary deepening module is used for reminding the learner to review and customizing review questions for the learner. Therefore, the problems that in the prior art, courses lack personalized adaptation, learning interaction is not deep, feedback lags and the like are solved.

The purpose is to easily manage abnormal program operation. [Solution] The abnormality management device 1 includes a first learning unit 11 that uses normal data indicating processing resource usage data in which the processing resource usage corresponding to each process ID is normal as training data from among a plurality of processing resource usage data each including the usage of processing resources used in executing a process corresponding to each process ID, and learns parameters of a probability model that outputs a posterior probability that the processing resource usage corresponding to each process ID is normal by maximum likelihood estimation, and a derivation unit 12 that derives a probability distribution of abnormal data indicating processing resource usage data including abnormal processing resource usage based on the posterior probability estimated by the learned probability model, the probability distribution of the normal data, and the prior probability of normality.

Processing circuitry of a learning platform may be configured to maintain a graph database describing student learners. Processing circuitry may obtain new student learner data and load the data into the graph database. Processing circuitry may receive an engagement or interaction from the new student learner and responsively extract new learnings about the new student learner which are loaded into the graph database. Processing circuitry may receive an inquiry from the new student learner and in response, extract the new student learner data and the new learnings from the graph database and contextualize, using a large language model, a learning unit from the educational content provided by the learning platform as a response to the inquiry using the new student learner data and the new learnings. Processing circuitry may further return the learning unit contextualized by the large language model to the new student learner.

The invention discloses a temperature dynamic control system for optical fiber optical wand processing. The temperature dynamic control system comprises an acquisition unit, a control unit, a learning unit and an integration unit. The acquisition unit captures production line material imaging data through a multi-macro camera array, acquires temperature, tension, humidity and speed data in real time by using a production line sensor network, and forms production state comprehensive data after edge computing node fusion preprocessing. The control unit establishes a dynamic mathematical prediction model based on the comprehensive data, predicts future temperature states of a plurality of temperature zones and generates an optimization control input instruction; the learning unit receives an optimization control input instruction and historical temperature data, and generates a self-adaptive control strategy by adopting a reinforcement learning algorithm; the integration unit integrates the control input instruction and the self-adaptive control strategy to form a final control instruction, the temperature of each temperature zone is adjusted, and the temperature control precision and the system stability are ensured.

The invention relates to an automatic image segmentation and recognition system for the radiology department, and relates to the technical field of intelligent medical treatment. Comprising the following steps: an image preprocessing module provides a high-quality image for subsequent analysis by applying a multi-scale transformation and adaptive filtering technology; the feature extraction and fusion unit accurately excavates features by means of a CNN and an attention mechanism; the disease diagnosis reasoning engine is combined with the RNN and a medical knowledge base to deduce disease details; and the personalized treatment scheme recommendation module customizes a scheme by using reinforcement learning according to diagnosis and patient information. In addition, the system also comprises a virtual case generation unit, a model interpretability analysis unit, a self-supervised pre-training unit, a continuous learning unit, a distributed cooperative training architecture and a cross-platform adaptive interface. The method comprises the steps of collecting and standardizing image data, inputting the image data into a system to obtain diagnosis and schemes, and visually displaying auxiliary decision-making. The system efficiently processes radiology department images, improves diagnosis accuracy, optimizes a treatment scheme, adapts to multi-center and multi-platform requirements, and is expected to promote intelligent development of radiology department diagnosis and treatment.

The invention discloses a laser accurate positioning system and method based on an improved PID control algorithm, and relates to the technical field of laser positioning, a holder laser module is constructed and comprises a laser emission unit and a holder execution unit and is used for executing a control strategy, and a multi-sensor array is constructed to collect parameter data of the holder laser module; a control module is constructed and comprises a dynamic bimodal PID unit, a space-time decoupling learning unit and a frequency domain anti-interference unit and is used for analyzing parameter data to generate a control strategy of a holder execution device, a fault processing module is constructed and is used for detecting whether a fault occurs in the process of a laser precise positioning system, and when the fault occurs, the fault processing module is used for processing the fault. Executing a three-stage shutdown process; according to the invention, by improving the PID control algorithm and the frequency domain anti-interference design, the anti-interference capability and the system stability of the laser positioning system in a complex dynamic environment are significantly improved.

The invention relates to the technical field of electric power asset management, and discloses an electric power equipment full life cycle management system based on a digital twinning technology, and the system comprises a causal knowledge graph construction unit, a strategy rule learning unit, a decision option value quantification unit, a combined decision optimization unit, and a maintenance decision voucher management unit. The method comprises the following steps: constructing a knowledge graph containing a causal relationship among equipment, environment and management activities; carrying out anti-fact deduction based on the atlas to dynamically learn strategy rules; quantizing the value of the potential maintenance decision by adopting a physical option model; performing combination optimization under resource constraints of budget, spare parts and the like to generate an optimal decision combination; and finally, generating a standardized maintenance decision voucher for the decision in the combination. According to the method, decision-making flexibility and environment uncertainty can be quantified into specific economic values, future-oriented and globally optimal resource allocation is realized, and a decision-making process and a decision-making result are solidified into traceable and manageable digital assets.

The invention relates to the technical field of sleep environment regulation and control, and discloses a sleep environment self-adjustment system based on multi-source heterogeneous data. An environmental parameter acquisition module and a physiological feature acquisition module of the system respectively acquire sleep space environmental parameters and user physiological feature data in real time; after the master control system receives the two types of original data, a fusion calculation unit cleans original environment parameters to generate a standardized environment data stream, and extracts features from the original physiological feature data to generate a physiological feature time sequence; the dynamic evaluation module divides the two types of data into a plurality of sleep stage data segments according to a preset rule, and calculates an environmental parameter fluctuation index and a physiological feature deviation index; the label distribution mechanism combines the two types of indexes to generate a comprehensive comfort label of each sleep stage; an adjustment decision engine generates a global sleep environment adjustment strategy according to all labels, and an actuator control module drives environment adjustment equipment to execute; and the feedback learning unit receives the adjusted data and updates the calculation benchmark of the dynamic evaluation module.

The invention relates to the technical field of power utilization monitoring, and discloses a photovoltaic user power utilization abnormity monitoring method and system based on artificial intelligence. The photovoltaic user electricity consumption abnormity monitoring system based on artificial intelligence comprises a data acquisition module which is used for acquiring photovoltaic power generation data, electricity consumption data and environment data of a user; the data preprocessing and feature engineering module is used for cleaning, aligning and normalizing the original data acquired by the data acquisition module and constructing a feature data set for model training and reasoning; and the artificial intelligence analysis engine module comprises an unsupervised learning unit, a supervised learning unit and a deep learning unit. According to the invention, the false alarm rate and the missing report rate can be effectively reduced, the accurate diagnosis of the abnormal type can be realized, and the intelligent and accurate operation and maintenance requirements of power grid enterprises on the power utilization monitoring of photovoltaic users are met.

A state signal generation unit that detects the state of a mechanical apparatus as a state signal, a condition signal generation unit that generates a condition signal by detecting the operating status of the mechanical apparatus as an operating condition, a state feature generation unit that generates state features from the state signal, a condition feature generation unit that generates condition features from the condition signal, an initial state learning unit that executes learning based on initial learning state features and outputs initial state learning results, an initial condition learning unit that executes learning based on initial learning condition features and outputs initial condition learning results, an anomaly degree calculation unit that calculates the degree of anomaly based on state learning results and detection state features, and an unknownness degree calculation unit that calculates the degree of unknownness based on condition learning results and detection condition features are included.

An object of a system according to an embodiment is to automatically learn a behavior pattern of a target person, detect an abnormality, and issue an alert.SOLUTION: A system according to an embodiment includes a GPS device, a AI learning unit, an abnormality detection unit, and an alert generation unit. The GPS device keeps track of the current location of the subject. The AI learning unit automatically learns a daily behavior pattern based on the position information of the target person acquired by the GPS device. The abnormality detection unit detects an abnormality by comparing the behavior pattern learned by the AI learning unit with the current position information. The alert issuing unit issues an alert based on the abnormality detected by the abnormality detection unit.SELECTED DRAWING: Figure 1

In a learning device (10), a first evaluation unit (14) performs quantitative evaluation on a plurality of pieces of image data and thereby acquires a plurality of first evaluation results for each of the plurality of pieces of image data, a teacher data generation unit (15) selects a second parameter from among a plurality of first parameters having values different from each other on the basis of the plurality of first evaluation results and generates a first set of teacher data including the selected second parameter, and a first machine learning unit (17) performs machine learning using first sets of teacher data and thereby generates a learned model that outputs a third parameter used for processing of image data of a processing target.

A machine learning device includes: a training data acquisition unit which acquires, as a training data set, data pertaining to a plurality of different filters which are applied to images in which a subject is imaged, and data indicating the state for each predetermined section of a plurality of filtered images which have been processed by the plurality of filters; and a learning unit which uses the training data set to generate a training model which outputs a synthesis parameter for synthesizing the plurality of filtered images for each corresponding section.

The invention discloses a medical image analysis method and system based on a visual language model, and belongs to the technical field of medical image intelligent diagnosis, and the system comprises an image preprocessing unit which carries out the down-sampling of an original retina OCT image to 256 * 256 and carries out the normalization of the original retina OCT image; the feature encoding unit comprises an image encoder based on RET Found in combination with LoRA optimization and a text encoder based on BioClinicalBERT; the class balance comparison learning unit is used for adjusting loss through class balance coefficients so as to relieve the class imbalance problem; the uncertainty estimation unit is used for calculating confidence quality and uncertainty scores based on Dirichlet distribution, and determining a threshold value in combination with an improved Youden index; and the model training unit adopts a total loss function of class balance loss and uncertainty loss, outputs a diagnosis result and an uncertainty score through transfer learning, and further comprises an image input module, a result display module and a data storage module. Rare disease classification performance and reliability are improved, training efficiency is improved through LoRA optimization, and an accurate and reliable scheme is provided for detection of the rare retina diseases.

A method for calibrating a machine-learning unit includes generating an analog calibration signal from an input sequence; generating a digital calibration signal by taking digital samples representing the value of the analog calibration signal at a predetermined sample rate; and applying the analog calibration signal as an input to a physical parallel array analog-to-digital converter (PA ADC) to produce a digital response. The method further includes producing an output by the machine-learning unit, at least in part based on the digital response; modifying a parameter of the machine-learning unit to reduce an error between the output and the digital calibration signal; and determining that the error is not less than a predetermined threshold.

To provide a technique for evaluating residual proof stress in a structure such as a pier.SOLUTION: A machine learning section 12 performs machine learning that uses teacher data using a deterioration degree of a pile and an external force condition to the pile as explanation variables and a damage state of the pile as an objective variable to generate a learned model. A recording section 13 records the learned model generated by the machine learning section 12. An input section 14 inputs the deterioration degree of the pile that becomes an evaluation target and the external force condition to the pile to the learned model recorded in the recording section 13 as input items. A determination section 15 determines use propriety of the pile when external force in accordance with the external force condition is applied, on the basis of the damage state of the pile obtained in accordance with the input to the learned model by the input section 14. An output section 16 outputs damage state data corresponding to the damage state of the pile obtained in accordance with the input to the learned model by the input section 14 and use propriety data corresponding to use propriety of each pile constituting a pier.SELECTED DRAWING: Figure 2

A host vehicle position estimation unit includes an absolute position estimation unit that estimates a first vehicle position based on absolute position information acquired from a GNSS. The host vehicle position estimation unit further includes a relative position estimation unit that estimates a second vehicle position based on relative position information acquired from outside the host vehicle; a traveling state determination unit that determines a change of a traveling state of the host vehicle based on vehicle information or satellite information; a difference computation unit that computes a time synchronized difference between the first and the second vehicle positions; a learning unit that accumulates the difference as time-series data for each traveling state and learns a correction amount of the first vehicle position for each traveling state based on the time-series data; and a position correction unit that corrects the first vehicle position based on the correction amount.

The purpose is to more reliably manage abnormal communications. [Solution] The anomalous communication detection device 1 includes a learning unit 11 that constructs a multivariate probability model using information on the presence state indicating whether the communication terminal 2 was present in the communication area of ​​each base station 3 during each time period, based on a history of presence information regarding the communication area of ​​the base station 3 in which the communication terminal 2 identified by the subscriber identifier was present, estimates parameters of the multivariate probability model, and generates a matrix indicating conditional dependencies between presence states, and a determination unit 12 that determines that an inconsistency exists in the presence information of the communication terminal when the value of a component in the matrix indicating the conditional dependencies between presence states exceeds a predetermined threshold.

A feature value generation device includes a generator configured to generate vectors whose elements are feature values of data items collected at a plurality of timings from a target of anomaly detection, so as to normalize or standardize the vectors based on a set of predetermined vectors; a learning unit configured to learn the predetermined vectors so as to output a learning result; and a detector configured to detect, for each of the vectors normalized or standardized by the generator, an anomaly based on said each of the vectors and the learning result. The set of predetermined vectors is a set of vectors with which no anomaly is detected by the detector, and the set of vectors is updated in accordance with no anomaly being detected by the detector.

In a learning phase, a communication unit (203) acquires communication data that includes a parameter value from which an operation state of a monitored system can be estimated, and that is to be communicated in the monitored system, as learning phase communication data. In the learning phase, a state input unit (204) acquires a learning phase operation state value that indicates a learning phase operation state which is an operation state of the monitored system. In the learning phase, a learning unit (210) performs learning using the learning phase operation state value and a learning phase parameter value included in the learning phase communication data, and generates a learning model (215) for estimating from an attack detection phase parameter value included in attack detection phase communication data which is communication data that is to be communicated in the monitored system in the attack detection phase, an attack detection phase operation state which is an operation state of the monitored system.

The invention provides a deep-sea mining system, and relates to the technical field of deep-sea mining. The deep-sea mining system comprises a decision-making layer, the decision-making layer is connected with a coordination layer through a quantum encryption underwater acoustic communication module, the coordination layer is connected with an execution layer through a bionic tentacle electromechanical interface, and the execution layer is connected with a compensation layer through an ecological data bus; the decision-making layer comprises a phase resource management module which comprises a marine rhythm analysis unit, an ore deposit respiration model engine and a mining window decision-making device; the biological contract agreement module comprises a microorganism symbiotic culture cabin, a sound wave induction generator and a tentacle mechanical learning unit; the coordination layer comprises a dynamic topology network which comprises a cluster self-organizing controller, a real-time damage evaluation unit and a topology reconstruction algorithm library. Through collaborative design of the decision-making layer, the coordination layer, the execution layer and the compensation layer, intelligent and efficient deep-sea mining is achieved, and the mining efficiency and the resource utilization rate are remarkably improved.

The invention provides a self-driven humidity detection system based on artificial intelligence. The self-driven humidity detection system comprises a humidity sensing converter module, an LED bulb, a self-driven module and a background processing module. The humidity sensing converter module comprises a humidity sensing module and an indication module; the self-driving module comprises a water storage device, a TENG, an instillation device and an integrated circuit; and the background processing module comprises a visual monitoring acquisition unit, a data analysis unit, a machine learning unit and a display interface unit. The system can continuously detect the humidity change in the environment, accurately measure the real-time humidity and realize the continuous visualization of the humidity signal. The humidity detection function is based on a pointer suite which deforms along with the change of environment humidity, a pointer of the indication module deforms, and an LED bulb fixed on the pointer of the indication module shifts along with the pointer. In the working state, water drops collected by the water storage device drop on the TENG through the instillation device, friction current is generated through impact, the LED emits light through a connecting circuit of the TENG and the LED, the position and relative displacement of the bulb are recognized and analyzed through the background processing module, and the relation between the position and the environment humidity of the bulb which are input in advance is compared. Therefore, the humidity signal is visualized and the environment humidity is accurately measured. Remote visualization of humidity detection is realized based on image identification and processing of artificial intelligence, sustainability and accuracy of humidity detection are guaranteed, and efficiency and convenience of humidity detection are improved.

The invention discloses a dynamic resource scheduling optimization method based on LSTM (Long Short Term Memory) and Q-learning. The dynamic resource scheduling optimization method comprises the following steps: step 1, acquiring historical load data X of equipment; 2, inputting the historical load data X into the LSTM load prediction model, and optimizing the LSTM load prediction model to obtain an optimal LSTM load prediction model; 3, predicting predicted load data Xw at a certain time point or within a certain time period in the future through the optimal LSTM load prediction model; 4, inputting the predicted load data Xw into a resource scheduling optimization module; and 5, calculating a Q value of a Q-learning unit of the resource scheduling optimization module, and updating the Q value. The dynamic resource scheduling optimization method based on the LSTM and the Q-learning has the advantages that the computing resources can be effectively optimized, the use efficiency of the computing resources is improved, the speed of the computing resources is increased, the use cost of the computing resources is reduced, and the like.