94results about How to "Self-directed learning" patented technology

The invention discloses a retinal vessel segmentation method based on combination of deep learning and a traditional method and relates to the fields of computer vision and mode recognition. According to the method, two grayscale images are both used as training samples of a network, corresponding data amplification, including elastic deformation, smooth filtering, etc., is done against the problem of less retinal image data, and wide applicability of the method is improved. According to the method, an FCN-HNED retinal vessel segmentation deep network is constructed, an autonomous learning process is realized to a great extent through the network, convolutional features of a whole image can be shared, feature redundancy can be reduced, the category of multiple pixels can be recovered from the abstract features, a CLAHE graph and a gauss matched filtering graph of the retinal vessel image are input into the network, an obtained vessel segmentation graph is subjected to weighted average, and therefore a better and more intact retinal vessel segmentation probability graph is obtained. Through the processing mode, the robustness and accuracy of vessel segmentation are improved to a great extent.

The invention discloses a virtual reality-based foreign language learning method and system. The system comprises a virtual reality terminal, a terminal adapter, a computer and a virtual reality database. The method comprises the steps of acquiring an action of a user through an input device in the virtual reality terminal; transmitting the action to the computer through the terminal adapter; processing and calculating the acquired data by the computer; performing retrieval, matching, modeling, mapping and calculation according to various foreign language learning scene models in the virtual reality database to generate corresponding virtual reality output results; and displaying virtual reality scenes in an output device of the virtual reality terminal. The user also can perform foreign language three-dimensional associative learning by utilizing 3D scenes. According to the virtual reality-based foreign language learning method and system, the computer is used for performing three-dimensional animation making, so that simulated scenes similar to real foreign work and life are generated; and the user can interact with a virtual environment in senses such as a visual sense, a hearing sense, a touch sense and the like, so that visual and intuitionistic learning feelings are formed.

The invention relates to the technical field of big data, in particular to a question and answer library knowledge management system based on deep learning and an implementation method thereof. The system comprises a global setting module, a knowledge management module, a knowledge learning module, a knowledge statistics module, a session simulation module and a terminal docking module, each module has setting functions of adding, deleting, modifying, checking and the like; the method comprises the following basic steps: (1) carrying out global setting; (2) performing knowledge management, including knowledge creation; (3)carrying out knowledge learning, wherein knowledge mining and problem clustering are carried out through an iteration process of big data information / historical sessions,and user annotation feedback is carried out; (4) carrying out session simulation: simulating a session of a real environment; (5) carrying out knowledge statistics, and high frequency and sudden increase problems in the application field are understood for improvement and decision making; and (6) carrying out multi-terminal adaptation and distribution on the packages by the modules. According tothe system and the method, the natural language habit of human beings can be known, the complex problem of a user can be known, and manual intervention is not needed.

The invention belongs to the technical field of education data mining, and discloses a multi-step hierarchical learner cognitive level mining method and system. The method includes: constructing a test question-knowledge cognition level matrix P by combining target classification and cognition psychology in the Bloom cognition field, constructing a learner ideal answer matrix by integrating the learner knowledge cognition level matrix and the test question-knowledge cognition level matrix, and mining a learner knowledge mastering candidate set by utilizing maximum likelihood estimation; and comprehensively judging the global expectation of the elements in the candidate set to obtain the final knowledge cognition level of the learner, and visually outputting the result by using a radar map.According to the invention, after the knowledge cognition level of the learner is mined, the mining result fed back to the learner by using the radar map is more intuitive and popular and easy to understand, and the learner is assisted to adjust the learning scheme in time; the hidden parameters of the test questions mined by the model can evaluate the quality of the test questions forming the test, and the test accuracy and credibility are improved.

The application discloses a message processing method and message processing equipment, which are used for realizing autonomic learning of IP (Internet Protocol) addresses by a matching condition and are suitable for a plurality of scenes taking message features as a static parameter and a dynamic parameter. According to the embodiment of the invention, the method comprises the following steps: acquiring an IP message, and judging whether or not the acquired IP message is a domain name system (DNS) response message; if the IP message is the DNS response message, resolving a domain name requested by a user terminal from the IP message; judging whether or not the domain name requested by the user terminal exists in the matching condition of a control strategy, wherein the matching condition comprises a domain name needing to perform message control according to the strategy action of the control strategy; and if the domain name requested by the user terminal exists in the matching condition, adding an IP address corresponding to the domain name requested by the user terminal in the IP message into the matching condition in order that the matching condition further comprises the added IP address.

The invention discloses a multi-unmanned aerial vehicle 3D hovering position joint optimization method and device and an unmanned aerial vehicle base station. The method comprises the following steps:firstly, acquiring state information of a heterogeneous network where unmanned aerial vehicles are located; inputting the state information to a deep reinforcement learning network constructed in advance, determining hovering positions at next moment through a current strategy, and obtaining reward function values of the hovering positions of the unmanned aerial vehicles at the current moment from the environment; determining a gradient strategy algorithm based on different strategy depths, calculating an updated gradient, and synchronously updating strategy parameters of the plurality of unmanned aerial vehicles; according to the updated gradient, iteratively executing the steps from obtaining state information to synchronously updating the strategy parameters of the plurality of unmanned aerial vehicles, so that an objective strategy function is gradually converged until an optimal strategy is obtained. The device comprises a state acquisition unit, a reward unit, a gradient updating unit and a training unit. The method can be executed by a processor of the unmanned aerial vehicle base station. According to the method, the device and the unmanned aerial vehicle base station in the invention, autonomous learning of the multi-unmanned aerial vehicle in the environment can be realized, and dynamic and non-stable environment changes can be adapted to.

The invention discloses an adaptive particle swarm algorithm-based grayscale threshold obtaining method and an image segmentation method, and belongs to the technical field of image processing. The grayscale threshold obtaining method is characterized by comprising the following steps of S01, performing population initialization on a grayscale value of an image; S02, calculating a fitness value of an individual in a population; S03, calculating an optimal position and a global optimal position of the individual in the population; S04, updating the optimal position and the global optimal position of the individual in the population; and S05, judging whether a stop condition is met or not, and if the stop condition is met, obtaining an optimal solution and obtaining an optimal grayscale threshold, otherwise, executing the step S02 to enter a next-generation population, wherein the optimal position and the global optimal position of the individual are dynamically adjusted by adopting an inertial weight in the step S04. The grayscale threshold obtaining method has autonomic learning property, adaptivity and relatively high robustness, can concurrently solve the grayscale threshold globally and better avoid local optimum, and is accurate and efficient.

The invention discloses an industrial production process target data prediction method of a multi-feature fusion deep neural network. Collecting other variable time sequence data related to the key variable in the industrial equipment through equal-interval sampling by utilizing a sensor, and performing predictive analysis on the time sequence data of the key variable in the process industry; inputting into a pre-designed and pre-constructed deep convolutional neural network for training; segmenting historical data of the key variables according to time steps and then inputting the segmented historical data into a deep gated recurrent neural network for learning; a multi-feature fusion method is utilized to fuse output features obtained by two networks and then input the output features into a full connection layer, network parameters are optimized through back propagation, and prediction precision is improved. According to the method, reliable and effective target variable parameter prediction is provided for process monitoring in industrial production, and the hysteresis quality of measurement of key variables such as the silicon content of molten iron in industrial production isrelieved.

The embodiment of the invention provides scene error correction method, device and equipment, and a storage medium. The scene error correction method comprises the following steps: performing semanticunderstanding on a training sample by utilizing scene knowledge of a target scene to obtain semantic features; mining associated knowledge of the scene knowledge; and training by utilizing the semantic features, the scene knowledge and the associated knowledge to obtain a scene error correction model, wherein the scene error correction model is used for performing scene error correction on an input text. According to the embodiment of the invention, the scene knowledge is fully utilized to carry out semantic understanding on the training sample, so that finer semantic features can be obtained, and the error correction model effect is improved; and moreover, scene knowledge is fully utilized, and associated knowledge is mined, and an error correction model is assisted to make a decision.

The invention discloses a learning data interaction method and a system based on group discussion in an SPOC platform. The method and the system are characterized in that the system comprises a server, a teacher terminal and multiple student terminals; and the method comprises steps that 1) the server searches and stores user characteristic data into a database; 2) the teacher terminal carries out grouping according to a preset analysis method and selects a group leader of each group according to a preset selection method; 3) group leader terminals monitor learning time and the corresponding student terminals carry out data interaction so as to form report data and then send the report data to the teacher terminal; 4) the teacher terminal carries out feedback; and 5) the server forms scores of all group members. According to the invention, autonomous learning of students is realized; intelligent management is realized for teachers; and a novel teaching and learning mode is developed.

The invention provides an intelligent regulation and control system and method for mine ventilation equipment based on digital twinning. The system comprises a digital twinning system physical entity, a virtual entity, twinning data and a ventilation equipment intelligent regulation and control system. The digital twinning system physical entity comprises a mine fully-mechanized excavation scene, an excavation working face, ventilation system entity facilities and various sensors; the virtual entity of the digital twinning system refers to a virtual modeling environment of the physical entity, truly maps data information of the physical entity and provides a virtual debugging environment for the ventilation equipment intelligent regulation and control system; the intelligent regulation and control system runs on the virtual entity and acts on the physical entity, and is a key module for realizing real-time interaction between the virtual entity and the physical entity; a ventilation system intelligent regulation and control system is responsible for processing the twinning data, generating a decision scheme through data information and providing intelligent decision information for ventilation system equipment.

The invention provides a super-resolution deblurring method of a generated antagonistic network, which comprises the following steps: S1, utilizing a DRCN network structure to form a super-resolutiondepth convolution network, and establishing an antagonistic network model; S2, combining the SRGAN network cost function to improve the countermeasure network performance; S3, selecting clear picture,adding Gaussian noise and motion blur to realize training. The present invention analyzes the characteristics of motion blur, Designing the artificial noise of samples, adding defocus fuzzy kernel and multi-direction motion fuzzy kernel, realizing the super-resolution deblurring of the blurred image with double magnification, and analyzing the blurred image taken by UAV, can greatly reduce the influence of professional knowledge and experience on the design of deblurring algorithm.

The invention belongs to the technical field of energy management, and discloses a railway station building construction equipment monitoring and energy management system, a program and a storage medium. The railway station building construction equipment monitoring and energy management system comprises a construction equipment monitoring and energy management system host, a central air conditioner energy-saving control subsystem, a fresh air system, an intelligent illumination control subsystem, an electromechanical equipment acquisition control subsystem, an environmental data acquisition subsystem, an energy data acquisition subsystem, a regional equipment control cabinet, various sensors and acquisition devices. The construction equipment monitoring and energy management system host comprises a central air conditioner acquisition control module, a fresh air system acquisition control module, an intelligent illumination acquisition control module, an electromechanical equipment acquisition control module, an environment data acquisition module, an energy data acquisition module, a passenger station brain data interface module and other system data modules. The system has the functions of autonomous learning and promotion, and the innovative technical level and economic indexes reach the domestic leading level.

A medical diagnosis management education system comprises: a database device for storing one or more simulated patient data; an operation interface device for providing the simulated patient data via an image display module and receiving a question indication message associated with the simulated patient data; a data processing device for obtaining, from the database device, standard diagnostic information of an expert associated with the simulated patient data to be displayed on the operation interface device; wherein the standard diagnostic information has one or more category blocks; and when the data processing device receives pairing information from the operation interface device, the pairing information includes pairing information of the question or instruction information and the one or more category blocks, and the data processing device generates a comment according to the pairing information.

According to a multi-agent electric vehicle battery replacement scheduling method based on the Internet of Vehicles, a vehicle-road cooperative service is deployed on an MEC platform, and man-vehicle-road cooperative interaction is realized by means of a Uu interface or a PC5 interface and communication modes such as VANET, 4G or 5G; according to a map around an electric vehicle with a battery replacement service requirement, a roadside unit divides a battery replacement station cluster with a high potential cooperation matching degree into a whole and gathers the battery replacement station cluster into a battery replacement area, and the battery replacement area with the maximum service capability probability shares a plurality of electric vehicles with the battery replacement service requirement at the same time; the service rate of each battery swap station is taken as an assessment target to mainly assess the self-service capability, self-service quality and sitting information ofeach battery swap station node and the current self-state of the electric vehicle with the battery swap requirement; and the optimal joint action of global electric vehicles is provided, so that theoverall service balance of each battery swap station is maintained, and the long-term performance of the Internet of Vehicles is improved. According to the invention, the electric vehicle can exchangepower as soon as possible, and each power exchange station can maintain service balance.

The related assist learning device for student experiment comprises: a MCU connected to the output interface of the experiment device and the experiment network linked with teacher terminal, a LCD, an input device, a mobile memory card interface, an external card, the USB interface on MCU bus, a network communication chip and interface, an IIS interface chip, and a standard serial interface. This invention has lots of function for student experiment and instructor teaching.

The invention provides an intelligent aircraft collision avoidance method based on reinforcement learning. The intelligent aircraft collision avoidance method comprises the steps of querying a state,receiving data, calculating a position and receiving a result. The invention relates to electronic equipment and a readable storage medium to execute an intelligent aircraft collision avoidance methodbased on reinforcement learning. According to the method, static and dynamic data of airways and aircrafts are obtained; spatial positions of the aircrafts at n time points in the future are calculated; if the distance between the aircrafts reaches a danger threshold value, a collision avoidance strategy is learned by using a reinforcement learning method; an aircraft flight adjusting instructionis outputted; the aircrafts perform the instructions and update the headings; and an effect evaluation value is calculated and returned. The reinforcement learning algorithm is updated through the effect evaluation value and thus the algorithm is continuously optimized in continuous iterative learning, so that autonomous learning and improvement of the intelligent agent are realized, the dangerous situation of collision is autonomously avoided, and a problem of collision between aircrafts when multiple aircrafts fly at the same time is solved.

The invention discloses an auxiliary network teaching system. The auxiliary network teaching system comprises an online teaching center, a learning terminal and an education terminal, wherein the learning terminal and the education terminal are connected with the online teaching center through the internet, and the learning terminal at least is used for a student to study through the online teaching center; the education terminal is used for a teacher to carry out course preannouncement and synchronous teaching through the online teaching center. With the auxiliary teaching system provided bythe invention, the student can realize independent study, meanwhile, by establishing the linking between the learning terminal and the education terminal, while teaching video synchronous living broadcast is carried out by the education terminal, the interaction between the teacher and the student is realized, further, the purpose of synchronous learning is realized, and the learning efficiency isimproved.

The embodiment of the invention discloses a voice data processing method and device based on a dialogue scene, a storage medium and electronic equipment, and belongs to the field of online education.The method comprises the steps of setting a dialogue scene; collecting voice data input by a user in a dialogue scene, and analyzing dialogue content of the voice data; under the condition that the dialogue content is not matched with the dialogue scene, displaying first prompt information, wherein the first prompt information indicates that the dialogue content of the voice data is not matched with the dialogue scene. According to the invention, unsupervised autonomous learning can be realized, the labor cost is reduced, and the learning efficiency is improved.

The invention discloses a control system based on an intelligent teaching robot, which comprises an input module, a classification module, a data display module, a data processing module, a controller, a reward control module, a reward box, a voice module, a database, a data calculation module, a data analysis module, a cache module, an induction module, a Bluetooth interconnection module and a prompt module, wherein the control system comprises an input module, a classification module, a data display module, a data processing module, a controller, a reward control module, a reward box, a voice module, a database, a data calculation module, a data analysis module, a cache module, a Bluetooth interconnection module and a prompt module. The invention adopts a data analysis module to analyzeeach test in the acquired period of time, the correct rate Hij of each item type is analyzed, and Ti and Li are obtained after calculation, According to Ti and Li, each numerical value size carries onthe independent determination to the skilled question type, After that, the proficient question types identified independently will be transmitted to the classification module, so that the classification module will automatically ignore the consistent proficient question types in the future classification process, so as to achieve autonomous learning in the continuous teaching process, in order to improve the learning efficiency of children.

The invention relates to the technical field of intelligent teaching, in particular to a shape-up exercise method, a shape-up exercise device, a shape-up exercise system, computer equipment and a storage medium. The shape-up exercise method comprises the following steps of obtaining a shape-up exercise action image, wherein the shape-up exercise action image is a human body whole body image subjected to optical sensing; calling a standard shape-up exercise image corresponding to the shape-up exercise action image; performing comparison on the shape-up exercise action image with the standard shape-up exercise action image; outputting comparison information in real time. The shape-up exercise action is collected by using an infrared sensing technology; the shape-up exercise action is subjected to analysis and comparison; the avoidance of influence due to light irradiation, background, garment and the like during the shape-up exercise can be ensured; the shape-up exercise action correctioprecision is improved; the independent study of a user can also be realized; the teaching by a teacher is not needed; the standard shape-up exercise action can be easily learned.

The invention discloses a biologically-inspired image meaning information autonomous extraction method and device. The method comprises the following steps of: using an image data set with labels as training samples for training a convolution neural network; clustering weight parameters of a training network, and carrying out convergence on the weight parameters according to the clustering result; and using the weight parameters after the convergence as new parameters of the convolution neural network, using the new network to extract meaning information characteristics of images, and carrying out identification and classification on the images according to the characteristics. According to the invention, the networked expression method of meanings is given, the autonomous learning and extraction of the meaning information are realized by the utilization of the network structure, the characteristic dimensions are substantially reduced without influencing the model effect, and the explanation capability of the model is simultaneously improved.

The invention discloses a vehicle distributed steering driving system control method based on reinforcement learning. The method comprises the steps of obtaining current state data of a vehicle; estimating the current side slip angle of the vehicle by adopting an unscented Kalman filter algorithm; solving an ideal side slip angle and an ideal yaw velocity; calculating to obtain a deviation value between the ideal side slip angle and the current side slip angle and a deviation value between the ideal yaw velocity and the current yaw velocity; calculating to obtain ideal lateral force and yawingmoment; learning a steering angle and torque distribution strategy by using a reinforcement learning method, and outputting four-wheel independent steering angle and torque; and executing the generated distribution strategy, and returning a reward value which is used for evaluating the quality of the executed action. According to the method, autonomous learning and improvement of four-wheel independent steering angle and drive distribution are achieved, the operation stability and tire durability of the automobile are comprehensively considered, and limitation of a traditional control methodis avoided.

The invention discloses a transformer substation personnel detection method based on deep learning. The method comprises the steps: training a transformer substation personnel track and behavior detection model based on deep learning, and carrying out the transformer substation personnel detection through employing the transformer substation personnel track and behavior detection model based on deep learning. The invention further provides a transformer substation personnel detection device based on deep learning, the transformer substation personnel detection device comprises a memory, a processor and a computer program which is stored in the memory and can run on the processor, and the processor implements the steps of the method when executing the computer program. Characteristic selection does not depend on manual experts any more, objects such as safety clothes, safety helmets and work permits can be efficiently and accurately recognized, and the capacity quite meets the requirement of an unattended operation mode of a current transformer substation.

The invention discloses a robot motion skill learning method and system. The method comprises the steps: acquiring a data sample set of human dragging teaching; performing dimension reduction processing on the data sample set based on a principal component analysis method; establishing a variable constraint condition in an implicit space, and screening the data sample set after dimension reduction processing in combination with the variable constraint condition to generate an implicit space data set; performing modeling learning on a hidden space data set by adopting a Gaussian mixture model and a Gaussian mixture regression method, and outputting a robot motion control training model; and predicting the robot motion control training model based on a recurrent neural network, solving a model optimization solution, and converting a model optimization solution into a robot actual control quantity. In the embodiment of the invention, the autonomous learning of the robot motion skills can be realized by using a small amount of human teaching data and considering the inherent constraints of the robot body, so that the generalization ability and programming efficiency of the algorithm are effectively improved.

The invention provides a power equipment state monitoring and fault identification method and system. According to the method, the sound source signals are collected based on the noise cancellation method, the sensitivity to the environment is reduced, and the sound signal processing precision can be improved. The power equipment is pre-monitored by judging whether the frequency spectrum value ofeach point of the field signal is within the confidence coefficient of the class center frequency spectrum value of each point of the normal sample signal or not, the field signal of the power equipment is screened, and a small number of field fault signals needing fault identification are screened out, so that real-time monitoring is facilitated, and the fault identification efficiency is improved. The algorithm processing speed is improved, the safety, efficiency and accuracy of power equipment fault detection are improved, sensors do not need to be arranged in a one-to-one correspondence mode, and the cost of power equipment fault detection is reduced.

The invention provides an intelligent desk for tutoring learning. The intelligent desk comprises a desk body, a support arranged on the desk body and a host. A plurality of cameras for acquiring stateinformation of learners, a scanning device for scanning homework information, a display device and a loudspeaker are arranged on the support; the host comprises a control system and a network communication module. The control system comprises a tutoring program module for transmitting scanned homework information to an external network server to complete correction and returning a correction result and exercise explanation; a supervision program module used for outputting a prompt according to the learner state information acquired by the cameras; and a report program module which generates an analysis report based on the operation data of the tutoring program module and the supervision program module and transmits the analysis report to an external monitoring terminal. The intelligent desk can assist a learner in completing homework correction and exercise explanation, and the learner is kept in a good learning state and can develop a good learning habit through reminding; and the learning effect can be timely and accurately fed back to parents.

The invention discloses a parallel intelligence-based chronic disease diagnosis and treatment system and diagnosis and treatment method. The system comprises an actual diagnosis and treatment system,a virtual diagnosis and treatment system and a virtual-real parallel execution system; the actual diagnosis and treatment system comprises an actual doctor, an actual patient and a process that the actual doctor diagnoses diseases for the actual patient and formulates an actual treatment scheme; the virtual diagnosis and treatment system comprises an intelligent doctor, an intelligent patient anda process that the intelligent doctor carries out intelligent diagnosis and makes an intelligent treatment scheme according to information of the intelligent patient. The virtual-real parallel execution system comprises interaction of disease diagnosis results and disease treatment schemes between the intelligent doctor and the actual doctor, and monitoring feedback and abnormal intervention of treatment effects in the actual patient treatment process. The chronic disease diagnosis and treatment system disclosed by the invention can ensure the accuracy and reliability of a diagnosis and treatment scheme, ensures the treatment effect of a chronic disease patient, and provides a reliable and effective solution for intelligent chronic disease diagnosis and treatment research.

The invention relates to an urban scene semantic segmentation fine-grained boundary extraction method based on a laser point cloud, and the method comprises the steps: designing a deep convolutional neural network model adaptive to an urban scene, training the model, carrying out the semantic segmentation of image data in obtained data through employing the trained model, and obtaining a fine-grained boundary of the urban scene. Obtaining a city scene preliminary semantic segmentation result based on the 2D image; carrying out fine boundary extraction through a post-processing conditional random field and an embedded conditional random field; and finally, calculating exterior orientation elements of the camera according to interior orientation elements of the camera through a direct transformation algorithm, obtaining overall mapping between the image and the corresponding laser point cloud, inputting the point cloud on the basis, and obtaining a refined semantic segmentation result based on the laser point cloud. The precision and the effect of semantic segmentation of the city scene can be improved.

The invention discloses a basketball training blanket based on a sensing technology, comprising a training blanket body, wherein a starting point is arranged on the training blanket body, and a take-off point is arranged on an end opposite to the starting point; taking the connection line of the take-off point and the start point as the central axis, the step changing point is symmetrically arranged on the training blanket body on both sides of the central axis; a simulation system includes a motion capture suit, a touch sensor, a motion capture camera and a master controller. The touch sensors are respectively arranged below the take-off point, the start point and the step change point; a main controller is respectively connected with a motion capture camera and a touch sensor, a human body model is stored in that main controller, a display screen is arrange on the main controller, and the main controller restores the movement of the trainer on the human body model according to the movement data and the data sent by the touch sensor, and displays the movement on the display screen. It can simulate the movement of training, so as to ensure the standardization of movement and improve the training effect.