72results about How to "Guaranteed learning effect" patented technology

The invention relates to a panorama mode online learning system and a pen special for the learning system. According to the panorama mode online learning system and the pen special for the learning system, a courseware making system obtains an explanation process and motion and position of data display, acquires voice and the time corresponding to the voice when the data are obtained, and forms coding courseware after the data are processed according to a courseware making standard; the standard courseware is decoded and played in a live broadcast mode or a recorded broadcast mode; homework assigned by the system and finished by students is uploaded to a central server and assigned to a teacher for amending, and the amended homework is then sent back to the students; a detailed experiment process or an experiment result is displayed; and the panorama learning pen is specially used for the system and facilitates electronic courseware making, homework amending and homework solution. According to the panorama mode online learning system and the pen special for the learning system, the explaining process, the voice and the courseware data are played synchronously for the students, and therefore the teaching purpose of panorama simulation of a real classroom is achieved.

A digitalized piano is disclosed and capable of network demonstration piano performing and real time video teaching. The digitalized piano comprises a mechanical piano, a computer host, a video information collection module, an audio information collection module, a key movement information collection module, a footboard movement information collection module, a key information and footboard information display screen, a touch display screen, a network data transmission module and a loudspeaker module. The computer host is connected with all modules via a USB bus. Via use of the digitalized piano disclosed in the invention, an entity teaching mode is replaced with a home self-learning mode and a network remote teacher guidance learning mode; piano learning demand in areas where high-quality piano teacher resources are not available can be met, and demand of users who hope to learn piano playing at home at low cost can be met.

The invention provides a mandatory interaction system and method of multimedia courseware. The method comprises the following steps: dividing the multimedia courseware into a plurality of fragments, wherein each fragment corresponds to a knowledge point; setting, at any place of each fragment, one or a plurality of interaction nodes which are related to the knowledge point of the fragment; generating an examination type interaction content on the interaction node in allusion to the knowledge point corresponding to the current fragment; setting the triggering manner of the interaction node as mandatory triggering; displaying the corresponding examination type interaction content when triggering the interaction node, and judging whether the examination is passed according to the feedback of the learner; and only when the examination of the examination type interaction content is passed, allowing to enter the next fragment. According to the mandatory interaction system and method of the multimedia courseware, the interaction with the learner can be mandatorily carried out while knowledge points are transferred to the learner; and only when the learner passes through certain examinations, the learner can learn the next knowledge point, so that the learning effect is ensured, and the learner can learn anytime anywhere without the limitation of time and place.

The invention provides a method and a device for selecting terminal equipment nodes in a federated learning system. The method comprises the following steps: determining a test accuracy optimization target model selected corresponding to a terminal device node; and obtaining current environment state data corresponding to each federated learning iteration process, and inputting the current environment state data into a terminal equipment node selection model to obtain a terminal equipment node selection strategy corresponding to test accuracy and time delay constraints output by the terminal equipment node selection model. By adopting the method for selecting the terminal equipment nodes in the federated learning system, the federated learning performance can be ensured, the joint optimization of the resource utilization rate and the energy consumption of the terminal equipment is realized, the terminal equipment set is reasonably selected to participate in the aggregation of the federated learning model, and the efficiency and the accuracy of the federated learning are effectively improved.

The present invention relates to a method and an apparatus for embedding and extracting an image watermarking based on BP and RBF neural networks. An original watermarking image is scrambled through affine transformation, then the scrambled watermarking image is encrypted through an improved BP neural network algorithm, after that, the encrypted watermarking image is embedded into an original carrier image; the encrypted watermarking image embedded in the carrier image is extracted, then the encrypted watermarking image is recovered through affine inverse transformation and a BP neural network algorithm, so as to obtain the original watermarking image. Through adoption of the method and the apparatus, secondary encryption for the watermarking image is achieved, the number of keys is increased, and degree of safety of the watermarking is improved.

The invention belongs to the technical field of fault diagnosis for main electric power equipment, and particularly relates to a transformer three-ratio fault diagnosis method improved by a B-spline theory. The method comprises the following steps of: collecting sample data of gas dissolved in oil to construct a sample library, forming a three-dimensional space by a three-ratio method to obtain the characteristic area of each fault mode, and establishing each fault interface equation expressed by a B-spline curved surface; calculating the position relationship between a sample and a fault interface to judge the fault type of a transformer; for the fault sample with wrong diagnosis, re-calculating a curved surface parameter by virtue of the inverse computation method of the B-spline curved surface, and bringing the sample with wrong diagnosis into a correct fault characteristic area; calculating the position relationship between the sample to be diagnosed and each interface curved surface after shape adjustment to realize judgement on the fault type of the transformer. According to the method disclosed by the invention, the fault characteristic area determined by improving the three-ratio method is taken as a basis, and a dynamic learning function is realized through flexibly and self-adaptively adjusting the interface curved surface shape in the fault area; the method is great in practicability and generalization performance.

The invention discloses a personalized online learning and occupation two-way recommendation method and system, and the method and system can recommend a course which needs to be learned most according to an interested job hunting target of a learner, enable a user to be targeted in a learning process, and guarantee the learning effect of the user. When the user does not specify the job hunting target, the most suitable occupation can be recommended according to the preset knowledge and the online learning condition of the user, the occupational planning consciousness of the user is cultivated, and the most suitable job seeker is selected according to the learning condition of the user for an enterprise, so that the method has very important significance for improving the online learning quality of the user and developing job seekers.

The invention discloses a wireless federated learning-oriented mobile user equipment clustering training method, which is characterized in that an edge service server and a plurality of mobile user equipment are arranged in a wireless network, and each user equipment belongs to a local data set. Firstly, a DBSCAN algorithm and an LEACH algorithm are used to divide users into a plurality of clusters and select cluster heads. Then, each user uses local data to train respective model parameters and uploads the model parameters to a cluster head of a cluster where the user is located for aggregation, and the cluster head further uploads the aggregated model to an edge server for aggregation; and the edge server distributes the aggregated model parameters to the cluster head, and the cluster head further distributes the aggregated model parameters to the user equipment in the cluster for the next round of training. And repeating the steps until convergence. By using the clustering training method, the communication overhead and the training time delay of wireless federal learning can be reduced.

The invention relates to the technical field of image recognition and provides a data sampling method and a data sampling device for solving the problem of long tail of a training data set in the prior art, and aims to solve the problem of incomplete network feature learning caused by the fact that training samples are the same in each network training process from the beginning to the end of network training, and the method comprises the following steps: acquiring the current iteration round of a network; updating the sample weight of each type of samples based on the current iteration roundof the network and the sample number of each type of samples, with the increase of the current iteration round of the network, the weight of the tail type of samples being gradually increased; and according to the updated sample weight of each type of samples, sampling samples meeting a preset condition as target samples. Through the increase of network iteration turns, the sample weights in the training data set are increased, but the increase amplitude of the weights of the tail class samples is large, the probability that the tail class samples are selected as the training samples is increased, the problem of long tails can be effectively relieved, and the network feature learning effect is good.

An English translation teaching management system is characterized by comprising a central computer, a student terminal group, a teacher terminal group, a compiling terminal group, a compiling server, a standard server, a unit test server, a question doing server and a teacher server. An English translation teaching management method comprises the following steps: a question screening step, an input step, a lexical and syntactic mapping relation compiling step, a question scoring step, a first unit test generation step, a student question making and feedback step, a question customization step, a subsequent unit test generation step and a periodic arrangement step.

The invention relates to an implementation method for self-adjusting of pen holding postures. The implementation method is characterized in that a pen container, a three-axis gyroscope and a telescoping mechanism are provided, the three-axis gyroscope is arranged in the pen container and used for inducing the inclination angle of the pen container, the telescoping mechanism is arranged in the pen container and used for driving refills to stretch out and draw back, three touch sensitive sheets are uniformly distributed on the periphery of the hand holding portion of the pen container, an LED lamp is correspondingly arranged at the position, located over each touch sensitive sheet, of the pen container, a first LED lamp bank is formed by the three LED lamps, four LED lamps are uniformly distributed on the periphery of the upper portion of the pen container and form a second LED lamp bank, different angles of the pen container are detected under three-dimensional coordinates so that guidance can be given to people holding pens, and therefore learning of pen holding modes is completed. The automation degree of an adjusting method is high, the requirement of special crowds for learning of the pen holding modes can be effectively met, and learning effects are guaranteed.

The invention provides an online education course management method and system based on artificial intelligence, and relates to the technical field of learning management. The method comprises the steps of inputting the first course plan image information into a training model; obtaining the output information of the training model, wherein the output information comprises first completion degreelevel information and first concentration degree information corresponding to the first course plan image; screening and filtering the first course plan image information according to first completiondegree grade information and first concentration degree information corresponding to the first course plan image to obtain standard course plan image information; obtaining first time planning information of a first user; judging whether the matching degree of the first time planning information and the standard course planning image information meets a first preset condition or not; and when thematching degree of the two meets a first preset condition, taking the standard course plan image information as a first course management standard of the first user.

The invention relates to an Internet-based discussion method, system and commercial platform. The method comprises the following steps: acquiring initial registration information of a user, and generating a member user and corresponding member information, wherein the member user establishes a discussion group within a certain range according to a preset establishing mode; analyzing and evaluatingdata information of the member user and the discussion group; and calculating a comprehensive score. According to the Internet-based discussion method disclosed by the invention, a discussion type teaching environment is established, so that the member user can discuss and learn with each other through the Internet without limitations of time and locations, and parents and teachers also can be invited into the discussion group for auditing and supervision so as to guarantee a discussion order and a learning effect; and furthermore, the Internet-based discussion method can provide comprehensive and accurate discussion data analysis for the member user, so that the member user further knows own discussion level, thereby directionally improving the comprehensive performance.

The invention belongs to the field of software quality guarantee, and discloses a software defect module severity prediction method based on an ordered neural network. The method comprises the following steps that a version control system and a defect tracking system where a sample software project is located are mined, measurement and defect severity marking are conducted on a program module of the sample software project, and a sample data set is constructed; based on the sample data set, an ordered neural network model and a Bayesian hyper-parameter optimization method are adopted to obtaina software defect prediction model, namely, the ordered neural network model with optimal hyper-parameters; and the defect severity of the program module in the software project is predicted by adopting the software defect prediction model. Compared with a conventional software defect prediction method, the method for constructing the software defect prediction model has the advantages that software defect prediction can be performed, the severity of defects can be predicted, and the prediction accuracy is higher.

The invention discloses an urban rail vehicle frame repair simulation learning system, which comprises a console and a display stand, and further comprises a supporting mechanism arranged at the top end of the display stand; two mounting frames which are arranged at the top end of the display stand and located on the two sides of the supporting mechanism respectively; a cover plate arranged at the top ends of the two mounting frames; a frame assembly movably arranged at the top end of the supporting mechanism; a display assembly arranged on the two mounting frames; a connecting rod, wherein the two ends of the connecting rod are movably arranged on the two mounting frames respectively; and a moving assembly arranged on the supporting mechanism. According to the urban rail vehicle frame repair simulation learning system, a real scene feeling of watching below a train can be realized, and observation dead angles are avoided by virtue of a scene synthesis technology of a panorama camera; and the safety protection performance is improved.

The invention discloses an intelligent teaching system, which relates to the field of intelligent education and comprises a cloud end, a plurality of classroom ends and a plurality of user ends, the cloud is used for storing the teaching video collected by the classroom end and allowing a user to register in the cloud and log in the cloud to watch the teaching video; the user ends are used for logging in the cloud and playing the teaching video stored in the cloud; and each classroom end comprises an intelligent blackboard, a teaching video acquisition unit, a plurality of teaching video watching units and a server. According to the system, a traditional teaching mode is improved, the traditional teaching mode is combined with online education, and the teaching quality and the learning effect are guaranteed.

The invention discloses a deep cluster neural network model construction method for aerodynamic data processing, and relates to the technical field of deep cluster neural network model construction in the aerodynamic data processing field. The method comprises the following steps: a, preparing and preprocessing a data set: firstly, obtaining an aerodynamic data set through a computational fluid dynamics method, and extracting main design parameters and response parameters in the aerodynamic data set; classifying the data, and dividing the data into a plurality of subsets; labelling each subset ; finally, dividing the pneumatic data set into a training set, a verification set and a test set; b, constructing a deep cluster neural network model; c, training a deep cluster neural network model; and d, verifying the deep cluster neural network model. When the cluster neural network obtained through training of the method is used for processing aerodynamic data, the defect of a traditional neural network model in the environment of insufficient sample quality caused by uneven sample sampling or uneven sample distribution can be overcome, and therefore the prediction precision of the model is improved.

The invention provides an online learning method and device, electronic equipment and a storage medium, and relates to the technical field of the Internet. The method can display a learning interfacethrough a graphical user interface of terminal equipment, and comprises the following steps: in response to a touch operation for a video playing control in the learning interface, obtaining and playing a learning video; after the learning video is played, splitting the learning content corresponding to the learning video in response to a touch operation on the split control, and displaying a plurality of split learning blocks; and in response to the operation on the learning control corresponding to each learning block, completing learning. In the learning process, the learning content can bepreliminarily understood and memorized through the learning video, and the learning content is subjected to block learning through the plurality of split learning blocks, so that the user can be further helped to understand the learning content block by block on the basis of the preliminary understanding and memorizing of the user, and the learning effect is ensured.

The invention relates to a model compression method and system and a computer readable medium. The model compression method comprises the steps of 1, obtaining a trained original model; 2, searching the optimal pruning rate of each layer in the original model, and then performing filter-level pruning on the model; 3, quantifying the pruned model; and 4, performing knowledge migration in a knowledge distillation mode, and then performing fine tuning on the model after knowledge migration on the training data set to obtain a compression model. Compared with the prior art, the invention has the advantages of automatic compression, adaptability to hardware environment, small model size, good performance, convenience, rapidness and the like.