60results about How to "Simplify the learning process" patented technology

A method of and a system for upgrading a universal remote control with a command code of an electronic appliance are provided. The command code is downloaded from an internet server into a personal computer or set-top box, which is coupled to an infrared transmitter. The universal remote control is placed in front of the infrared transmitter and the command code is transferred to the universal remote control, for example in response to the user indicating a key, shown in a displayed picture of a control panel of the electronic appliance and associated with said command code.

The invention provides an electronic nose gas identification method based on source domain migration extreme learning to realize drift compensation. According to the source domain migration extreme learning to realize drift compensation, a source domain migration extreme learning machine framework is proposed from the perspective of machine learning and used for solving the problem of sensor drift instead of direct correction for single sensor response; a source domain data set and a target domain data set are built according to labeled gas sensor array sense data matrixes collected by an electronic nose before drift and after drift respectively and are taken as inputs of an extreme learning machine for training an identification classifier of the electronic nose, so that the tolerance performance of the identification classifier on gas identification after the electronic nose drifts is improved, and the purposes of drift compensation and gas identification precision improvement are achieved; besides, technical advantages of the extreme learning machine are kept, and accordingly, the method has better generalization performance and migration performance. Therefore, based on the source domain migration extreme learning machine framework provided by the invention, one learning framework with good learning capacity and generalization capacity is built.

The invention provides a target domain migration extreme learning-based electronic nose heterogeneous data identification method. A domain migration extreme learning machine framework is put forward from an angle of machine learning and is used for solving a problem of sensor drift. Via gas sensor array sensing data matrixes which are collected via an electronic nose before drift and labeled and non-labeled gas sensor array sensing data matrixes which are collected via the electronic nose after drift, a source domain data set, a target domain data set and a data set for a domain to be tested are respectively established and are respectively used as input for an extreme learning machine, an electronic nose identification and classification machine can be learned, gas identification tolerance performance of the identification and classification machine can be improved after the electronic nose is drifted, an aim of drift compensation and improvement of heterogeneous data sample identification precision in gas identification is attained, technical advantages of the extreme learning machine are kept, good generalization and migration performance of the method are realized, and the method can be widely applied to different electronic nose products for identifying different gases.

The inVention discloses a learning attention eValuation system based on an EEG signal, which comprises an EEG acquisition subsystem, which is used for acquiring an EEG signal for preprocessing to remoVe power frequency interference and a baseline, and transmitting the EEG signal to an EEG analysis subsystem; the EEG analysis subsystem, which is used for analyzing the acquired EEG signal, extracting the time domain characteristic and the frequency domain characteristic as the feature Vector of the attention classification, and inputting the feature Vector into an attention classifier to obtainthe attention leVel; a learning state display reminding subsystem, which is used for feeding back the attention leVel obtained by the EEG analysis subsystem to students and teachers, thereby achieVingreal-time monitoring of a student's attention and solVing the technical problem that the student's learning state is difficult to eValuate objectiVely.

The invention provides a quality control chart pattern recognition method based on improved genetic algorithm optimization. The quality control chart pattern recognition method comprises the followingsteps: simulating various pattern characteristics of a control chart by using a Monte Carlo method; generating data of a corresponding mode through the parameter values; adopting the PCA principal component analysis method to carry out dimension reduction and denoising on the original data, main features of the data are extracted, shortening the training time of the model and improvintg the recognition accuracy; establishing a probabilistic neural network model, and carrying out pattern classification recognition by utilizing the characteristics of simple structure and convenient training; optimizing a main parameter smoothing factor of the probabilistic neural network by virtue of an improved single-objective optimization genetic algorithm; searching possible abnormal reasons from different aspects according to the identification result.The method solves the problems that all abnormal conditions cannot be monitored and recognized when an existing enterprise carries out quality control, effective abnormal information is difficult to find from a control chart, and appropriate measures cannot be taken to correct the abnormal conditions in the production process.

The present invention provides an image classification method based on a concise unsupervised convolutional network, and belongs to the image processing and deep learning technology field. The method of the present invention utilizes a classic unsupervised clustering algorithm K-means to cluster the image blocks of a training image set, the obtained each clustering center is a convolution kernel in a network model, and the time-consuming process in the conventional convolutional network of utilizing a stochastic gradient descent algorithm repeatedly to obtain the convolution kernels is abandoned. In addition, the present invention provides a probability pooling method to enhance the robustness of the network to the image deformation. By a concise unsupervised depth convolutional network classification model provided by the present invention, the model training time can be shortened effectively, at the same time, the identification capability of the model to the changeable scene pictures is improved.

The invention relates to a miniature circuit breaker fault analysis method based on a probabilistic neural network. The method comprises the following steps: (1) constructing a probabilistic neural network model; the system comprises an input layer, a mode layer, a summation layer and an output layer; (2) training and testing a probabilistic neural network model; respectively simulating each groupof vibration signals of the miniature circuit breaker under N fault types, carrying out classic modal decomposition on each group of vibration signals, and extracting a fourth-order intrinsic mode function IMF component; calculating the correlation dimension of the original signal and a fourth-order intrinsic mode function IMF component by using a G _ P algorithm; and (3) outputting a test resultthrough the trained probabilistic neural network model in the step (2). The probabilistic neural network is simple in learning process and high in training speed; classification is more accurate, andit can be guaranteed that the optimal solution under the Bayesian criterion is obtained; fault tolerance is good, and training data is allowed to be increased or decreased without long-time training.

The invention relates to a power transformer fault early warning system based on data mining, and the system comprises a power transformer full-dimension original data set module, a seamless embedded data application interface module, a power transformer full-dimension original data set cleaning module, a power transformer high-quality sample data set module, and a power transformer core algorithm module. a power transformer fault early warning model analysis module, and a human-computer interface display module. Compared with the prior art, the system has the advantages that massive electric power big data generated in intelligent power grid construction is fully utilized, implicit and potential value information is extracted by adopting a data mining method, the modeling speed is high, the learning process is simple, and a complete fault knowledge base is not needed; the defects of low modeling speeds, low accuracy and the like of traditional modeling methods are overcome, the problems of low maintenance pertinence, incapability of distinguishing primary and secondary maintenance, rigid maintenance mode and the like in planned maintenance of the power transformer are solved, and the requirement of fault early warning of the power transformer is met.

The invention discloses a method of constructing a smart vehicle EPS-used AC motor anti-interference smart controller. A d-q axis current decoupling control module, a vector control module, a voltage coordinate conversion module, a PWM regulation module and the AC motor are connected in series, together with a current coordinate conversion module and a disturbance detection module as a whole, an EPS motor system is formed; a neural network controller, an optimization controller, a robust controller and a robust controller parameter optimization module are connected in parallel and then, together with an angle setting module and a filter tracking error model, an anti-interference smart controller for the EPS motor system is formed, and thus, the static control performance and the anti-interference control for the EPS motor system are improved, and the neural network controller control precision is ensured.

The invention relates to an intelligent agent control method and device based on reinforcement learning. The method comprises the steps of obtaining current local observation of an intelligent agent;taking the current local observation of the intelligent agent as the input of a reinforcement learning model, and obtaining the current execution action of the intelligent agent output by the reinforcement learning model; controlling an intelligent agent to execute the current execution action of the intelligent agent; according to the technical scheme provided by the invention, the strategy learning process in a large-scale multi-intelligent system can be effectively simplified, the number and types of intelligent agents are easy to expand, and the method has potential value in large-scale real world application.

The invention discloses a demagnetization detection apparatus and detection method of a permanent magnet synchronous linear motor. The demagnetization detection apparatus of a permanent magnet synchronous linear motor is characterized in that a linear motor with a demagnetization detection apparatus is used as a detection motor; the demagnetization detection apparatus takes an aluminium block as apedestal; a gaussmeter probe is fixedly arranged on the aluminium block by means of a clamp, so that the probe is arranged in the space air gap of the linear motor to be detected; adjusting bolts areused to adjust the positions of three detection points of the gaussmeter probe in space air gap of the linear motor to be detected; and the demagnetization detection apparatus is used for detecting distribution of air-gap flux density at the positions of the three detection points of the linear motor to be detected so as to realize demagnetization detection. The demagnetization detection apparatus and detection method of a permanent magnet synchronous linear motor can determine the difference distribution between the flux density to be detected and normal flux density distribution, and can extract the demagnetization fault characteristics. The demagnetization detection apparatus and detection method of a permanent magnet synchronous linear motor can realize accurate identification of thedemagnetization fault by means of a PNN classification algorithm.

The invention belongs to the technical field of power supply equipment and particularly relates to portable power supply equipment which comprises an energy storage battery device, a pressure and electric energy conversion device, a voltage and current integration device and a frame supporting device. The energy storage battery device is fixedly connected with the frame supporting device; the pressure and electric energy conversion device and the frame supporting device are connected with each other and arranged above and / or below the energy storage battery device; and the voltage and current integration device is connected with the energy storage battery device and the pressure and electric energy conversion device. Compared with the prior art, under the condition that the portable power supply equipment is not provided with an external power supply for supplementing the electric energy, the piezoelectric material can convert the mechanical energy into the electric energy directly through shaking and other mechanical actuations, the electric energy is integrated by the voltage and current integration device, stored into a battery and converted into the chemical energy. Under the condition that no standby battery or power supply needs to be prepared, the electric energy of a power supply can be supplemented so that electronic equipment can work normally for a long time.

The invention provides a target tracking method based on spatial offset learning, which relates to the technical field of computer vision. The target object to be tracked is selected and determined, and the target selection process is automatically extracted or manually specified by the moving target detection method. Spatial migration learning network includes four parts: image data extraction, depth neural network, multi-layer perceptron MLP and spatial migration output. In real-time processing, the video image collected by camera and stored in storage area is extracted as the input image tobe tracked. In the case of offline processing, the captured video files are decomposed into image sequences composed of multiple frames. Short-time tracking adopts particle filter, in which a particle represents a possible target image block, and the corresponding target attention area ROI is used as the online training set of the spatial offset learning network, and the stochastic gradient descent (SGD) method is used to online train the spatial offset learning network and update the network parameters. Target Location and updating are carried out.

The invention discloses a security system which comprises a mainframe, a radio network module, radio alarm equipment, a radio alarm signal receiving module and a monitoring terminal, wherein the radio network module and the radio alarm signal receiving module are respectively connected with the mainframe; the monitoring terminal is networked with the mainframe through the radio network module; and a radio signal decoding single chip microcomputer is connected between the mainframe and the radio alarm signal receiving module and is used for encoding data signals sent out by the radio alarm signal receiving module into a uniform ID (identification) number and transmitting the ID number to the mainframe. Through arranging the radio signal decoding single chip microcomputer, an operator can control the mainframe through the monitoring terminal to willfully study different types of alarm sensors, so that the type and the quantity of the radio security alarm sensors can be willfully extended, and the extensibility of the security system can be improved.

The invention discloses a speed and precision balanced steel product surface defect detection method, and belongs to the technical field of steel surface defect detection, and the method comprises the following steps: S1, obtaining a hot rolled strip steel surface defect typical image sample from a database, and preprocessing the sample; s2, on the basis of a Center Net target detection model, designing a hot-rolled strip steel surface defect target detection model comprising skip connection and a pyramid feature fusion module, and selecting a backbone network as a feature extractor of the target detection model; s3, initializing parameters of a backbone network structure part in the target detection model by adopting parameters trained on an ImageNet data set, and training the target detection model by using a training sample; and S4, testing a test sample by using the trained target detection model, and outputting a detection result. The invention is high in accuracy of detecting the target position of the surface defect of the hot-rolled strip steel, high in detection speed and capable of being effectively applied to field real-time detection of the surface defect of the hot-rolled strip steel.

The invention discloses a self-adaptive network suitable for high-reflection bright spot segmentation in a retinal optical coherence tomography image. The self-adaptive network comprises a feature coding module, a self-adaptive SA module and a feature decoding module, the feature coding module comprises a feature extraction unit and a dual residual DR module, the dual residual DR module comprisestwo residual blocks, the adaptive SA module comprises a feature input end, a deformable convolution layer, matrix multiplication and pixel-level summation, and the feature decoding module reconstructshigh-level features generated by the adaptive SA module, gradually performs feature splicing with local information guided by the dual residual DR module through 2 * 2 deconvolution layer deconvolution, and uses a result obtained through 1 * 1 convolution layer convolution as output of the feature decoding module. The method can simplify the learning process of the whole network and enhance gradient propagation while enhancing feature extraction, and can adapt to segmentation targets of different sizes.

A classification-based diagnosis for detecting and predicting faults in physical system (e.g. an electronic circuit or rail switch) is disclosed. Some embodiments make use of partial system model information (e.g., system topology, components behavior) to simplify the classifier complexity (e.g., reduce the number of parameters). Some embodiments of the method use a Bayesian approach to derive a classifier structure.

The invention discloses an NLP-based weblog processing system. The weblog processing system comprises a natural language processing assembly and a database. A classification lexicon, a preset word meaning library and a language processing model are constructed in the database; the classification lexicon sets a plurality of literal translation words specifically corresponding to equipment types orkeywords taking high-frequency words derived after word segmentation processing as standards; the classification word bank is in a mapping relation with a preset word meaning bank, and the preset wordmeaning bank is associated with a language processing model; the natural processing component is used for carrying out classification on syslog source data and log files of the equipment, and analyzing and determining meanings contained in natural language statements. According to the method, the defect that an undefined log cannot be analyzed by a conventional template-based method is overcome,availability of the system is improved, and usability of a user is improved.

The invention discloses a deep learning wind power alarm information analysis method based on small sample data. The method includes extracting characteristic parameters, wind speed and output powerof all components of a fan as the training samples, adopting a contrast divergence algorithm to train a restricted Boltzmann machine model, carrying out characteristic quantity extraction on data, carrying out statistical classification on the extracted characteristic quantity by using a probabilistic neural network algorithm, dividing a corresponding fault risk probability, and carrying out earlywarning according to a fault risk level, thereby realizing fault warning of the wind turbine generator set. According to the method, characteristic quantity extraction is carried out on the data sample through the restricted Boltzmann machine, the effect of dimension reduction is achieved, a good foundation is provided for subsequent data analysis, and the complexity of data analysis is reduced;and the problem of local minimum is solved, and the accuracy of the fault probability risk level is improved.

The invention discloses a human muscle movement perception based menu selection method for a human-computer interaction interface. The method comprises the steps of setting initialing parameters; determining a user front space through a kinect sensor; dividing and displaying the user front space; catching the user movement information; determining the manual selection result of the user human-computer interaction interface. According to the method, the 3D physical contact technology is adopted, and the kinect sensor is used for detecting the body front space and performs a series of division for the space in front of the body, including vertical plane division and longitudinal horizontal division; each divided space is calibrated through a space coordination system; the manual selection at the human-computer interaction interface is performed by selecting the space by human hands; the corresponding instruction and operation are sent to the machine to avoid complex memory of a user to the action and gesture.

The invention discloses a paper cloud interactive language teaching system. The system comprises a data processing and storage module, and a paper cloud handwriting interactive module and a voice playing and collecting module connected with the data processing and storage module. The paper cloud handwriting interactive module records user interaction data including written contents; and the voiceplaying and collecting module plays audio contents and collects user interaction data including voice data. The system provides interactive services including content-on-demand, intelligent paper marking, dictation practice and listening-reading practice; and the data processing and storage module identifies the voice data and compares the voice data with standard pronunciation to give a pronunciation score. A paper cloud interactive language teaching method comprises the steps that the paper cloud handwriting interactive module and the voice playing and collecting module are used for matchingprinted learning materials to use learning languages including reading, hearing, spoken language and writing. The system has the characteristic that the deficiency of effective spoken language training in language learning in the prior art is supplemented.

The invention provides a method and equipment for constructing a calculation material simulation platform. The method comprises the following steps: configuring a server underlying architecture so that application software used for a material module can be installed and compiled on a server; taking the VASP software as an interface to connect the corresponding software interface with the interfaceof the VASP software to form a software set, installing the corresponding software in the software set in a server, and optimizing the software based on the underlying architecture of the server; keywords are set, the keywords are searched in an input file of the VASP software to obtain secondary application software needing to be called by the VASP software, and all the application software canbe mutually called; and configuring a graphical visual interface in the server, and associating the visual interface with corresponding software in the software set. By using the scheme of the invention, a scientific researcher can carry out a whole set of analysis and simulation on the platform, the time cost of software installation, deployment, learning, use and updating is greatly saved, and the working efficiency of the scientific researcher can be effectively improved.

The invention discloses a language studying method and a system thereof. The method comprises the steps of according to the student grade of a logged-on student, acquiring a to-be-studied course whichcorresponds with the student grade; analyzing the to-be-studied course, acquiring preview problems which correspond with the to-be-studied course, and pushing the preview problems to a student clientwhich corresponds with the logged-on student, thereby making the logged-on student perform previewing of the to-be-studied course based on the student client; acquiring the finished preview problemsof the logged-on student from the student client, performing scoring on the finished preview problems, acquiring a scoring result, obtaining a preview problem score and transmitting to the student client for displaying, thereby realizing observation by the logged-on student; if the preview problem score reaches a preset score, supplying a course reserving entry on the student client, thereby realizing online course reserving by the student based on the course reserving entry. The language studying method and the system thereof have advantages of simplifying a studying process, improving studying efficiency and preventing unnecessary time waste.

The invention discloses an electronic tag Bluetooth opening method and device. The method comprises the steps that the current state of an electronic tag is detected; and starting a Bluetooth functionof the electronic tag according to the current state. By adopting the technical scheme of the invention, the learning and operation process of a user on the Bluetooth opening mode after installationis simplified, the use interaction of the electronic tag is greatly optimized, and the user experience is improved.

The invention discloses a Pinyin-form Chinese character input method. The method includes a form character mapping step, namely mapping form characters of Pinyin-form Chinese characters to a keyboard or keys of a screen according to Pinyin codes corresponding to form character name sounds; a form character inputting step, namely acquiring the form characters of the Pinyin-form Chinese characters to be input through the keyboard or screen; a Pinyin characters inputting step, namely acquiring Pinyin characters of the Pinyin-form Chinese characters to be input through the keyboard or screen; and a display step, namely displaying the Pinyin-form Chinese characters to be input and / or corresponding current Chinese characters after analyzing the acquired form characters and Pinyin characters. The method is easy to learn and use and is a completer Chinese character input manner, the Chinese characters can be input directly and rapidly while people can be assisted to learn and use Chinese and Chinese characters, and the method has positive affect in development of the Chinese and Chinese characters and promotion the Chinese and Chinese characters to the worldwide.

The invention discloses a method for realizing multifunctional operation by using a single knob and a single knob device. The method is used for controlling an electronic product to execute multifunctional operation through a single-knob device, wherein the single-knob device comprises a knob, an indicating element and a control component; a key is integrated in the knob; and the knob and the indicating element are connected with the control component. The method comprises the following steps that: when the knob is operated, an operation signal comprising a rotation direction, a rotation angle and a key state is generated; and a control part identifies the operation signal and makes a response action, wherein the response action comprises controlling the electronic product to execute a corresponding function operation and driving an indication element to make a corresponding indication. According to the method of the invention, the single knob is used for realizing multifunctional operation on the electronic product, so that the operation can be simplified, the product is simplified, the user experience is improved, and the practical value is higher.

The present invention provides an electronic nose gas identification method based on source domain migration extreme learning drift compensation. It proposes a domain migration extreme learning machine framework from a machine learning perspective to solve the sensor drift problem, instead of directly correcting a single sensor response, using The labeled gas sensor array sensing data matrices collected by the electronic nose before drifting and after drifting were used to construct source domain data sets and target domain data sets respectively as inputs to the extreme learning machine, and the recognition classifier of the electronic nose was carried out. Learning to improve the tolerance performance of the recognition classifier for gas recognition after the electronic nose drifts, to achieve the purpose of drift compensation and improve gas recognition accuracy, and maintain the technical advantages of the extreme learning machine, making this method have better Generalization and transfer performance. It can be seen that the source domain transfer extreme learning machine framework proposed in the method of the present invention establishes a learning framework with good learning ability and generalization ability.

The invention relates to the technical field of electronic nose calibration, and particularly discloses an electronic nose drift general calibration method based on convex set projection and an extreme learning machine. From the perspective of machine learning, a constraint network net1 and a calibration network net2 are established based on the extreme learning machine, the constraint network net1 is trained by using an electronic nose in a feature data set X without drift, network parameters are stored, then a drifting feature data set Xd is taken as an input of a constraint network net1, iterative adjustment is performed on the input Xd of the net1 network based on convex set projection to obtain a calibrated sensor feature data set Xc, and then the feature data set Xc is taken as a label of a calibration network net2; and the feature data set Xd is input for common training so as to calibrate an unknown gas response signal so that the tolerance performance of gas identification after drifting of the electronic nose can be improved, and a network obtained by training can achieve a drifting compensation effect on an unknown gas sample. Therefore, the gas recognition precision of the electronic nose after other gas sensors drift is improved.

The invention provides an electronic nasal heterogeneous data recognition method based on target domain migration extreme learning. From the perspective of machine learning, it proposes the domain migration extreme learning machine framework to solve the problem of sensor drift.Gas sensor matrix sensor sensor matrix and label and label -free gas sensor arrays collected after drift are constructed from source domain data sets, target domain data sets, and data sets to be tested, respectively.The input of the machine is learned to the electronic nose recognition classifier to improve the tolerance of the identification classifier after drift of the electronic nose.The purpose, with the technical advantages of extreme learning machines, enables this method to have better generalization and migration performance, which can be widely used in different electronic nose products for different gas recognition applications.

The invention provides a communication interference existence detection method based on residual connection and expansion convolution, and the method comprises the following steps: 1, constructing an interference existence detection data set, and simulating system signals under different interference-to-noise ratios and system signals with interference; 2, establishing an interference existence detection model for detecting interference in system signals, wherein the interference existence detection model is based on residual connection and expansion convolution; 3, training the interference existence detection model in the step 2 by using the interference existence detection data set in the step 1, so that the interference existence detection model is converged, and a final interference existence detection model is obtained. The method has the characteristics of automatic extraction of interference characteristics in the signal, high accuracy of interference detection, and fast interference detection speed.