33results about How to "Increase training methods" patented technology

The present invention is one kind of computer-aided language learning system and the user completes the answer to some question based on the prompt. The present invention includes at least followingsteps: establishing at least one sentential form sample in the sentential form data base, proposing some question with the question generating module, outputting the question in phonetic or text modeto the learner with the sentence making module and separating the corresponding answer into several units for recombination to form one new sentence which is delivered to the learner, accepting the phonetic or text input of the learner, comparing the input and the several answers in the sentence making module and delivering the comparing result to the learner.

The invention relates to artificial intelligence, and provides a system anomaly detection method and device, equipment and a storage medium. The method comprises the steps of respectively inputting amarked log, an unmarked log and an extended log of a to-be-detected system into three same training models in a training model set for training, and outputting probability distribution of abnormal levels of the three training models; calculating cross entropy loss and consistency loss output by the training models; predicting abnormal levels of the unmarked log and the expanded log according to the consistency loss, and iterating the training model set according to the cross entropy loss until the training model set converges to obtain a log anomaly detection model; and finally, detecting an abnormal log in system operation through the log anomaly detection model. In addition, the invention further relates to a blockchain technology, and the marked log, the unmarked log and the extended log can be stored in the blockchain. By optimizing a model training mode, model over-fitting is prevented, and the detection difficulty of the detection model on abnormal points in a system is reduced.

The invention provides a power transmission line foreign matter detection method based on an improved Faster RCNN, and the method comprises the following steps: carrying out the preprocessing of a collected foreign matter image, and increasing the scale of a training set; carrying out network improvement on the shared convolutional network part in the Faster RCNN; adjusting the size ratio of the anchor in the RPN network in the Faster RCNN; performing end-to-end joint training on the improved network; and carrying out foreign matter detection on the picture by using the trained detection model, framing the position of the foreign matter and identifying the type of the foreign matter. The invention relates to a power transmission line foreign matter detection method based on an improved Faster RCNN, which combines image processing and deep learning, and adopts the improved Faster RCNN to improve the accuracy of power transmission line foreign matter detection.

The invention discloses an electric energy metering device simulation training platform. The electric energy metering device simulation training platform comprises a cabinet, an acquisition terminal, a main three-phase electric energy meter, an auxiliary three-phase electric energy meter, a main meter junction box, an auxiliary meter junction box and terminal blocks. A 220V alternating-current power supply supplies power to the acquisition terminal, the main three-phase electric energy meter and the auxiliary three-phase electric energy meter; the main three-phase electric energy meter and the auxiliary three-phase electric energy meter are respectively connected with the acquisition terminal through a 485 communication interface; the acquisition terminal, the main three-phase electric energy meter, the auxiliary three-phase electric energy meter, the main meter junction box, the auxiliary meter junction box and the terminal blocks are connected to form A-phase, B-phase and C-phase voltage and current circuits. The electric energy metering device simulation training platform has the advantages that actual operation conditions of the onsite electric energy meters and the acquisition terminal are simulated under off-site test room conditions, all work contents of metering and acquisition operations of a substation are covered, requirements on staff training and defect elimination simulation are met, and safety and stability in operation of electrical metering equipment and efficiency in stable acquisition of electric quantity data are improved; further, equipment movement is facilitated owing to wheels mounted at the bottom of the cabinet.

The invention discloses a parameter estimation model training method and device, equipment and a storage medium. The method comprises the following steps: inputting a face image training set into a pre-constructed neural network model, estimating reconstruction parameters specified by three-dimensional face reconstruction, inputting the reconstruction parameters into a pre-constructed three-dimensional deformation model, and reconstructing a three-dimensional face corresponding to a training sample; calculating loss functions between the three-dimensional face and the training sample under multiple items of two-dimensional supervision information, and adjusting the weight corresponding to each loss function; and generating a corresponding fitting loss function based on each loss function and the corresponding weight, and performing reverse correction on the neural network model by using the fitting loss function to obtain a trained parameter estimation model. According to the technicalscheme provided by the invention, the parameter estimation model is trained through the loss function under multiple items of two-dimensional supervision information, so that the reference information in the training process is more comprehensive, and the estimation accuracy of the reconstruction parameters adopted during three-dimensional face reconstruction is improved.

The invention discloses an ultrasonic breast tumor automatic segmentation method based on an attention-enhanced U-shaped network. The method comprises the following steps: (A) constructing an attention-enhanced U-shaped network model for ultrasonic breast tumor automatic segmentation; step (B), establishing a mixed attention loss function of the attention enhancement U-shaped network model; and step (C), according to the attention-enhanced U-shaped network model and the mixed attention loss function, training the attention-enhanced U-shaped network model through a coarse and fine combination strategy to achieve segmentation of the breast ultrasound image lesion area. The method can be used for extracting the focus area of the mammary gland ultrasonic image, can effectively improve the accuracy of mammary gland tumor segmentation, is used for assisting a doctor in quickly and accurately positioning the focus area, reducing the workload of the doctor and relieving the defects of insufficient clinical experience and the like of young doctors, and has very important research value and application prospect for modern medicine.

The invention belongs to the technical field of distributed energy generation control, in particular to a wind power generation power prediction method based on a parallel wavelet neural network, comprising the following steps: 1) acquiring condition parameters of a wind farm and wind power output data; 2) constructing a training data set and a test data set of a prediction sample; 3) constructingand training a parallel neural network model; 4) after completing the training of the parallel neural network model, putting into use. The invention shortens the training time of the neural network,not only retains the advantages of the traditional wavelet neural network in carrying out the time series prediction, but also solves the shortcomings that the parameters of the wavelet neural networkare difficult to be determined, and improves the prediction accuracy.

An agile ladder structure for football training, comprising a fixed ladder frame, a hurdle and a storage mobile car; the structure body is provided with a fixed ladder frame; The distance adjustment placement groove and the ladder frame connecting groove are formed; the side wall of one end of the fixed ladder frame is provided with a ladder frame connecting block; the other end of the fixed ladder frame is provided with a ladder frame connecting groove; the upper end of the fixed ladder frame Ladder distance adjustment placement slots are symmetrically opened; ladder bars are installed inside the ladder distance adjustment placement slots; the upper end of the fixed ladder frame is also symmetrically connected and installed with hurdles; wheel composition. The agile ladder structure used for football training is provided with a ladder frame connecting block that fits closely with the ladder frame connecting groove, and the ladder frame connecting block can be embedded and installed inside the ladder frame connecting groove. The side is inserted into the connecting groove of the ladder frame for fixed connection, which is convenient for the lengthening and use of the structure body.

The invention provides an online micro-grid virtual simulation training system, comprising an online micro-grid virtual simulation training platform and an operation module, wherein the online micro-grid virtual simulation training platform comprises a scene creation module, a mode selection module, a feedback module, a background module, a discussion module and an information release module; theoperation module is used for operating the behavior of a virtual character, controlling the travel of the virtual character, operating device buttons and switches, and operating contents in micro-gridcontrol software. The online micro-grid virtual simulation training system provided by the invention improves the micro-grid training mode by means of the online micro-grid virtual simulation training platform, makes the training environment be more vivid and controllable, improves the learning motivation of the students, achieves virtual and real combination and virtual and real interaction between the virtual training and the real training, and realizes the gradual and hierarchical learning.

The invention discloses a second-hand commercial vehicle condition detection method based on an improved Faster RCNN prediction model, and the method comprises the steps: collecting an image of a to-be-detected part of a second-hand commercial vehicle, and dividing the image of the to-be-detected part into a training set and a test set according to a proportion; using an image marking tool to mark the category and the position coordinate of an accident point in the training set; and naming the training set as an image file in a preset format, scaling the image file in the preset format into a preset size suitable for network training through a bilinear interpolation method, and then performing data enhancement operation on the image file to obtain a preprocessed image. The method has the advantages that the deep learning technology and the image processing technology are combined, the accident point on the second-hand commercial vehicle in the complex scene is detected and recognized through the deep learning method, and the detection speed and precision are improved. The network structure and the training mode of the Faster RCNN prediction model are improved, the optimization algorithm and the data augmentation method are improved, and the detection effect is improved.

The invention relates to a sports sandbag. The sports sandbag comprises a sandbag part, a transmission part, a control part, a support part, a sandbag body of the sandbag part, a spring, a sensor, anda rubber plate. The transmission part comprises a single chip microcomputer, a battery, a Bluetooth, a terminal device, a voice broadcast device, and a sealed box. The control part comprises a controller, a wire and a power supply. The power supply controls the Bluetooth and the voice broadcast device. The wire is connected with the single chip microcomputer and the controller. The controller outputs signals to change the data processing mode of the single chip microcomputer and controls the on-off of the power supply. The controller is provided with a switch, a strength intensity button anda frequency button. The support part comprises a support body and an iron chain. The sandbag body is hung on the support body through the iron chain. According to the sports sandbag, by hitting the sandbag part, the hitting frequency and strength can be monitored, signals are transmitted to the terminal device through the transmission part, and the support part can maintain the stability of sandbag. By hitting the sandbag, the purposes of real-time monitoring and hitting frequency and strength adjustment by trainers can be achieved.

The invention provides a post competence training method, equipment and system, and relates to the technical field of business management. The post competence training method comprises the steps that,an operating instruction sent by a user terminal is received, and the operating instruction is used for instructing post competence training parameter information corresponding to the user terminal;according to the operating instruction, topic information corresponding to the instructing post competence training parameter information is selected and sent to the user terminal; the answer information input by the user terminal according to a topic is received; and the answer information is judged to obtain a corresponding post competence training result. The topic information of training personnel at various posts is set through self-defining, the master situation of various skills of the training personnel can be supervised and understood, the aim that the training personnel is trained isachieved, so that the management is more convenient, training methods of the training personnel is more convenient, the pressure of the training personnel is reduced, the waste of manpower resourcesis reduced, and the efficiency of business management is improved.

The invention discloses a Wushu arm power training device capable of preventing muscle atrophy. The training device comprise a cement base, wherein upright rods are symmetrically mounted at the top end of the cement base; a suspended cross rod is mounted at the tops of the upright rods; first fixed pulleys are rotatably connected in the middle of the suspended cross rod at equal distances in an embedded manner; a training belt is slidingly connected to the top end of each of the first fixed pulleys; a first handle is hanged at a bending part of each of the training belts; a counterweight blockis slidingly connected to one side of the middle of a counterweight rod, which corresponds to a traction belt. The training device is scientific and reasonable in structure, and safe and convenient to use; the training belts, the traction belt, the first handles, the counterweight rod and the counterweight blocks are arranged, so that the different quantities of counterweight blocks are placed atthe two ends of the counterweight rod, and a user pulls each of the first handles with different arm power. Therefore, the problem that the user has weaker arm power before onset of muscle atrophy issolved. In addition, after the user is trained for a long time, the arm power of the user can be effectively improved, and the muscle atrophy can be effectively prevented.

The invention discloses a limb auxiliary exercise rehabilitation device for disabled persons. The device comprises a device base, a lifting platform, a seat, a cylinder and a gear, a lifting table ismounted on one side of the top end of the device base; a lifting cavity is formed in the lifting table; the seat is arranged above the lifting table; racks are fixed on two sides of the bottom end ofthe seat; a support plate is fixed on the surface of the device base on one side of the lifting table; the cylinder is fixed to the surface of the support plate, a fine adjustment mechanism is fixed to the output end of the cylinder, a training frame is fixed to the end, away from the lifting table, of the device base, a sliding block is fixed to the bottom end of the training frame, a rotating disc is hinged to the interior of the training frame, pedal frames are fixed to the two ends of the rotating disc, and fixing bandages are sewn to the two sides of the surfaces of the pedal frames. Thefunction of adjusting the training intensity of the rehabilitation device is achieved, the training modes of the rehabilitation device are increased, and the convenience of adjustment of the rehabilitation device is improved.

The invention relates to a linkage type shoulder pushing and lifting exerciser. The device comprises a support frame. The supporting frame comprises two supporting rods and a longitudinal rod. A firsthinge seat is arranged on the cross rod. The first hinge seat is connected with a first bent pipe. The first bent pipe comprises a vertical section and a horizontal section. The free end of the longitudinal rod is connected with a second hinge seat. The second hinge seat is connected with a second bent pipe. The first bent pipe is hinged with the second bent pipe. Arm supports are horizontally arranged on the left side and the right side of the connecting rod. One end of each arm support is hinged with the corresponding supporting rod. Each arm support is connected with a first bent pipe below through a vertical pull rod. A counterweight part is arranged on the first bent pipe. By means of the device, when a trainee pushes the shoulders, the back of the trainee can incline along with pushing of the shoulders, the linkage effect of the shoulders and the back in the training process is achieved, the body of the trainee is fully unfolded, the overall training effect is improved, and thedevice is suitable for industrial production and has high practicability.

The invention discloses a targeted upper limb rehabilitation training system. The upper limb rehabilitation training system comprises a server, a processor, a computer, an input device and a display device. The input end of the processor is in communication connection with the output end of the computer through a signal line; a medical worker inputs patient information and upper limb conditions through the computer and uploads the patient information and the upper limb conditions to the server through the processor; a medical worker selects a line segment needing to be drawn by the patient according to the upper limb condition of the patient on the computer and then uploads the line segment to the processor; the input end of the input device and the display device are in signal connectionwith the output end of the processor through signal lines; the input device comprises a digital panel and a pressure sensing pen which are in signal connection with the USB interface of the processor;and the processor is in signal connection with a server through the Internet. The system has the advantages of being simple in structure, convenient to use and capable of evaluating recovery of a patient, so that medical staff can refer to the system conveniently, and a reasonable training method is customized.

The invention relates to a deep neural network method based on principal component analysis and clustering analysis, and the method specifically comprises the following steps: S1, dividing label datainto training data and test data, and the training data being used for the training learning of a model; S2, performing feature dimension reduction on all initial data features of the training data byutilizing PCA, and extracting a new principal component; S3, performing K-Means clustering analysis on all the training samples according to the principal components extracted by PCA; and S4, forminga single-layer neural network for training by taking an upper-layer data result as an input and combining the tag obtained by clustering, thereby obtaining a network weight parameter. According to the deep neural network method based on principal component analysis and clustering analysis provided by the invention, a statistical feature learning method and the neural network are combined and applied, a training mode of a traditional multilayer neural network is optimized in a training process, and a better test effect is obtained for learning of a common deep neural network.

The invention discloses an auxiliary device for treating a cerebrovascular patient, and effectively solves the problems that an existing auxiliary treatment device is single in structure and simple in mode, and cannot perform coordinated exercise on the whole body of the patient, so that the auxiliary treatment process is boring, the enthusiasm of the patient cannot be improved, and the treatment effect is very small. The device comprises a support, one side of the support is fixedly connected with a bottom plate, the bottom plate is rotationally connected with two first pedals capable of being treaded in a reciprocating mode, the support on the upper sides of the pedals is rotationally connected with two first swing rods capable of swinging in a reciprocating mode, push rods are hinged to the first swing rods, the push rods are movably connected with connecting blocks, and fixing screw rods are installed between the connecting blocks and the push rods. The connecting blocks are rotationally connected with a lifting rod synchronously moving with the pedal, the lifting rod is located in the support and slidably connected with the support, the other side of the support is fixedly connected with a seat plate, the bottom of the seat plate is provided with a rotatable second pedal capable of being adjusted front and back, and the upper portion of the seat plate is provided with a second swing rod synchronously moving with the second pedal.