20results about How to "Alleviate scarcity" patented technology

The invention relates to the technical field of artificial intelligence, and discloses a network security situation awareness method based on artificial intelligence, which comprises the following steps: acquiring network security data; performing feature extraction on the network security data by adopting a full-connection neural network based on dynamic group evolution optimization to obtain a first feature; performing feature dimension reduction on the first feature by using a self-encoding neural network based on contrast entropy to obtain a second feature; constructing a classifier model by using a fractional order neural network based on complexity perception, and training the classifier model by using the second feature; and performing network security situation awareness classification by using the trained classifier model. According to the method, the capability of distinguishing the complex network security situation is improved, a more accurate classification effect can be realized, and the applicability of the model in the network security data and the classification accuracy are enhanced.

The invention relates to a method for converting Lao characters into phonemes based on transfer learning, and aims to realize efficient phoneme conversion by utilizing a model pre-trained on other languages and finely adjusting limited Lao data. According to the method, the problem of data scarcity can be relieved, the training process of the model can be accelerated, and the generalization ability of the model on new tasks is improved. A more robust Lao character transphoneme model can be constructed through transfer learning, and a foundation is laid for subsequent speech synthesis. As Thai with high language similarity with Lao is also a low-resource language and cannot be used as a pre-training language, Chinese is selected as the pre-training language. A Chinese open source data set is firstly used for training a Chinese G2P model based on a Transform architecture, decoding end parameters of the model are initialized to a student model, fine tuning training is performed by using a Lao data set with the scale of ten thousand, and the accuracy of a Lao character-to-phoneme model is improved.

The invention discloses a rainstorm disaster event argument extraction method and device based on event text enhancement, and the method comprises the steps: constructing a rainstorm disaster event template, and defining an event type and an argument role; extracting data related to the rainstorm disaster from the microblog platform; preprocessing the data; labeling the preprocessed data to obtain a data set; generating a problem template; performing event text enhancement; integrating the problem template and the enhanced event text as model input; encoding the integrated text by using a pre-trained Chine-BERT-WWM model, and obtaining a hidden layer representation; deep features are extracted through an intermediate linear layer, and starting and ending positions of an input sequence are predicted by using a full connection layer; according to the method, the rich MRC data set can be directly utilized to enhance event extraction, the problem of data scarcity is relieved, and possibility is provided for zero sample event extraction of invisible event types.

The present invention relates to the field of artificial intelligence technology, and discloses a network security situation awareness method based on artificial intelligence, including obtaining network security data; using a fully connected neural network based on dynamic population evolution optimization to extract features from network security data to obtain a first feature; using an autoencoder neural network based on contrast entropy to reduce the first feature to obtain a second feature; using a fractional-order neural network based on complexity perception to construct a classifier model, and using the second feature to train the classifier model; and using the trained classifier model to classify network security situation awareness. The present invention improves the ability to distinguish complex network security situations, can achieve more accurate classification effects, and enhances the applicability of the model in network security data and the accuracy of classification.

The invention provides a diffusion model magnified endoscope image generation method based on pathological information guidance. The method comprises the following steps: acquiring a real image, a Mask image and a pathological information text prompt of a magnified endoscope, and constructing a training and test data set; constructing a magnified endoscopic image generation model based on pathological information guidance, and training through the training data set to obtain a trained magnified endoscopic image generation model; the test data set is used for testing, the performance of the test data set on unknown data is evaluated, and if the test data set is qualified, the test data set is a finally obtained amplified endoscopic image generation model; and applying the finally obtained magnified endoscopic image generation model to magnified endoscopic image generation work to obtain a magnified endoscopic image generated by the model. The method is suitable for generating amplified endoscope pictures, amplified endoscope pictures of different pathologies are accurately generated in a pathological information and background Mask picture guiding mode, and the problem of data imbalance of downstream segmentation and classification tasks of the amplified endoscope pictures is effectively solved.

The invention relates to a semi-supervised SAR target identification method and system based on contour enhancement and cooperative training. Specifically, according to the method, a to-be-recognized SAR image is input into a trained target recognition model to generate a corresponding target prediction result. The training process is divided into two stages: in the first stage, contour extraction and enhancement processing are carried out on an SAR image sample without a label, and then semi-supervised learning is carried out by using a FixMatch framework; and in the second stage, the weak augmented images of the label samples in the first stage are input into the cooperative training network for training, judgment results of different rounds are summarized, and the reliability of each judgment result is judged through an entropy value. The common judgment results of the cooperative training network form a sampling pool, a certain number of high-confidence samples are extracted from the sampling pool to serve as pseudo-label samples to expand a label training set of the first stage, and then the expanded label training set is used for continuing to train FixMatch. According to the method, the accuracy, the reliability and the robustness of SAR target recognition are remarkably improved.

The present invention is a method for detecting abnormal sounds that uses metadata to generate audio to predict unknown abnormalities, which solves the problem that the abnormal audio samples of the target machine are invisible to the existing abnormal sound detection model, resulting in the inability to select hyperparameters for the abnormal sound detection model and the limited performance of the abnormal sound detection system. The abnormal sample prediction strategy provided by the present invention uses the alignment of audio features and metadata information features to construct an audio generation method based on metadata information, and then predicts the abnormal sound characteristics of unknown target machine types, and generates predicted sound samples of the target machine that can be used for model training in the First-Shot scenario for the existing abnormal sound detection method. With the help of the predicted sound samples, the abnormal sound detection method can determine the optimal model in complex real-life scenarios for abnormal sound detection, thereby improving the performance and versatility of the abnormal sound detection method. At the same time, the unified sample generation model constructed by the present invention alleviates the problem of scarcity of abnormal samples in real-life scenarios, and effectively reduces the difficulty of industrial deployment of the abnormal sound detection method.

The application discloses a battery external short circuit diagnosis method based on multi-scale fusion, and belongs to the technical field of lithium ion battery safety monitoring and artificial intelligence fault diagnosis. The method only collects battery terminal voltage time series data, and constructs event level labels after preprocessing; a virtual and real hybrid sample generation module based on real voltage samples is established to generate virtual external short circuit samples and difficult negative samples; further, multi-scale voltage features such as statistics, segmentation, context, transient and morphology are extracted, and through input projection, residual feature enhancement, multi-token reconstruction, multi-head attention fusion and hierarchical prediction head, the output is the external short circuit occurrence probability, fault scene, severity and alarm confidence. The application does not depend on the measured current and temperature signals, and can alleviate the problems of real external short circuit sample scarcity, normal disturbance false alarm and high resistance early fault identification difficulty.

A multi-view finger vein image acquisition device includes a camera module, a lighting module, a driving module and a supporting module. The supporting module is provided with a rotating track. The driving module includes a motor and a rotating handle. The motor is mounted on the supporting module, and the rotating handle is mounted on the motor. The camera module and the lighting module are arranged relatively and connected to the rotating handle respectively. The camera module and the lighting module are also mounted on the rotating track respectively through pulleys. A space for placing a finger is provided between the camera module and the lighting module. The motor drives the rotating handle to rotate, and the rotating handle drives the camera module and the lighting module to rotate synchronously along the rotating track. The camera module performs multi-view photography of the finger, and the lighting module provides illumination of the corresponding viewing angle to the finger during multi-view photography. The present invention uses only a single camera to realize multi-view all-round flexible acquisition of finger veins, greatly reduces the investment cost, simplifies the equipment structure and control program, and realizes high-quality and easy-to-operate multi-view finger vein acquisition at a low cost.

The present invention provides a lithium battery surface defect detection based on multi-task learning and a model construction method thereof, which belongs to the technical field of lithium battery defect detection based on computer vision. A domain generalization neural network is constructed based on the lithium battery surface image and the steel surface defect image, and a multi-scale feature extraction module, a feature fusion module based on residual attention, and the first deep convolutional layer in the trained dual-task learning deep convolutional neural network are deployed on the background computer; an industrial high-precision CMOS industrial camera is used to collect the surface image of the lithium battery on the production line and transmit it to the background computer; the lithium battery surface image is sequentially passed through the multi-scale feature extraction module, the feature fusion module based on residual attention, and the deep convolutional layer to obtain the lithium battery surface defect detection result. The present invention solves the problem of poor generalization ability and data scarcity of existing defect detection models, and at the same time enhances the model's perception of tiny defects on the surface of lithium batteries and steel surfaces.

The application relates to the technical field of computer vision and artificial intelligence, and provides a remote sensing image target detection method based on an improved YOLOv9s, which comprises the following steps: acquiring a remote sensing image and performing a pretreatment operation; introducing a kernel selection feature fusion (KSFF) module into a neck network of a YOLOv9s model, and replacing at least one Concat structure in the original model; adding a cross-space multi-scale attention (CSMA) module in the neck network; replacing at least one SPPELAN structure in the original model with a parallel pooling feature modulation (PPFM) module in the neck network; and training, testing and evaluating the improved YOLOv9s network. Through the comprehensive application of network structure improvement, data pretreatment optimization, a semi-supervised learning strategy and an advanced training method, the improved YOLOv9s model has higher detection precision in a remote sensing image target detection task.

This invention discloses a satellite fire detection method based on dynamic exponential features and deep learning. The method includes: acquiring multi-band spatiotemporal observation data from geostationary meteorological satellites; filtering candidate pixels from the full-disk satellite data using multiple threshold conditions; extracting spatiotemporal input features and channel input features of the candidate pixels based on the multi-band spatiotemporal observation data from the geostationary meteorological satellites; and inputting the spatiotemporal input features and channel input features of the candidate pixels into a pre-trained fire spatiotemporal network model to obtain the fire detection result. This invention effectively solves the problem of high false alarms and false negatives caused by cloud cover, high-temperature ground surfaces, and vegetation interference in traditional methods, overcomes the difficulties of weak fire signal extraction and spatial positioning in mixed pixel backgrounds, and significantly improves detection accuracy, stability, and computational efficiency. It can provide key technical support for forest fire monitoring and emergency decision-making.

The present application discloses a Transformer-based image classification method, system, device and storage medium, which can make full use of a small amount of labeled data and feature aliasing technology for model training, and quickly build a high-performance image classification model in the case of scarce labeled samples to achieve accurate classification of images. The present application method includes: obtaining a collection of labeled training images, wherein the training image collection contains training images of multiple categories; training a Transformer model and a first feature classifier according to the training image collection to improve the first feature representation capability of the Transformer model and the second feature representation capability of the first feature classifier; linearly interpolating the hidden layer features of the two training images to obtain the target hidden layer features; decoding the target hidden layer features, and calculating the cross entropy loss with the category label of the training image; training the Transformer model according to the cross entropy loss to improve the third feature representation capability of the Transformer model.

The invention provides an intention recognition method and system for programming technical issues, and belongs to the technical field of software engineering.The method comprises the steps that the programming technical issues of a target platform are collected and screened; performing intention annotation on the programming technical problem by adopting an iterative statistical sampling mode, and constructing a multi-level intention classification system to obtain a basic data set; performing data enhancement on samples in the basic data set based on a large language model to obtain an enhanced data set; and training an intention recognizer through the enhanced data set, and carrying out intention recognition on the to-be-tested programming technical problem through the trained intention recognizer. According to the method, the multi-level intention classification system is constructed and the intention recognizer is trained, so that understanding of deep semantics of programming technical problems is realized, the problem of category imbalance in a data set can be relieved through a data enhancement mechanism, and the intention recognition accuracy is improved.

The invention discloses a target-level sentence simplification method and system based on label confidence coefficient weighting, which can capture nuisance of sentences on different simplification levels by training a multi-level simplification classifier, and the multi-level classification not only improves the classification accuracy, but also enables the subsequent sentence simplification process to be more meticulous and flexible. The source sentences and the target sentences of the paraphrasing data set are marked by utilizing the multi-stage simplified classifier, a large amount of pseudo training data can be automatically generated, the problem of scarcity of high-quality marked data can be effectively relieved, the cost of data collection is reduced, and when a sentence simplification model is trained, the training efficiency is improved. By introducing a label confidence coefficient weighting scheme, the influence of error label samples on model training can be remarkably reduced, the robustness and the performance stability of the model can be improved, the finally obtained target-level sentence simplification model can receive a sentence to be simplified as input and output the simplified sentence, and the sentence simplification efficiency is improved. The output simplified sentences are generally high in quality and accuracy.

The present application relates to the field of display technology and human-computer interaction, and provides a flat panel display visual comfort prediction method and system based on a multi-modal fusion model, wherein the method comprises the following steps: step one, collecting an original image set, for each original image, using a generative adversarial network model to convert an input random sequence into enhanced materials, and then fusing the enhanced materials and the original image at a preset ratio to obtain a plurality of new images; step two, collecting physiological features and physical features to obtain training samples; step three, training a multi-modal fusion model based on a stacked ensemble framework; and step four, inputting test data to obtain a visual comfort prediction result. The present application aims to solve the problems of a lack of high-quality labeled samples, a lack of coupling relationship in single-modal analysis, and poor scene adaptability in the field of flat panel display visual comfort prediction.

The present invention provides a method for generating magnified endoscopic images based on a diffusion model guided by pathological information, including obtaining a real image of a magnified endoscope, a mask image and a text prompt of pathological information, and constructing a training and test data set; constructing a magnified endoscopic image generation model based on pathological information guidance and training it through a training data set to obtain a trained magnified endoscopic image generation model; using a test data set for testing to evaluate its performance on unknown data, and the one that is qualified is the final magnified endoscopic image generation model; applying the final magnified endoscopic image generation model to the magnified endoscopic image generation work to obtain the magnified endoscopic image generated by the model. This method is suitable for the work of generating magnified endoscopic images, and uses the method of guiding pathological information and background mask images to accurately generate magnified endoscopic images of different pathologies, effectively solving the data imbalance problem of downstream segmentation and classification tasks of magnified endoscopic images.

Embodiments of the present disclosure provide an image registration method, apparatus, device, and computer-readable storage medium. The method provided by the embodiments of the present disclosure extracts feature maps of the image pair to be registered at multiple scales, and converts the displacement vector of each voxel in the moving image of the image pair to be registered at each scale into a combination of multiple displacement basis vectors determined based on the feature vector corresponding to the voxel in the corresponding feature map, so as to learn the displacement field from the paired feature maps, and generate the final displacement field based on the displacement fields at all scales, so as to make full use of the image feature information at different scales to achieve efficient and accurate image registration. The method of the embodiments of the present disclosure can learn the displacement field from coarse to fine based on different scales, so as to make full use of the potential information of different representation subspaces by fusing the displacement fields at all scales without introducing dense continuous network calculations.

The present application relates to the technical field of automatic driving end-to-end perception, in particular to an end-to-end automatic driving long-tail recognition method based on contrast learning pre-training, first, a synthetic image data with long-tail distribution characteristics is generated through a conditional diffusion model, then a fine-grained scene classifier is used to systematically organize and semantically label the generated samples, and a structured multi-modal graph-text alignment dataset is constructed; finally, the enhanced dataset and the original training set are fused, the visual-linguistic joint embedding space is optimized through a multi-task contrast loss function, and the parameter update of the pre-training model is realized. The method innovatively establishes a closed-loop optimization mechanism of generative data enhancement and contrast learning framework, effectively alleviates the data scarcity problem under the long-tail distribution scene, and significantly improves the cross-modal representation ability and downstream task generalization performance of the model on low-resource classes.

The invention discloses a hydroelectric generating set stability prediction method and system, and aims to solve the problem of insufficient generalization ability of stability prediction under extreme working conditions such as a flood period and a dry season in the prior art. The method comprises the following steps: preprocessing historical operation data containing extreme working conditions of a target hydroelectric generating set; acquiring extreme working condition data of the associated cross-domain unit; constructing a stability prediction model fusing physical mechanism embedding and cross-domain transfer learning; and inputting real-time data, and outputting prediction results of the vibration amplitude, the throw value and the winding temperature rise. The system correspondingly comprises a unit historical data processing module, a unit screening and data acquisition module, a physical mechanism-cross-domain migration prediction model construction module and a real-time data input and prediction output module. The method can supplement extreme working condition training data, improves the prediction precision under the extreme working condition, guarantees the stable operation and maintenance of the hydroelectric generating set under all working conditions, and is high in practicability.